👋 Howdy to the 1,784 new legends who joined since our last edition! You are now part of a 522,532-strong tribe outperforming the competition together.

LATEST POSTS 📚

If you’re new, not yet a subscriber, or just plain missed it, here are some of our recent editions.

⚡️ The Fintech King & BILL’s 20-Year Race To $1 Trillion. Serving small businesses all the way to 1% of GDP.
🔮 Middle-earth, Big Brother & A Philosopher King: The Palantir Story. The interesting story of Alex Karp and Palantir.
🏹 Unfinished Business, Product Intelligence & The AI Moment. An interview with Jen Taylor, CEO at Mixpanel.

PARTNERS 💫

Meetings aren’t the problem. The challenge is everything that comes after them. With Granola, the AI Notepad for people with back-to-back meetings, it's no longer about scrambling to keep up.

Take notes the way you always have. Granola works in the background, turning conversations into clear summaries, action items, and next steps. Before, during, or after a meeting, you can chat with your notes to quickly understand what needs to happen, write follow-ups, or share context with others.

Framer helps teams design, build, and launch their marketing sites lightning fast. With the ability to publish hundreds of CMS pages in a single click, operate at a global scale with seamless localization, and even host unified content across multiple domains, teams have never been able to ship faster.

Trusted by companies like Miro, Bilt, and Perplexity.

_{Interested in sponsoring these emails? See our partnership options here.}

HOUSEKEEPING 📨

Our new Intelligence product at Athyna is on fire. It's exciting because, for the first time in a while, it’s a bet that can radically change the company's trajectory. There are a number of ways we are working with partners today: recruiting AI talent (this is our existing model), done-for-you post-training (RLHF, SFT, red-teaming) with the best STEM PhDs and Masters grads, and building custom datasets.

Train your AI.

The even cooler thing is the ICP for this product is really Big Tech: the LLMs and the biggest and best application layer players in the space today. If you are an ambitious AI company and want to improve your frontier models, hit us up today!

INTERVIEW 🎙️

Luke Des Cotes, CEO at Metalab

Luke Des Cotes is the CEO of Metalab, a product agency headquartered in Victoria, British Columbia, that has been fully remote for 20 years and operates across 21 countries. Since its founding, Metalab has worked on some of the most consequential products in tech, including early versions of Slack, Coinbase, Uber, and Midjourney. In 2022, Metalab launched Metalab Ventures, a $15M venture fund, which the firm completed deploying in early 2026 and is now raising a second fund.

Luke.

Luke joined Metalab nearly 11 years ago, long after the brand had already earned its reputation, which means his job has never been to build it from scratch. It's been to protect it, extend it, and often that has meant resisting the temptation to do too much with it. That's a different and arguably more challenging task, especially for an agency whose whole value proposition is helping other companies define theirs. The venture fund adds a natural extension: Metalab was already doing diligence on founders before deciding whether to take them on as clients, now they're writing cheques too.

What does Metalab build that other agencies can't?

I think agencies are a unique beast unto themselves. People who join agencies generally love the idea of constantly building new things. You have to have that builder's mindset. Where Metalab has been able to play in a way that's incredibly unique and puts us in a different spot is the constant wave of inflection point work we do. Beyond just designing products in moments of change, often we’re being asked to reshape the actual business behind the product in our process. Very few agencies in the world actually get trusted with that process.

Source: Metalab.

I sometimes say that within six months a designer, developer, or product person who joins Metalab—after they've had two or three projects under their belt—probably has more experience shipping zero-to-one or huge monumental inflection point work than a lot of people do in their entire careers. After a few years at Metalab, you've got incredible experience in that. It's a very different way of thinking and handling the symbiotic and yet sometimes contradicting nature of business needs, complex user needs, systems thinking around UX patterns that are needed to shape new categories, while at the same time, building something that's beautiful and going to be beloved. Challenging work.

What's the biggest myth about running a top agency?

Probably that everything's easy, that once you've established yourself, you can keep going and building. I think what I've developed a newfound respect for over the years is just how both simple and complex the idea of brand is.

People believe that once you have a strong brand, everything just unlocks and you sail off into the sunset. What I've recognized is how easy it is to wreck a great brand, which I’ve seen many great companies and agencies do over the years. That's something that keeps me up at night. Building a brand is an incredibly challenging thing. The Metalab brand foundations were built before I actually joined the company, about 11 years ago.

Doors, duly opened. Source.

Since then, and especially since stepping into the CEO role, my focus has been on how you maintain a brand, how you keep what's already there, and how you look around corners to see what might become a threat to it. If you do all the right things, doors open, and it's incredible. But you need to be careful about doing too much or taking what you have for granted and destroying what you've created.

What separates good product design from world-class?

So much of that comes down to care, attention, and the harmony between brand, user experience, features, and the business itself. The same things that go into making a truly world-class product are the same things that people often can't put into words when a product feels slightly off. There are a lot of great businesses that have brought themselves to market and established themselves as trusted, but when you pull the product open to use it, it just doesn't quite feel like that brand promise was there. A lot of what we focus on is how to connect brand with the experience and product, how to deeply understand the promise of what that business is: what's the value they're trying to bring to market, why do they exist,, and then how do you translate that into both a seamless expression of that. Finding those moments to wow and delight.

Source: Metalab.

Source: Metalab.

Early Slack designs (Source).

Still, it's not just about confetti and joy. Think back to the work we did on Slack, with the sassy robot and all that - sometimes that's exactly what's needed to bring people out of their almost expected position, give them something different, make them feel uncomfortably wonderful. There's always an opportunity to do more. A lot of what we do at Metalab is figure out what the symptoms or the systems are that keep us at ‘good enough’ but not incredible, and then try to break those down. We're constantly trying new things, whether that's building in extra time for what we used to call polish and delight or finding ways through our process to drive more divergent thinking and innovation. We're constantly trying to inject new shocks into our system to force us to go beyond what is and find something new.

How do you decide which clients Metalab takes on?

That's a big one and one I'm directly involved in. There are many parts of the business where I create the systems and let them work, but our portfolio and client makeup are a big part of what I do day to day. A lot of it comes back to brand. When we look at which clients to take on, I want us thinking about what will create the portfolio of the future we want to be known for. Part of that is looking for ways to continue enhancing our skills. If there's a product challenge we just haven't seen in a while, and our team needs to push on that area, that excites us. Sometimes it's about impact and reach. Sometimes it's about matching up passion areas with the right people on the team, since we know that often produces the best outcomes.

Ultimately, for us as a services business, we're always trying to ensure our teams are passionate and excited about the work they're doing. They're here to think through something differently, to wrestle with ambiguity and the meaty challenge, not for the thousandth iteration of something that's been done so many times and just needs execution.

Source: Metalab.

So it comes down to the people, the passion, the driving conviction they have around that product and solution. Also, a lot of times it's their belief in the design process, whether they truly believe design will be a differentiator. Typically, people who come to Metalab already believe that, but we're always looking for it.

What does a typical week as Metalab's CEO look like?

Always curveballs. The areas I focus on most are probably our portfolio and hiring. On the portfolio side, I'm involved a fair bit in our pipeline, looking at the clients coming in. I'll sometimes be directly involved with clients, but we've got a great team that works with them, so a lot of what I'm doing is looking at the overarching makeup and making sure we have the right balance of interesting venture clients alongside awesome Fortune 500 enterprise clients. Also looking at how we can challenge and stretch ourselves in different ways.

On hiring, as a services business, there's nothing more important than being in tune with the type of people, personalities, and culture you're building through who you bring in. I put a lot of trust in our team and leaders to make ultimate decisions, but I spend quite a bit of time getting to know everyone we're evaluating. Overseeing whether we’re bringing the right skill sets, if we’re adding joy to our experience of building an incredible product together, or whether there are any flags to watch out for. And finally, when the teams let me, I'll get involved in product strategy. I love to dust off the old cobwebs and get involved in that side of things, but I don't spend as much time on it as I used to.

Where is product design headed over the next five years?

I think two things are happening simultaneously. There's a deepening respect for the craft, taking things from good to world-class, and I think that delta is becoming more evident and visible. At the same time, there's a deepening value placed on systems that can create huge efficiencies in the world of product and design. Where AI is having some of its biggest impacts is in the speed to translate ideas into something on paper that you can feel and work with, and also in taking what's been shaped and building systems so you can drive consistency and roll things out. So product design, along with product thinking, is heading toward a place where there will be a lot more time and energy available to put into the work of going from good to great or truly world-class.

Situationship.

Most of us internally at Metalab are really excited about the opportunity. What we've seen is a kind of rising tide. The bottom of product has risen. We're at a place where the lowest level of product should be good pretty soon. That creates an enormous opportunity to spend your time on something different. Rather than spending bits of time here or there to focus on innovation, what if the majority of your time was spent on that? How incredible could your products be? That's the drive that pushes us toward the next iteration of our mission: to push the product industry forward. How do we build products that are truly worth talking about, that people want to keep coming back to and dissecting, saying, “How did they do that?” You can't just expect that to happen. You need to build the workflows, systems, and a clear mandate for it to become a core part of your DNA.

Why launch Metalab Ventures with a $15M fund now?

I've always believed that major launches of new areas of a business should feel ambitious but natural. When teams try to force their way into new things, that's where it goes wrong. A venture fund for us felt completely natural. Our DNA has always been in the world of venture. In the early days, 15 or 16 years ago, we built incredible relationships in Silicon Valley, and startups started coming to us. We built relationships with venture funds where we could advise and help partners work with their portfolio companies to shape better experiences. Over the years, two things happened. One, we worked with incredible companies that went on to do great things (early days of Coinbase, Slack, Uber). We saw them come through our doors, worked with them, and maybe had opportunities to take equity but chose to take cash instead. There's a little regret there, but we're not going to live life in a state of regret.

Source: MetalabVC.

I still have the email from a fellow named Kris Marsalek in my inbox, excitedly telling me they'd just secured the URL for crypto.com and were building a business. Pre-seed, earliest days. We worked with them, launched the brand, the product, and now, every NBA season, I get to see my team playing against the Los Angeles Lakers at none other than Crypto.com Arena. Those are the types of opportunities I hang onto as evidence that we should be doing more than just building together. At the same time, when I talked about gatekeeping our portfolio for the future, I realized we were already doing much of the diligence work. We were already spending time understanding what makes a great founder, how they think about product, where their conviction comes from, how their internal systems work, and why they have a reason to win. We were already going through that process, and then quickly realized there was an opportunity to formalize it and take it to the next level.

We talked about doing it off the balance sheet, but ultimately decided that going out and fundraising would be the fastest way to take it seriously. It's one thing to invest your own profits.

It's another thing to take someone else's money, look them in the eyes, and say, “We're going to build a portfolio that delivers real returns.” We launched about three years ago, finished deployment of Fund One a few months ago, and we're about to start raising Fund Two. We've built internal systems so our teams actually share in the fund's upside, creating a highly aligned incentive model. The teams that work on projects that go on to be incredibly successful share in that upside.

How do you get the best out of yourself?

For me, I'm ultimately a brand builder and a people person. I care deeply about people and their growth.

I've always found that the opportunity to create growth in someone who has potential is such a unique thing—something everyone should look for. A lot of what motivates me in my work is seeing people come to Metalab, join the team, experience that growth, try new things, and challenge themselves in new areas.

So a lot of what I'm looking for in terms of the types of clients and the types of people we bring on is people who have incredible talent but have potential that needs to be unlocked, who can do more. And it's clients who have incredible potential and great ideas that need some real sifting through to help them reach it. I think that's what I do to motivate myself: put us in a position where we want more and to be better. To do that, you have to look for potential more than you look for the finished product. That excites me. The day when we're doing the thousandth version of a thing and feeling like we've been doing this for five years and it's still just happening over and over, that's the day I shouldn't be working in this space.

Extra reading

• Inside Metalab, Silicon Valley’s Best-Kept Design Secret - February, 2026

• AI Delivers Mediocre Results—By Design. So How Do You Stand Out? - February, 2026

• Designing a New Relationship with AI - November, 2024

And that’s it! You can follow Luke on LinkedIn or check out Metalab on their website to keep up with what they’re building!

BRAIN FOOD 🧠 

TWEETS OF THE WEEK 🐣 

Source.

Source.

Source.

TOOLS WE RECOMMEND 🛠️

Every week, we highlight tools we like and those we actually use inside our business and give them an honest review. Today, we are highlighting Mochary Method—an executive coaching firm that specializes in helping growing startups.

See the full set of tools we use inside Athyna & Open Source CEO here.

HOW I CAN HELP 🥳

P.S. Want to work together?

1. Hiring global talent: If you’re hiring tech, business or ops talent and want to do it 80% less, check out my startup, Athyna. 🌏

2. See my tech stack: Find our suite of tools & resources for both this newsletter and Athyna here. 🧰 

3. Reach an audience of tech leaders: Advertise with us if you want to get in front of founders, investors, and leaders in tech. 👀 

That’s it from me. See you next week, Doc 🫡  P.P.S. Let’s connect on LinkedIn and Twitter.

👋 Howdy to the 1,406 new legends who joined since our last edition! You are now part of a 520,748-strong tribe outperforming the competition together.

LATEST POSTS 📚

If you’re new, not yet a subscriber, or just plain missed it, here are some of our recent editions.

🔮 Middle-earth, Big Brother & A Philosopher King: The Palantir Story. The interesting story of Alex Karp and Palantir.
🏹 Unfinished Business, Product Intelligence & The AI Moment. An interview with Jen Taylor, CEO at Mixpanel.
🧞 Deepfakes, Trust & Selling Video To The Mainstream. An interview with Andrew Mok, CMO at HeyGen.

PARTNERS 💫

79% of enterprise buyers say generic AI voices damage brand perception—and that risk starts long before recording begins. It's in how the voice is sourced, licensed, and built.

Voices' free guide breaks down the 5 AI voice myths costing CX platforms their biggest deals. Get it before your next enterprise conversation. 

You're paying for a CRM, a sequencer, an enrichment tool, a call recorder, and an agent builder. Lightfield replaces all five.

One agentic CRM with built-in agents that build your pipeline, prep your meetings, send follow-ups, and keep records current. Connect email and calendar, and it starts working in your first meeting.

_{Interested in sponsoring these emails? See our partnership options here.}

HOUSEKEEPING 📨

If you’ve been a subscriber to this newsletter for a while, you will have seen me shout out the YouTube channel Channel 5 with Andrew Callaghan before, but I can’t stress how much I love this channel. Andrew interviews and breaks down the most controversial, NSFW, and just overall interesting topics.

Some examples are his recent trip to Palestine, inside Mexico’s cartels piece, Clavicular interview, and breakdown of the Irish language revival; I don’t know what it is about this type of storytelling, but I absolutely love it. Appointment viewing.

ZERO TO ONE 🌱

The Fintech King & BILL’s 20-Year Race To $1 Trillion

In Game of Thrones lore, there was a king who sat fifty-five years atop the Iron Throne. Known as the Wise, the Conciliator, and the Old King, Jahaerys I Targaryen reigned so long because he was fair-spoken, open-handed, and chivalrous. He was also flexible enough to adapt to conditions around him and universally loved by his people. I see similarities between Jaehaerys of House Targaryen, the First of His Name, and the Founder & CEO of BILL, René Lacerte.

Not only has René lorded over fintech for decades, but just as Jaehaerys was a fourth-generation king of Westeros, René is a fourth-generation entrepreneur. The Lacerte name has been synonymous with accounting software since the 1960s. His family built Lacerte Tax Software, which Intuit acquired for $400 million in 1998. Lacerte is to accountants what Targaryen is to dragonriders.

King René first of his name.

The story of BILL is very much the story of René, but not only René. It’s also the story of Henrique Ceribelli, whom I interviewed for this piece, beginning his career as a PM, to 20 years later running large parts of BILL’s GTM, along with Tara Wellington and Ariege Misherghi, young product leaders looking to push BILL to the next level, whom I also interviewed.

As of today, BILL processes more than 1% of GDP in the United States, and, if you ask René and his team, they are only just getting started. Today, I will break down for you where they came from, how they succeeded, and what their plans are for the future. Let’s dive in!

Familial values

While interviewing René in the lead-up to this piece, one thing became really clear to me: His family background was central to the story. Both René’s parents and grandparents had been in financial services for SMBs and accountants since the 1960s. He didn't stumble onto this problem; it was a problem he inherited at the dinner table.

Henrique recalled to me the company in the early days, when it was operating under the name CashView, and its homepage showed a simple calendar: money in, money out, bills to pay, invoices sent. A simple idea, but one that came from a long-held desire to help small businesses. Those constantly managing the balance between extending payables and bringing in receivables just to survive.

It wasn’t only a thorough understanding of what makes a small business tick that René learned from his parents. He also picked up two character traits that have served him well, all these years as an entrepreneur: grit and perseverance.

Research god mode = unlocked.

“My dad exemplified that in a very personal way,” he would tell me. “He was born with six fingers, four on his left hand and two on his right. He wanted to play piano, but nobody would teach him. So he taught himself over the course of one summer. He did it by playing every single day until his fingers bled. By the end of that summer, he had developed something remarkable; he could play any song he could hum.” As I listened to René tell this story with warmth and love, I could tell how proud he was of his father.

His mother, too: ”My mom also had a strong appreciation for being present in the moments that matter. Whether it was watching the sunset every night or stopping to look at a flower, she had that quality of actually being there.”

Beyond that, one of the core values at BILL is authenticity, something René traces directly back to his parents. “Even though they came from a different time and held different political views, there was never any judgment about people. They were always welcoming and open. I remember one moment specifically where I must have said something judgmental as a young person, and my dad turned to me and said, ‘Don't be so damn judgmental, René. It doesn't help you in life. You don't have to agree with someone, but do not pass judgment.’”

This culture, the culture of acceptance, respect, and openness, is something that radiated through my interviews for the piece, something we’ll touch on more deeply in a later section.

René’s early career

René not only had a family heritage of small business and accounting, but he also cut his teeth in the tech industry before becoming a founder. First, there was his stop at PwC in 1989, back when they were just Price Waterhouse, where René realized that he wanted to be in “high tech.”

This was followed by a role at CertiData Corporation, and subsequently Intuit, where he spent half a decade setting strategy and leading teams in various PM roles. René would join Intuit when it was just himself and development partner Martin Gates on the team. By the time he left—coincidentally, with Gates, to found PayCycle—the team was over 300 people across four sites.

PayCycle, founded in 1999, was René’s first dive into the turbulent waters of entrepreneurship, where he would spend six years, the first five as CEO, and the final year as CFO. As per his LinkedIn, the major achievements during this time leading the United States first online payroll company were: “Over 25k small businesses; 105 great employees; 5-star awards from PC Magazine and numerous accountant trade publications; customer numbers and revenue doubling every year; partnerships with Microsoft, Intuit, ATX, Digital Insight,” among loads more achievements.

Source: PCMag (2008).

Source: PCMag (2008).

Interviewee Ariege actually worked in customer service at PayCycle with René in her very first role in tech. “I accepted it because I wanted to be around smart people and because I needed the money,” she would tell me. “But I was shocked by how much I loved helping customers. Because it was a payroll company, the stakes were real. If someone's employees don't get paid, somebody can't pay their mortgage. You'd take a call with someone really fired up, and our charter was: help every customer, figure it out, because the stakes are high. That freedom to work the system in favor of the customer gave me a lot of joy. That was the environment René created.”

The DNA of BILL was a carryover of this first venture. The idea of helping the little guy do his best business in the world and to build for what René today calls the ‘Fortune 5 Million.’

Like all great stories, this one would come full circle, with PayCycle eventually being acquired by Intuit in mid-2009.

Source: Reuters.

Day zero decisions

When I asked René to reverse-engineer BILL’s success for me, he broke it down in a very simple way: “I think any software company would say it has to have three different areas of expertise. One is domain, the second would be technology, how you're going to actually execute on the domain, and the third would be how you're going to build customer trust.”

Luckily for the team at BILL, René had a healthy helping of all three: his family, his career trajectory, and the fact that he’d actually felt a large portion of the painful back-office pain while running PayCycle. Like all the best ideas, René was scratching his own itch. Armed with their new market, small team, and shiny mission statement, the question now was, how to find a wedge that would allow them to start making progress.

Incumbents in the space at the time had accounts receivable (AR) covered; FreshBooks had ~500K users and QuickBooks Desktop ~5M, so the opportunity was to enter the market by building automated accounts payable (AP) solutions for small businesses. At the time, this really was category-creating; they essentially had to educate the market that it was something they needed. Their competition was paper checks and online banking, which had simply taken the consumer banking experience and applied it to business without building any of the necessary processes around it.

The decision to focus on AP was a smart one, meaning they could build without any real competitors for years, at which point AR would come into the picture but would be added as part of a much bigger story. In the early days, René and BILL had a few aces up their sleeve that helped drive the company forward.

_{*Note: I am adding this video below not so much because it adds to the narrative, but because it’s great to see René speaking on behalf of BILL seventeen years ago.}

The accounting channel

If your accountant recommends something, you can bet your bottom dollar you’ll consider it. The trust is pre-established. To get accountants even more comfortable recommending BILL, René and the team invested very early in SOC 1 and SOC 2 certifications, non-negotiable given the nature of money movement.

The network

Businesses would receive a payment through BILL.com and then get a very convenient call: ‘Hey there. How would you like to use this to automate your own accounts payable?’ The person on the receiving end already knew who BILL was, so the familiarity led to much higher close rates.

The partnerships

René's instinct was always to partner rather than build everything himself. Their first key partnership came in 2008 with CPA.com, AICPA's division, roughly 18 months into founding the company. Today, BILL serves as the backend for JPMorgan's commercial clients, among many others.

Founded in early 2006, BILL would grow like a rocket within its first few years, until 2008, when the world would be hit by what would be called the Great Recession and the Global Financial Crisis.

The darkest chapter: 2008

Cracks in global markets began to appear in mid-2007, with the collapse of the U.S. subprime mortgage market, and intensified through to its zenith in 2008. This was an incredibly difficult time for BILL, of which I’ll dive into shortly, but it was also a heartbreaking time for René on a personal level, with his father being diagnosed with cancer in February and passing away in September. The Lehman Brothers bankruptcy, a low-water mark in market sentiment, and the stress that put René under coincided almost to the day with a moment of deep personal sadness.

Click to watch.

The next month, esteemed Silicon Valley venture firm Sequoia Capital released their ‘R.I.P. Good Times’ presentation, urging founders to make cuts to survive what looked to be a “nuclear winter” in venture funding.

See full deck.

See full deck.

It was at this point that René knew that he had to move. “I read it, and the very next day I went to the head of product and said I thought we needed to act,” René would tell me. “I asked him to think through how we would do it. Then, literally two days later, I went to the board and said I believed we needed to lay off 40% of the company and that, if we did, it would extend our runway by six months. We executed the layoffs, and even after doing that, we came within about two months of running out of money before I was able to close a new round. So it was absolutely the right call, because without it we would have been out of business entirely.”

Layering on top of this emotional pain was the loss of his father, and the fact that the parents of his wife were also sick. “She was flying to Hawaii every three weeks to be with them. We had kids who were five and three years old. As a family, we were seeing each other roughly every three weeks.” How René was able to hold it together through the difficulties was by ‘creating separation’ from the business: “Exercise has been one way to do that. Hiking is another. I also play Ultimate Frisbee once a week when I am in town and available. The point is having things that are just for you; not for the company, not for your family, just for you.”

From survival to 1% of GDP

Two months out from bankruptcy, with a grieving founder, a lean team of 15 early BILL operators went to work figuring out how to survive in a market that had utterly collapsed. What they would decide to double down on would either make or break the burgeoning technology powerhouse.

After hearing and summarizing these interviews, it seemed there were a handful of themes explaining why BILL is still here and processing 1% of GDP today.

1/ Build like you're already at scale

Most startups optimize for the short term, whereas René optimized for what BILL would need long into the future. In a money-movement business, that means compliance, controls, and audit infrastructure built ahead of demand, not after.

”We were never just thinking about what we needed this year; it was always about what we needed for the endgame,” René would tell me. “One example of that is the way we approached moving money from day one. Moving money is a trust-based business… we treated that with the utmost care, operating as though we were already moving 1% of GDP and needed to do it the right way from the start.”

2/ Show product discipline

Henrique originally joined BILL to work on receivables, which was put on ice for a full three years while they focused on dominating the payables product. “The team was small enough that we could only tackle one thing at a time,” Henrique would go on to tell me.

The same discipline can be seen in the fact that they’ve never moved out of financial operations; they could easily have moved into bookkeeping, payroll, tax, but they didn’t. They stayed laser-focused. And again, you see this discipline in their ICP. They are focused on Main Street, not Wall Street (or Sand Hill Road), and it’s remained that way forever.

Staying true to the small business ICP meant that as the product surface expanded, the scope remained the same. It would have been very tempting throughout the journey to release the self-imposed shackles on a defined ICP and start building for the enterprise. But they didn’t; René and the team were building for the little guy.

3/ Leveraging their network effect

Next is a very unique distribution lever BILL has been able to pull on since the very early days: the network effect. This is something I’ve thought deeply about recently, thanks to my mentoring role as part of the Startmate S25 cohort. The company I’ve been working with is FlowGo, a specialized AI agent swarm (dolphins have pods, lions have prides, agents have swarms) for financial operations in Asia. The company is only four months old and growing like a rocket by leveraging the same effect BILL has used for decades.

Demo Day.

Network (from FlowGo deck).

In the early days, BILL saw that their customers wanted to pay suppliers electronically (ACH); the issue was that suppliers were skeptical of handing their bank details out to random payers. Henrique's idea was to take the receivables product (which they'd already shelved) and make it free. Let suppliers sign up for an account to receive payment without exposing their bank info. As a bonus, they could send invoices through it, too. That free product is now an eight-million-entity network; one billion documents stored; trillions of dollars moved.

To put a direct line through the compounding lesson: when BILL goes on customer visits today, half the room has already been a payee on the platform. The network sells the network.

4/ Own customer touch points

BILL learned the hard way, early, that something as important as customer touch points should stay as close to home as possible. Originally, BILL needed to integrate with every major accounting platform (because the network is software-agnostic by design). They brought in third parties to build those integrations so they could move faster, but the integrations were inconsistent, and customers refused to deal with anyone but BILL when something broke. They pulled it back and rebuilt from scratch—painfully—mid-journey.

They also tried to lean on accounting partners to sell into firms, but without aligned incentives, they found that this GTM didn't scale. To fix this, they built an internal sales team that owns both the sale and the implementation directly with firms. “When the product is complex, and the customer relationship is critical, you need to maintain direct control rather than delegating it out and hoping for the best,” Henrique would tell me.

These experiences led BILL towards only ever outsourcing what's commoditized, but owning what touches the customer.

5/ Acquire if and when necessary

Lastly, in recent years, BILL has not shied away from a strategic acquisition. Divvy was the first move. While René and the team had long owned the bill-pay side of cash-out, they didn’t own the card-side spend. And in this case, rather than build it, they acquired Divvy to bring their customers and intellectual horsepower in-house. Their next addition was the Australian, mobile-first, accounts receivable firm Invoice2go. This strengthened the AR side and gave them greater international reach (especially within APAC).

And most recently, Finmark, which was acquired for an undisclosed amount, giving them new financial planning and forecasting chops. This was a smaller tuck-in to extend BILL into the planning layer above transactions.

With five compounders and 20 years of payment volume, document data, and supplier identity stacked on top of all of it, you start to understand why BILL's AI bet has the data underneath. Which is exactly where we're going next.

AI = Don't automate. Delete.

While every public company scrambles to fit AI into their earnings calls, many are soon to find out that their data and distribution don’t actually help them in the AI world. BILL, on the other hand, has served customers for twenty years, building a rare combo of: 1) a serious data moat, 2) an ever larger trust moat. The goal for BILL is larger than building a handful of helpful workflows and insights. René and the team are looking to delete work altogether.

One easy example that anyone doing business in the U.S. can attest to is the W9 collections agent. In the old world, a customer asked their supplier for a W9, that supplier sent it, and the work helped exactly one customer. With BILL's network, the supplier uploads it once into their free account, and every other customer of that supplier instantly has it on file. The agent's work multiplies across the network. As Henrique would put it, "The agent effectively solved the problem for multiple payers at once because of the network multiplier effect.”

Source: BILL.

Same logic on invoice coding. BILL learns from how the AP team codes a vendor, but they also learn from how everyone paying that vendor codes it. Collective intelligence example numero uno. Plus, BILL has a global vendor directory built up over years of processing documents. When a new customer joins and drops in a doc, BILL already knows the vendor, payment methods, context, and more. Collective intelligence numero dos.

Source: BILL.

So, I hear you asking, ‘How does Bill actually build AI internally, you ask?’ Well, allow me to explain. BILL runs its AI strategy across two tracks. First, the transformational layer: rethinking the entire product to be AI-native, not just AI-enabled. Holistic, cross-org, owned by an AI platform team. And second, the feature layer: every individual squad asking how AI changes their corner. Each squad is super lean, ~6 engineers, dedicated to one outcome. No matrix sprawl. No AI czar.

Underneath the two tracks sits a very specific operating model. And it's the same one BILL runs across all of product, starting with planning. Every year, BILL runs a three-tier cycle: a three-year strategic view at the top; a one-year software plan in the middle; team-level objectives at the bottom. Each layer feeds the one below it.

Each year then breaks down into quarters, with each having a six-week pulse check (are we executing, are we hitting the metrics, do we need to pivot?), then a full reset at the end of the quarter. The inputs feeding all of this aren't just product team intuition. Tara walked me through it:

Every initiative at BILL is run by a single dedicated squad. The core is a trio: an engineering manager, a designer, and a PM. Around them sit the engineers, usually about six. Research and product marketing partner in, but don't embed. This is how you set your org to build a world-class product, folks.

Want more on AI?

By the way, if you’re interested in understanding more deeply the AI usage in the world of finance, BILL released a pretty slick report, which you can enjoy below.

Download the report.

Culture, and why people stay

As a culture-obsessed founder myself, I feel as if I have a reasonable radar for the culture of an organization, along with at least a snippet into the mind of its leader. And although BILL has been through difficult periods in the past—2008, COVID, other rounds of layoffs driven by market forces—I got the real feeling that BILL, under René, was a great place to work. Here are a few notes from the interviews.

Playbook / how you can apply this

• Build like you're already there: René ran BILL as if it already moved 1% of GDP from day one, and that bar set the company.

• Bet the company on accountants: They backed the accounting channel in 2006 and never blinked. Today, over half of BILL's customers still arrive through firms.

• Build the network sideways: Suppliers wouldn't share bank details, so BILL made receivables free. That pivot turned into eight million network nodes.

• Treat trust as infrastructure: SOC 1 and SOC 2 came absurdly early for a startup. In 2008, cloud skepticism was the difference between yes and no.

• Make the CEO the chief customer: René still uses BILL daily and emails product teams about bugs he spots. The obsession trickles all the way down.

• Own what touches the customer: They outsourced integrations and sales early. Both broke. Both got rebuilt in-house. Both then worked.

• Lead with authenticity, even at scale: René cried with the team after the 2008 layoffs and still does at all-hands. Twenty years on, the trust holds.

Summary

Thinking back on the long, successful reign of King Jaehaerys, one thing stood out to me above all else: his ability to endure and determination to survive. The wise old king’s time was full of personal tragedy, political crises, rebellions, and great sicknesses. But through it all, he stayed alive and kept moving forward, something that sounds eerily similar to the protagonist of today’s story.

Business is often the ability to stay alive longer than the competition; out-survive first, then thrive second. BILL has passed their survival stage, and are now set in thrive mode, with 4-5% market penetration, allowing for 1% of the U.S. gross domestic product. But they are not done. Nor are the unsung heroes of this story: small business owners. So long as people are out there taking the risk to build their business, René Ledgerborn, the First of His Name, The Unbankrupt, King of the Fintechs, and Protector of the Fortune Five Million, will be there to serve them.

Fun facts

• Cashboard → CashView, BILL: Cashboard couldn't get the URL. CashView didn't quite stick. The third name turned out to be the trillion-dollar one.

• BILL.com is an OG YouTube channel: If you want to go and browse their back catalog, you can find videos from as far back as 14 years ago.

Source.

• The ‘holy shit’ demo: René's dad watched the early demo and said, "Holy shit, René, you've come up with something I've needed my whole life."

• The bank's bank: JPMorgan runs BILL as the backend for its commercial clients, managing financial operations. Yes, that JPMorgan.

Extra reading / learning

• Cybersecured: Vanta’s Zero To One - March, 2025

• Hyperbound Goes Hyperdrive: A Zero To One - February, 2026

• Tracksuit’s Bootstrap To Series B - February, 2026

And that’s it! You can keep up with René, Henrique, Tara, and Ariege on LinkedIn or check out BILL on their website.

BRAIN FOOD 🧠 

TWEETS OF THE WEEK 🐣 

Source.

Source.

Source.

TOOLS WE RECOMMEND 🛠️

Every week, we highlight tools we like and those we actually use inside our business and give them an honest review. Today, we are highlighting Framer*—the site builder trusted by startups to Fortune 500.

beehiiv: We use beehiiv to send all of our newsletters. Apollo: We use Apollo to automate a large part of our 1.2M weekly outbound emails. Taplio: We use Taplio to grow and manage my online presence.

See the full set of tools we use inside Athyna & Open Source CEO here.

HOW I CAN HELP 🥳

P.S. Want to work together?

1. Hiring global talent: If you’re hiring tech, business or ops talent and want to do it 80% less, check out my startup, Athyna. 🌏

2. See my tech stack: Find our suite of tools & resources for both this newsletter and Athyna here. 🧰 

3. Reach an audience of tech leaders: Advertise with us if you want to get in front of founders, investors and leaders in tech. 👀 

That’s it from me. See you next week, Doc 🫡  P.P.S. Let’s connect on LinkedIn and Twitter.

👋 Howdy to the 1,293 new legends who joined since our last edition! You are now part of a 519,342-strong tribe outperforming the competition together.

LATEST POSTS 📚

If you’re new, not yet a subscriber, or just plain missed it, here are some of our recent editions.

🏹 Unfinished Business, Product Intelligence & The AI Moment. An interview with Jen Taylor, CEO at Mixpanel.
🧞 Deepfakes, Trust & Selling Video To The Mainstream. An interview with Andrew Mok, CMO at HeyGen.
🗝️ Marketing Has Too Many Humans In The Loop. An interview with Doug Tallmadge, Co-Founder & CEO at Gradial.

PARTNERS 💫

Framer helps teams design, build, and launch their marketing sites lightning fast. With the ability to publish hundreds of CMS pages in a single click, operate at a global scale with seamless localization, and even host unified content across multiple domains, teams have never been able to ship faster.

Trusted by companies like Miro, Bilt, and Perplexity.

Your customers want to make their agent workflows event-driven, so you're losing deals if your API lacks reliable webhooks. Webhooks lead to deep product engagement, stronger retention, and integrated workflows, but setting up and managing webhooks is more complex than it seems.

Svix has solved all the hard problems so you don't have to: retries, idempotency, security, compliance, and more.

_{Interested in sponsoring these emails? See our partnership options here.}

HOUSEKEEPING 📨

Just stumbled upon the trailer for Onslaught by A24, and it looked very gory, but also super interesting. It had me thinking: horror is having a real moment. Weapons was awesome (and is looking to become a franchise), Sinners was even better—I fucking loved that movie—and my pipeline still has Backrooms, Obsession, and a handful more horror flicks to come.

Funny how things like that work. I never thought I’d be devouring the horror genre like I am today. Widow’s Bay was awesome, too. Anyway, shoot me a reply if you have any other film recommendations. Otherwise, onto today’s piece!

DEEP DIVE 🕵🏻

Middle-earth, Big Brother & A Philosopher King: The Palantir Story

Palantír, etymologically speaking, means ‘to keep watch from afar.’ To some Palantian bulls today, this can be loosely translated to being a guardian angel. A benevolent entity that’s forever keeping us safe. To the bears, it means Big Brother. And no, not the one with the intruders. The dystopic, Orwellian kind right out of the pages of 1984. To me, the jury is still out.

The founder of Palantir, Alex Karp, is akin to Cosmo Kramer with a PhD in philosophy and clearance to brief the Pentagon. He appears, makes a bold or baffling statement, then disappears again—part CEO, part myth. You’re never quite sure who he is or what he does. And maybe that’s the point.

Today, it’s my job to change that. My aim is to demystify both Palantir and some of the grey matter inner workings of its founder, Alex. In preparation for this deep dive, only moments ago, I finished Alex’s new book: The Technological Republic. An ode to an American and technological landscape unlike anything I have ever read. More on that later.

For now, let's delve into the fascinating story of Palantir, one of the most contentious companies in recent living memory.

Palantír = ‘palan’ (far), ‘tir’ (keep watch)

When I mentioned etymology in our introduction, you may have wondered what I was talking about. You won’t find the term in a regular Google Translate search. The name comes from the Quenya language, one of J.R.R. Tolkien’s invented Elvish languages. In the world-building of Tolkien’s Middle-earth, Quenya isn’t any old Elvish language; it’s the language of the High Elves.

During the time of The Lord of the Rings, it was an old language, mainly used for ceremony, lore, and poetry. Sindarin was more common for everyday Elvish conversation. Think of it in the way that we speak English today, but with Latin in some ceremonial examples. The Palantír itself is a seeing stone—a lava lamp-looking, magical crystal orb that allows users to communicate across vast distances and see events (past, present, and sometimes future) in other parts of the world.

Interestingly, like our subject today, the Palantír allows rulers to see far-off places and events, but the signal can be manipulated, often dangerously so. Palantir of 2026, for example, gives military and government agencies supercharged data visibility. But like the seeing-stones wielded (willingly or not) by the likes of Aragorn, Saruman and Peregrin Took, more data doesn’t always mean more clarity.

If you are like me, you are probably sick of this analogy by now. So let’s journey from the make-believe back into reality, with a college friendship that would prove meaningful to the future of our sovereignty.

A PayPal Mafia connection

Alex Karp recently stated that, "The death and pain that is brought to our enemies is mostly, not exclusively, brought by Palantir.” A stunning statement from the CEO of a Fortune 500 technology company. But how did we get here? It all started in 2003, with a young Peter Thiel, fresh off PayPal’s sale, setting out to build software that could “reduce terrorism while preserving civil liberties.”

In 2004, a then-35-year-old Thiel funded the creation of a prototype of what Palantir may look like, built by PayPal engineer Nathan Gettings and Stanford students Joe Lonsdale and Stephen Cohen. That same year, Thiel would hire Karp, a former colleague of his from Stanford Law, as the firm's first chief executive. All five men would be considered the co-founders of Palantir.

Alex Karp, Peter Thiel.

For a company that is now one of the darlings of the technology world—the current price-to-earnings of Palantir at the time of writing is ~150x (Google’s is 17x by comparison)—they struggled to find investors early on. Silicon Valley legend and Sequoia Capital chairman Michael Moritz was reported to have been so bored with the initial pitch that he drew doodles in his notebook through the entire conversation.

The team thought they might struggle to stay afloat until they had their first real external backer, the CIA. Yes, that’s right, the Central Intelligence Agency of the U.S. of A has a venture fund. It’s called In-Q-Tel, and it's used to fund innovation that keeps the CIA and other intelligence agencies at the cutting edge of technology.

Obviously joking.

Alongside their first external backing, Thiel would go on to write a much larger check of $30 million in fresh funding on behalf of himself and his now vaunted venture firm, Founders Fund.

By this point, Palantir, a company that builds data platforms to combat bad actors, was well and truly established and was quite literally inside the walls of the intelligence community itself.

The early product, in a nutshell

A lot of people have now heard the name ‘Palantir,’ but what they do is still somewhat of a black box. In the early days, Palantir had a singular focus: to sell products to governments to keep the United States and its allies safe. That, in the very early days, looked like a product called Gotham, an ultra-secret platform for spies, soldiers, and police. Gotham was essentially an intelligence OS helping connect siloed databases to visualize networks of bad actors. Head to their landing page for Gotham, and the real value proposition will be clear to see: Powering the kill chain.

_{*Palantir gets credited (rightly or wrongly) for helping to take out Osama Bin Laden. While Palantir Technologies' software and data analysis tools are reported to have played a role in the intelligence gathering that led to the location of Osama bin Laden, the company did not directly kill him, nor do they discuss the level of their involvement.}

When I take a look at their product frames today, it looks like Google Maps meets a bad guy CRM with a left-to-right, pipeline-staged flow of targets. It’s like an enterprise full of death. This is now, but we are talking about then, when their product’s value was in ingesting vast amounts of data and making it all make sense. The engineers were working at the edge of what was technically possible with Gotham—mixing relational databases, graph tech, and UI customization tools years ahead of their time.

The product was often built by what would come to be known as Forward-Deployed Engineers (I will use FDE for the remainder of the post today). This would be an incredible engineering talent, embedded with a Palantir customer. They would often sit in the office four days a week, working alongside in-house teams to provide the ideal solution. It was always bespoke.

In the early days, analysts often viewed Palantir as a consulting firm. ‘EY of war’ or some such notion. The idea was that a FDE would come in and build a novel solution that would ingest, connect, and visualize datasets from a staggering variety of formats: PDFs, chat transcripts, phone logs, GPS records, etc.

The magic was in link analysis, drawing connections between these disparate data points, like suspects, phone calls, locations, and bank accounts. If this sounds like the KPMG of the killing zone, it was. Palantir, though, would over time roll any useful-for-all technology it built into its global platform, giving the product more universality and moving Palantir further away from Deathloitte (consulting) and closer to Slaughterforce (platform).

_{*If you are expecting me to stop with the consulting analogies here, you are in for a long read. I came prepared.}

This early product looked more like middleware than something a Figmanian product design wizard would create; one part data cleaning, one part knowledge graph, and one part UI hackathon. The learning curve was quite brutal. But it worked. And once you were in, the product made you feel like you had superpowers, connecting dots others missed. Dots that would likely save countless lives in the process.

Finding PWF (Product-Warzone-Fit)

A mid-2010s TechCrunch piece, citing leaked documents from 2013, showed that Palantir were working with a who’s who of governmental and military organizations. From the CIA, DHS, and FBI, to the Marine Corps, Air Force, and even the National Center for Missing and Exploited Children.

Source: Palantir.

They were expanding their footprint rapidly, and the once VC-repellent startup, off the success of their first winning product, Gotham, was now starting to become the belle of the private markets ball. And they were just getting started. By the midway point of the decade, with the high-level client list growing, the budding intelligence juggernaut had raised a total of somewhere north of $1.2 billion and was rumored to have had a private market valuation of $20 billion. Here is a snapshot of the rounds raised in the formative years from the mid-2000s to mid-2010s.

By this point, Palantir wasn’t just raising big rounds—it was landing mission-critical wins in the field. In Afghanistan, for example, their platform was used by U.S. troops to uncover IED networks. By pulling together phone logs, safehouse locals, vehicle movements, and prior attacks, soldiers could finally make sense of enemy patterns. One commander reportedly called it “the only software that actually works in theatre.”

But it wasn’t just military actions where Palantir were excelling. In New Orleans, Palantir (controversially) partnered with local police to try and stem gang retaliation shootings. The system helped detectives identify which affiliations were most likely to escalate, and which weren’t worth chasing.

During the devastating earthquake in Haiti in 2010, Palantir helped gather victims' names, using their data integration tech to organize and plan during the humanitarian crisis. And at the National Center for Missing and Exploited Children, Palantir’s system helped triage massive volumes of unstructured case data—PDFs, chat logs, images, emails—surfacing patterns in child trafficking investigations that would’ve otherwise taken weeks to detect.

It was messy, expensive, and barely scalable. But it worked. In that moment, Palantir wasn’t a product company. It was a human-plus-machine strike force embedded at the edge of government operations, outperforming legacy vendors one case file at a time.

From Gotham to Foundry (From spies to Salesforce)

By the mid-2000s 2010s Palantir looked like a success by any metric. Growing revenue, deepening influence, and investors lining up to take part. But their next move was unconventional. If they cut their teeth with Gotham, their ultra-secret platform for spies, soldiers, and police, the follow-up was to turn to the commercial market.

Their next act would be Foundry, a newer platform that took Gotham’s DNA and repackaged it for enterprises like Airbus and Morgan Stanley, promising any big company ‘Palantir-grade’ data integration without needing a room full of forward-deployed mega-nerds to make it work.

The early stages of growth in the commercial space proved challenging, though, with their first product, Palantir Metropolis, an “unmitigated failure” (even if Joe Lonsdale disagrees). Palantir also had to learn how to function in the enterprise GTM market. They needed to manage long procurement cycles (which I guess they were somewhat used to), the need for glossy UIs, and the easy integrations, areas where upstarts like Snowflake or Tableau had already been excelling.

Since no one builds technology for governments and militaries, you can often get away with a duct-taped product approach.

This was part of the reason Palantir needed the fleet of forward-deployed geeks in the first place, even flying devs at times to Afghan war zones with laptops to help troops on the ground.

“To do things that do not scale,” as Brian Chesky would put it, can be powerful, but to enter the commercial market, they needed some level of scalability. The key unlock for Palantir’s commercial ambitions wasn’t just Foundry; it was a lesser-known tool called Apollo. Think of Apollo as the silent operator behind the scenes: it lets Palantir deploy software anywhere—cloud, on-prem, air-gapped servers in bunkers—and update it continuously. In other words, it turned Foundry from a big, clunky install into a flexible, enterprise-grade platform. Suddenly, companies didn’t need to host Palantir’s FDEs to make it work. That was the moment things started to scale.

Source: Palantir.

And it paid off. Airbus rebuilt its entire digital factory system on Foundry (see the full impact study here). “We are saving no less than one aircraft out of service per day, largely based on predictive alerting, but also in relation to expediency and heightened accuracy,” the Director of MX Ops Planning & Analytics would say. If that sounds like gobbledygook (/ˈɡɒb(ə)ldɪɡuːk/) to you, don’t worry, me as well. I asked ChatGPT to reframe this quote as if it were needing to be explained to a 3-year-old, and the results were quite good.

That’s not all. BP ran Foundry across its hundreds of rigs to optimize oilfield output. And Palantir even landed a spot inside Ferrari’s F1 operations team. Not bad for a company that used to only brief three-letter agencies behind closed doors.

Source: Palantir.

Source: Palantir.

The new Military (Tech) Industrial Complex

If the European Renaissance between the 14th and 17th centuries was a period—or ‘rebirth’—of interest in art, literature, and philosophy, then the period we are entering today may be a similar resurgence in interest in defense-related startups. Defense was a dirty word in technology in the early days of Palantir; they were a relative outlier when courting the Pentagon.

But now other players have entered the space (and space), such as the likes of Anduril—founded by ex-Palantirians, and also named in Quenya, after Aragorn’s sword, Andúril, the Flame of the West—SpaceX, and even OpenAI. If Lockheed Martin, Raytheon, Boeing and Northrop Grumman owned the defense budgets of the early part of this century and the latter part of the last, then you can absolutely expect the new big four of Palantir, Anduril, SpaceX, and OpenAI will control a large part of the next.

These firms are now beginning to blur the line between where Sand Hill Road ends and the military-industrial complex begins. In 2025, Palantir joined SpaceX and Anduril in pitching a futuristic ‘Golden Dome’ missile-defense shield–a SpaceX-led plan for hundreds of sensor satellites and AI targeting, sold on a subscription model rather than the Pentagon’s traditional procurement. Not only is there money to be made for these new entrants into the space, but the chances are the celebrity combination of Elon Musk, Sam Altman, Palmer Luckey, and Alex Karp are likely to inspire the next generations of builders to turn create the next generation of Team America: World Police, rather than an app that helps you checkout from your online shopping cart 0.001% faster.

Palantir’s journey is often framed as David vs Goliath—a startup taking on defense giants of the past. In fact, Palantir had to sue the U.S. Army in 2016 just to be allowed to compete on a major intel system contract, arguing the Army unfairly favored its usual vendors. Palantir won that case, a landmark that cracked open the door for tech companies in defense. I, for one, am all for it. I had ChatGPT whip up a quick comparison chart for us below.

These new kids aren’t just playing the same game better—they’re playing a completely different game. While the old primes build once and bill forever, the new breed ships code weekly, with AI, autonomous systems, and global-scale infrastructure. And while the old guard relies on 1970s procurement doctrine, the new players run like software startups: agile, iterative, accountable to outcomes, not process.

In a world where warfare is turning invisible (cyber), autonomous (drones), and predictive (AI), it won’t be the company with the biggest missile that wins. It’ll be the one with the best model weights. The New Age of Defense has begun.

Alex Karp, Philosopher King

The New York City-born son of a Jewish pediatrician and an African American artist, Alexander Caedmon Karp (born October 2, 1967) is one of the most fascinating men in the world. He is a dyslexic, half-black, half-Jewish, populist-left, single at 58, ski-obsessed, modern-day philosopher.

Dr. Karp, as his Palentirian (they call them that) disciples know him, is part-man, part-monk. A man who spends his days doing Tai chi, skiing across a number of his resort-side holiday homes, and killing bad guys (allegedly). He is a man just as comfortable discussing German poetry as he is analyzing geopolitical tensions. Time magazine, after including him in the 2025 edition of the world's 100 most influential people, would call Dr. Karp "the embodiment of a new kind of Silicon Valley billionaire: an unashamed techno-nationalist who evangelizes Western power."

Alex was raised in a heavily progressive, ‘leftist’ family, and took much of his political leanings from his parents, who were activists in his childhood, fighting for civil rights by day and social justice by night. For someone who has spent his formative years working side by side with right-wing figure Thiel, you’d be surprised to know he still leans left, having donated to the Harris campaign during the 2024 election.

Alex Karp’s politics are unique. As I mentioned in the introduction, in preparation for this piece, I read Dr. Karp’s book, The Technological Republic, and to break the fourth wall for a moment here, he may be my spirit animal. I believe he is one of the true leaders today who is principled enough to stand up for what they believe in. You may not like what he says, but I, for one, am glad that at least someone is out there creating conversation. Creating debate.

To understand Alex Karp’s politics, you have to start in the seminar halls of Frankfurt, at one of Europe’s most serious philosophy schools, where people sit around debating big questions like power, truth, and what makes a just society. His main teacher? A famous German thinker named Jürgen Habermas, basically the Yoda of political philosophy.

It wasn’t your typical college education. Think less 'frat partying' and more 'arguing about morality in four languages.' That time shaped Karp’s brain in a big way. He came out of it not as a left-wing activist or a right-wing hardliner, but as someone who sees the world through his own bespoke lens.

One of the things I found interesting about Alex is his harsh assessment of the current technology landscape. He believes that we are not building anything of substance anymore. His book hammers this point home and really acts as a call to action for people who want to build something meaningful in the world.

I can understand why the average young Stanford grad may not want to work at Palantir, which uses terms like “kill chain” on its website. But would it be a better, more values-driven life to go work at Meta, Google, TikTok, or Netflix? If you are acting with intellectual honesty, I’d say no. We can pretend as if bad guys in the world do not exist, but the fact of the matter is that they do, and the more I think about it, the more I believe in the mission of Palantir.

Another shockingly raw statement from Alex is his thoughts on ‘The West,’ an idea he frequently discusses in his book.

Alex is a progressive. A true left-leaning hippie. But this statement would be shot down in many liberal circles around the world. But is he wrong? And if he is wrong, how wrong is he? As someone who is liberal myself, I like to be accepting of all other peoples and their religions. And I have always somewhat subscribed to an idea I believe I heard from Sam Harris many years ago about the difficulty of the idea of ‘truth.’ The truth is really in the eye of the beholder.

But most of you reading this will also believe, I’m guessing, in basic human rights, an idea that Western values hold dear. And some other areas do not. Regardless of his stance and what you think of it, I appreciate that he has the courage to say it. And to be fair, I may agree with him.

“I actually am a progressive,” Karp has said. “I want less war. You only stop war by having the best technology and by scaring the bejeebers—I’m trying to be nice here—out of our adversaries. If they are not scared, they don’t wake up scared, they don’t go to bed scared, they don’t fear that the wrath of America will come down on them, they will attack us. They will attack us everywhere.”

Palantir’s long list of controversies

If you thought Alex Karp was controversial, he’s nothing compared to the broader societal and philosophical quandaries that Palantir stares down the barrel of on a daily basis. One of the first controversies arose in the early 2010s, when Palantir’s tools were used internally at JPMorgan Chase to monitor employees’ emails, calls, and movements.

Around the same time, it was leaked that Palantir pitched cyber operations to discredit WikiLeaks and its supporters, including Glenn Greenwald. The proposal sparked an uproar, and Palantir quickly issued a public apology.

Two of the more contentious aspects of Palantir are its work with Israel and the U.S. Immigration and Customs Enforcement agency, better known as ICE. Palantir won a $41 million contract to build ICE’s Investigative Case Management system way back in 2014.

A system designed to track, arrest, and deport undocumented immigrants. This has now turned into their recently announced ‘ImmigrationOS.’

Palantir has also supported the Israeli Defense Forces with targeting and operational software. Its role in the Gaza conflict drew scrutiny amid allegations of war crimes and AI-enabled surveillance of civilians.

Source.

But the controversy doesn’t end there. To go really deep in my research, I even pivoted from controversy to conspiracy, listening to tin-hat foil wearers from the Candace Owens podcast to Sam Tripoli to understand what the kooky side of the internet thinks. And let me tell you, it’s dark.

I think it seems, from the outside looking in, that the conspiratorial-minded believe that Palantir founder Peter Thiel planted JD Vance into his seat in the White House as his meatspace avatar that he uses to control us by way of election nudging secretly, mass surveillance, military AI, and elite influence ops—all part of a Palantir-powered cathedral of soft authoritarianism. Dark AF, right? Hell, some people think Peter Thiel is the literal Antichrist.

Source.

It doesn’t take a conspiracy theorist to paint Palantir in a pretty dark brush. I am very libertarian, but I am also not a total doomer when it comes to Trump and Palantir’s Big Brother-esque, Orwellian surveillance state.

At the end of the day, Palantir is either the last line of defense for Western civilization, or the software stack for an authoritarian future. Maybe both. But in a world where the threats are increasingly asymmetric, invisible, and real-time, it’s no surprise that the most serious governments want Palantir in their corner.

The success playbook

• Win trust from the most demanding customer: If you can build tools that work in war zones and counterterrorism units, selling to a Fortune 500 looks like child’s play.

• Embed inside your customer’s operations: Don’t just sell software, sit next to the analyst using it. Palantir’s FDEs co-built the solution, making the product indispensable.

• Hire the zealots: Palantir hired not only the best engineers on the planet but also those who are ideologically aligned with the mission. These are super-employees.

Source: Lenny’s Newsletter.

• Secure long-term contracts: Focus on contracts that are sticky, high-margin, and nearly impossible to rip out. Defense and spending are recession-proof.

• Build products for problems: Gotham and Foundry weren’t designed to please users; they were built to solve mission-critical chaos. Function first, UX later.

• Run toward complexity, not away from it: While Silicon Valley chased consumer dopamine, Palantir went after hard problems; terrorism, pandemics, logistics at war-scale.

• Back yourself with ideological capital: Peter Thiel’s network gave Palantir more than cash; it gave them air cover, conviction, and powerful allies who share the mission.

• Create a mythos: Palantir cultivated secrecy, intellectual elitism, and anti-consensus swagger. The result? A company people argue about and remember.

• Play the long game: Palantir stayed private for 17 years. No quarterly earnings pressure. Just slow, methodical, relentless product compounding.

Looking into the Palantír

Palantir was admitted into the S&P 500 in 2024, and it seems that the company’s future cannot be brighter, with a market cap at the time of writing of $298 billion dollars. It’s a good time to be polishing up that kill chain for the Palantirians today.

Source.

But I also do not expect them ever to be far from controversy. A constant meme that reminds us of just how closely intertwined capitalism is with the afterlife today. I leave you with a note from Dr. Karp himself: “The future belongs not to those who style themselves as creators, or even to the merely brave, but to those who actually build something substantial and lasting.” Whether you like them or not, Alex Karp and Peter Thiel are doing just that.

Fun facts

• Middle-earthen-English: The name Middle-earth comes from the Old English “Middangeard”, meaning ‘the inhabited world between heaven and hell.’ The same root word that gave Norse mythology its “Midgard.”

• Welcome to The Shire: The company’s Palo Alto office features Lord of the Rings décor and is nicknamed the Shire (the Washington office is Rivendell).

• No sales team required: For years, Palantir bragged about having zero salespeople. Just engineers in suits pitching the Pentagon.

• IPO after 17 Years: Most unicorns IPO in 8 years. Palantir? Took 17. Karp thought markets were “too short-termist.”

Extra reading / learning

• How Tech Giants Turned Ukraine Into an AI War Lab - Feb, 2024

• Alex Karp Has Money and Power. So What Does He Want? - August, 2024

• Palantir - Is it a bubble? - January, 2025

And that’s it! Alex doesn’t have social media accounts, but if you’d like to learn more about Palantir, you can check them out here.

BRAIN FOOD 🧠 

TWEETS OF THE WEEK 🐣 

Source.

Source.

Source.

TOOLS WE RECOMMEND 🛠️

Every week, we highlight tools we like and those we actually use inside our business and give them an honest review. Today, we are highlighting Granola*—an AI-powered notepad that takes, summarizes, and organizes meeting notes without using intrusive recording bots.

beehiiv: We use beehiiv to send all of our newsletters. Apollo: We use Apollo to automate a large part of our 1.2M weekly outbound emails. Taplio: We use Taplio to grow and manage my online presence.

See the full set of tools we use inside Athyna & Open Source CEO here.

HOW I CAN HELP 🥳

P.S. Want to work together?

1. Hiring global talent: If you’re hiring tech, business or ops talent and want to do it 80% less, check out my startup, Athyna. 🌏

2. See my tech stack: Find our suite of tools & resources for both this newsletter and Athyna here. 🧰 

3. Reach an audience of tech leaders: Advertise with us if you want to get in front of founders, investors and leaders in tech. 👀 

That’s it from me. See you next week, Doc 🫡  P.P.S. Let’s connect on LinkedIn and Twitter.

👋 Howdy to the 1,895 new legends who joined since our last edition! You are now part of a 518,049-strong tribe outperforming the competition together.

LATEST POSTS 📚

If you’re new, not yet a subscriber, or just plain missed it, here are some of our recent editions.

🧞 Deepfakes, Trust & Selling Video To The Mainstream. An interview with Andrew Mok, CMO at HeyGen.
🗝️ Marketing Has Too Many Humans In The Loop. An interview with Doug Tallmadge, Co-Founder & CEO at Gradial.
🛡️ The Agentic CRM For AI‐Native Companies. An interview with Keith Peiris, Co-Founder & CEO at Lightfield.

PARTNERS 💫

ZeroClick lets you sell your product to AI agents. One in two visitors to your site is already an agent, but most businesses have no way to sell to them.

ZeroClick changes that by turning any API or offering into an agent-purchasable service with x402 & MPP support, a machine-readable storefront, agent transaction analytics, and robust pricing controls.

Founding a company is easy, scaling it is hard. Operational systems break, people problems increase, and your calendar keeps getting fuller.  Mochary Method is an executive coaching firm founded by Matt Mochary—coach of Sam Altman and Brian Armstrong, and writer of The Great CEO Within—that specializes in helping growing startups.

The Mochary coaching team, who coaches companies like AngelList, Lovable, and Midjourney, is offering a complimentary coaching session to founders and executives for a limited time.

_{Interested in sponsoring these emails? See our partnership options here.}

HOUSEKEEPING 📨

I know we are still a long way from the 2028 U.S. election, but I’m placing my bet now on the eventual winner: Jon Ossoff from Georgia. 39-year-old Ossoff speaks with the gravitas of a leader and has been a shining light pushing back on the corruption of the Trump administration.

There is a lot of time between now and 2028, but Vanca lacks the charisma and is too attached to Trump’s administration to win. Rubio could do well if he ends up with the Republican nod; something I could stomach, even as a liberal. The Democrats may run AOC, or Newsom, or someone else, but if they run Ossoff, he wins in a landslide. You heard it here first.

INTERVIEW 🎙️

Jen Taylor, CEO at Mixpanel

Jen Taylor is the CEO of Mixpanel, a San Francisco-based product intelligence platform founded in 2009 that serves around 29,000 customers globally, including LG, Experian, and Airbnb. Backed by $277M in total funding and valued at $1.1B following a $200M Series C in 2021, Mixpanel tracks how users behave inside digital products at a company that reached an estimated $210M in annual revenue in 2025. Jen joined in September 2025, succeeding Amir Movafaghi after his nine-year tenure. Before Mixpanel, she was President at Plaid and before that Chief Product Officer at Cloudflare, where she helped take the company from $100M to $1B in revenue. Her career also includes roles at Salesforce, Facebook, and Adobe.

Jen.

What makes Jen's situation genuinely interesting is that she walked into something already working. Mixpanel had double-digit annual growth, consistent profitability, and a product freshly transformed from standalone analytics to an integrated platform. Her mandate is acceleration, full stop. She spent two decades as a product and marketing leader who used tools like Mixpanel every day and always felt they fell short: too hard to start, too limited, too easy to ultimately ignore. Now she's running one, at exactly the moment AI is about to blow the category open.

What problem is Mixpanel solving for product teams today?

Mixpanel helps product teams understand how their users experience their application, so they can identify what's working and what's not, and continue to improve the product, delight their users, and grow their business.

Take a ride-hailing app, for example. They'd instrument their mobile application to understand how new users engage when they first open the app, how they convert through the onboarding funnel, and then, over time, how individual users or cohorts continue to drive engagement and retention. That foundation lets you layer in experimentation, so the same app might use us to run an A/B variation of its onboarding funnel.

Source: Mixpanel.

And then there's session replay, where if they identify an error somewhere in the application, they can actually watch short video clips of what the user was experiencing to help them debug and figure out what to do next.

How does the AI-powered side of Mixpanel work?

If you build on that ride-hailing example, historically, all of that instrumentation, dashboard creation, and analysis has been manual. What we've embraced with AI falls into three buckets. The first is AI turbocharging our core value proposition. Within the application today, people can use chat to query, create a dashboard, instrument their product, and summarize session replays. A lot of those use cases are directly inside the product. The second is using AI technologies like MCP to integrate Mixpanel's capabilities wherever people already work, whether that's Claude or Slack. So users, regardless of their technical understanding of analytics, can get the information and insights wherever they live.

The third is the new frontier: helping people who are building and delivering AI-based applications figure out how to define and measure success. Think about the transition from web to mobile. You had different user paradigms and different definitions of what success looked like. We're seeing the same shift happen with AI-powered applications. In a mobile app, success might be a purchase flow. In an AI-powered application, it's more non-deterministic, and you have to identify different signals along the way. So those are the three buckets. AI turbocharging the analysis, democratising access to it, and helping people define success for AI-powered applications.

What's the most challenging part of converting that into real-world analytics?

Being totally transparent, we're still in the very early days. This is the stuff happening in the lab right now. Within that, there are three things we're helping people who are building agent experiences think about. The first is about how an agent experiences their site or application. Today, we've all been building largely for humans and not really thinking about what it means when a technology-driven, non-human user is the one experiencing an application. In a world of AI, you may need to optimize for that. So, as you do the analysis, you need to be able to understand both things separately.

Source: Mixpanel.

The second is looking for different signals. This is where it gets more speculative. A lot of how you look at different flows and usage patterns comes back to helping people think about what success could look like before you can even start to measure it. And the third is providing an ongoing, living feedback loop so people can continue to optimize and adjust the weightings of the models or data sources they're using underneath, to keep improving the application.

What excited you about the Mixpanel opportunity? And how did it come about?

I was excited about this opportunity because I have unfinished business. Most of my career has been as a product and marketing leader in B2B SaaS, in places like Adobe and Salesforce, where the ground truth was always the data and the analytics. I've used tools like Mixpanel my entire career, but I always felt they underdelivered. Too hard to get started, the analysis was too limited, and, frankly, they were too easy to ignore. The moment we're living in with AI really changes all of that. Both from what's now possible and the urgency behind it. People are shipping faster with AI, and they desperately need insights to figure out whether what they're building is any good. So we're on the edge of a massive opportunity, and for me, it's my life's work to figure out how to make this happen for people, because I've wanted it my whole career. As for how it came about, I've spent a lot of time in and around the industry.

I started putting it out into the universe that I was interested in stepping up into a CEO role and looking for opportunities to have more impact and challenge myself further. I was fortunate enough to get connected to this opportunity through a friend.

Do CEO roles at interesting tech companies come up often? Is there a lot of competition?

You're spot on. When you filter for companies where the underlying business is strong, where you have genuine passion and fit, and where you're bumping into the right people at the right time, it really is a needle in a haystack. A lightning strike moment. That's why I feel so fortunate to be here. A lot of people who step into the role I stepped into walk into an organization, open a cupboard, and have skeletons fall out.

A few of the 29k.

I'm immensely grateful to my predecessor, who stepped out to spend time with his family and follow his own passions. He built a company that's a tank. I walked into a business with 29,000 customers around the globe, solid financials, and a healthy, growing operation. For me, the mandate was to come in with my experience and my lens and figure out how to drive this AI transformation and be a catalyst for innovation.

What do your weeks look like?

No two weeks are the same. My first few months were really about a listening tour, but the structure hasn't changed much since. I love customers.

They are the whole reason I've done everything I've done in my career, and they're the whole reason I'm in this company. About a third of my time is with customers in some form, whether that's here in the Bay Area, across the US, or globally. Another third is really with the team. The most important work I've been doing in my first nine months is building and rebuilding the leadership team, bringing in new people, and setting those leaders up for growth and scale. And then the last third, which varies, is spent on the craft of the industry.

Experimenting and tinkering myself, whether that's playing with AI or riffing on ideas with the product team, trying to stay on the edge of what's happening in the space and bringing that energy back into the organization.

What operational cadences have you brought to Mixpanel? And what have you kept?

The way I was brought up as a leader, and the way I like to lead, is to identify goals and objectives, then empower and delegate to the leaders under me to figure out how to get there. One of the best methods I've found for doing that is built on the back of the Salesforce V2MOM. That's the framework Salesforce has used since the beginning to set the highest-level goals within the organization and cascade them down. You set the goals, objectives, and metrics at the start of the year, the CEO signs up for it, staff signs up for aspects of it, and it flows all the way down. I've brought some version of it into almost every organization I've worked in, and it's one of the things I really brought to Mixpanel to kick off the year. It creates strategic clarity, the right level of alignment, and most importantly, it gives teams the empowerment and autonomy to figure out how to get there themselves.

The Salesforce V2MOM, defined. Source: Visible.

The other thing I focused on was introducing Monthly Business Reviews. I really value that cadence of getting together as a leadership team and looking at the business as a whole. You'll have engineering looking at what's happening in sales, marketing looking at what's happening in operations, and that comprehensive, collaborative view has been really important. Outside of those, the ritual that already existed and that I've kept is the bi-weekly company All Hands. I've seen it in almost every organization I've been in, and for distributed teams especially, the ritual of coming together, sharing quick wins, and talking about where you're struggling is incredibly powerful.

How do you manage implementation from a change-management perspective?

The first thing was adapting it to the organization. I was fortunate to have some Salesforce alumni at Mixpanel who helped articulate what was possible and helped me think through the adaptations we'd need to make it culturally appropriate. We don't call it the V2MOM here. Steph Robotham, our interim CMO, adapted it for us. She calls it the ‘Go Plan’, and I love that. Steph took the first pass at drafting it. At our staff offsite in Q4, the leadership team discussed our financial and strategic goals and objectives. Steph then translated that into the Go Plan framework. My new VP of Operations started around the same time and was critical in operationalizing it, creating the right accountability mechanisms and the right level of visibility. We're still tweaking and refining as we go. Lotte, our VP of Operations, is working with Claude and other tools to build an automated dashboard that shares updates on this with the whole organization.

Interestingly, I don't think Mixpanel had a strong heritage of company-level goals. One of the first things I heard when I joined was that people really wanted to understand our strategy and objectives.

The Mixpanel leaders.

They wanted to understand what winning looked like. So the organization was actually hungry for something like this. You always have to be careful bringing frameworks in from the outside and look at the room you're in, but the reception has been genuinely great. I'd say we launched the Go Plan this year. Fully adopting and living it will probably take two cycles before it really soaks into the fabric of the organization.

Has anything surprised you since stepping into the CEO role?

The most surprising thing about my job is that it's the one I interviewed for. Going back to what we said about the needle in a haystack, many of my peers who've stepped into a CEO role have walked into an organization, opened a cupboard, and had skeletons fall out. My predecessor left a genuinely healthy organization, and the conversations I had with the board during the process were very direct, very clear-eyed, and truthful in both directions. There've been very few surprises and nothing of substance. Also, I was surprised by the office. I hadn't actually visited before my first day, since the whole interviewing process is confidential. The day before I started, our Head of People said, "Come meet me, I'll show you around." I walked in and thought, ‘Wow’. It's right next to the Ferry Building, right on the pier. You walk outside, and there are seagulls and ferries. It almost feels like Australia.

What was the first big thing you decided to tackle once you got your feet on the ground?

The first thing I did was hire an interim CMO. As I was onboarding, it was clear that the organization really didn't have a marketing function and hadn't had a strong heritage of one. That played out in near-term and longer-term ways. In the near-term, there was no clear articulation of who we were or what we stood for in the market. No amplification of that out to the world. People didn't really know what Mixpanel felt like out there.

Source.

So I brought Steph in, and she's been absolutely transformational. In October of last year, she sat me down and said, “Here's the scoop. Next spring, we're going to do a series of customer events around the globe, starting in San Francisco, to relaunch Mixpanel and launch all of our new AI capabilities.” I thought she was crazy. We hadn't even built the products yet. I wondered what narrative we were going to create and how we were going to go around the globe. And here I am, four events into a 16-city global tour, doing exactly that. Those big, audacious goals—that's what I needed from day one. Someone to create an anchor and barrel through the organization with me.

How do you think about leadership?

My leadership style has always been to hire people who are so smart I'd ultimately want to work for them, and then figure out how to delegate, empower, and make them successful. I've always been a team athlete. I grew up playing sports and genuinely love being on teams. I love the diversity of thought and perspective. What I'm trying to create in my leadership team is a place where people feel ownership and accountability, but also feel empowered to stretch their legs.

I believe firmly in a collaborative culture where people feel safe to disagree. But they have to disagree, commit, and move on. I always say the privilege of a bad decision is that you usually get the opportunity to make it again. Very few one-way doors. More broadly across the organization, I really try to hire for and encourage curiosity. Curiosity is how you take a diverse environment and make it a genuinely inclusive one, because people are curious about understanding each other's points of view and feel invited to bring that to the table. It is also the core of how you drive innovation. It's the ‘how might we do this?’ or ‘what would happen if we did it this way?’ questions that tend to become the springboards.

Who's on your direct team, and what do your management cadences look like with them?

My direct reports are the Chief Revenue Officer, Chief Marketing Officer, CTO, General Counsel, Head of People, Head of Finance, and VP of Operations. With each of them, I do weekly one-on-ones, in person whenever possible. A good portion of our leadership team is based in San Francisco, and we all come into the office every Wednesday. They may come on other days to spend time with their teams, but Wednesday is the constant. We also have a weekly hour-long staff meeting, which the VP of Operations sets and runs. Outside of that, I'm essentially on Slack with people all the time. Right now, I'm spending more time with the engineering, product, and design team because we're short on a CPO. But I'm in Slack channels with each leader individually and in groups across different topics.

Mixcursion time.

Great Place to Work in India.

There's a running thread with my CTO and CRO where we're constantly talking about what we're hearing from customers, what we're thinking about building, and passing each other podcasts, wherever we happen to be in the world. On top of that, we do monthly MBRs: a corporate MBR, a people team MBR with the e-staff, and I'll sit in or listen to recordings of the QBRs within the go-to-market organization.

What are Mixpanel's midterm goals? And where do you think the category is headed by 2030?

I'll start with where the industry is headed, because it provides context for where we're going. The industry has already moved from a world of best-of-breed standalone tools, where you'd have a separate product analytics, a separate experimentation, a separate session replay offering, to integrated platforms that give customers a more complete picture. We made that transformation over the past couple of years, and it was already underway when I walked in. What I think is happening now is the next shift. As the roles of product managers, engineers, and designers get blurrier and blurrier, as everyone starts to occupy different parts of each other's workflow, people are building things faster with AI. But it's absolutely critical that as they're building, they understand whether or not what they're building is any good.

That's the transformation we're making, from product analytics to a product intelligence platform. We're going to be the solution that helps all of those builders understand what's working, what's not working, and what to do about it. Powered by AI, but also through ongoing product innovation to keep deepening the understanding of users. Two shifts are driving this: from product analytics to product intelligence, and from SaaS to AI native. We're already starting to see both play out.

Source: Mixpanel.

Looking further out, I increasingly see Mixpanel as a critical data infrastructure for building and delivering products. And you'll see us integrate more deeply with additional data sources, including CRM solutions and revenue systems, so that the intelligence we can provide gives a much deeper context—not just the user, but the product and the business as a whole. Where I hope all of that gets us is to continue accelerating growth and delighting users, from large enterprises to innovative new startups.

How do you personally use AI in your CEO workflow?

Claude and Cowork are basically attached to me. That's where I spend a lot of my time. I create projects within Cowork for different areas: personal coaching and monitoring, helping me mentor and coach my team, and organizing and delivering projects. For example, I have a CEO coach project where I've built in context and asked it to give me very specific feedback, taking into account the Go Plan and how I've shown up on Slack, and basically giving me weekly updates on how I'm doing, what's working, what's not, and where the opportunities are for me to step in and deliver more as a leader.

Source.

I also have a project for the open CPO search, with the job description and interview guides. Every time I have an interview, I come back to it with information about the candidate to think about how to tailor the conversation. And then I use AI as a sparring partner. If I'm writing something like a board memo, I'll use it to challenge myself about what I'm missing, where I'm going to get caught. I've also fed in the personas of the people I work with regularly, so I can ask what I'm likely to hear from a particular person and think around corners before I walk into the room.

How do you get the best out of yourself?

Three things. The first is spending time with customers. I do what I do because I love customers. Every conversation energizes me and grounds me. I genuinely love helping people think through how to build and grow their business through the products they deliver. The second: I work with people I love. Life is short, man. I've gotten to a point where I really care who I work with.

I want to have fun, I want to enjoy it, but beyond that, we're building this thing together. There are hard days and exciting days, and I want to be doing both with people I genuinely care about. The other thing I need in the people around me is someone who will tell me what I need to hear, not what I want to hear. When you're a CEO, a lot of people really, really want to tell you what you want to hear.

Energised.

The third: I try to set aside some time every day when I'm not thinking. I'm a morning person. Whether it's a dog walk or a run, I do my best thinking when I'm not thinking. My brain needs to off-gas, to release some of what I'm carrying. And almost always, I'll come back from a dog walk with a voice memo or a post-it note that I may never look at again. But that's where a lot of the synthesis happens, just giving my brain a chance to marinate and digest everything.

Extra reading

• Mixpanel Introduces Mixpanel AI, Delivering Always-On Product Intelligence - May, 2026

• Why I joined Mixpanel as CEO: A new era in analytics - February, 2026

• A New Chapter for Mixpanel: My Conversation with CEO Jen Taylor - September, 2025

• 7 Leadership Principles for Product Managers Lessons from Jen Taylor - June, 2021

And that’s it! You can follow Jen on LinkedIn or check out Mixpanel on their website to keep up with what they’re building!

BRAIN FOOD 🧠 

TWEETS OF THE WEEK 🐣 

Source.

Source.

Source.

TOOLS WE RECOMMEND 🛠️

Every week, we highlight tools we like and those we actually use inside our business and give them an honest review. Today, we are highlighting BILL—Your AI-powered financial operations platform.

See the full set of tools we use inside Athyna & Open Source CEO here.

HOW I CAN HELP 🥳

P.S. Want to work together?

1. Hiring global talent: If you’re hiring tech, business or ops talent and want to do it 80% less, check out my startup, Athyna. 🌏

2. See my tech stack: Find our suite of tools & resources for both this newsletter and Athyna here. 🧰 

3. Reach an audience of tech leaders: Advertise with us if you want to get in front of founders, investors and leaders in tech. 👀 

That’s it from me. See you next week, Doc 🫡  P.P.S. Let’s connect on LinkedIn and Twitter.

👋 Howdy to the 2,698 new legends who joined since our last edition! You are now part of a 516,154-strong tribe outperforming the competition together.

LATEST POSTS 📚

If you’re new, not yet a subscriber, or just plain missed it, here are some of our recent editions.

🗝️ Marketing Has Too Many Humans In The Loop. An interview with Doug Tallmadge, Co-Founder & CEO at Gradial.
🛡️ The Agentic CRM For AI‐Native Companies. An interview with Keith Peiris, Co-Founder & CEO at Lightfield.
📡 The Last Industry Software Hasn't Eaten. An interview with Hermann Frank, Co-Founder & CEO at Gigs.

PARTNERS 💫

Security teams need better outcomes, not more tools. Learn what modern MDR should deliver, what questions to ask, and how to evaluate providers with confidence in this practical buyer’s guide from Arctic Wolf.

Athyna is smart hiring for teams that need the best talent now—not next quarter when the role’s gone cold. The difference is AI-assisted matching paired with proper human vetting, so you get people who fit the role and the team.

Less back-and-forth, less cost, built to scale. That’s it.

_{Interested in sponsoring these emails? See our partnership options here.}

HOUSEKEEPING 📨

It’s an exciting time to be an NBA fan right now, with the Knicks being back at the peak of basketball life, the rise of Victor Wembanyama upon us, and a brand new team—and accompanying set of narratives—for LeBron James. For me, LeBron’s move to the Sixers was a shock, but a super exciting one.

Source.

For those of you who are pure NBA nerds like I am, I want to see where you think this Sixers team will stack up; I’ll start by biasing the answer with my prediction: the NBA Finals, but losing to either the Spurs or the Thunder. But what about you…

Whichever way it goes. It will no doubt make for great drama along the way, and you can bet your bottom dollar I’ll be down for it. Now, onto today’s interview!

INTERVIEW 🎙️

Deepfakes, Trust & Selling Video To The Mainstream

Andrew Mok is the CMO at HeyGen, the AI video generation platform that recently surpassed $200M in ARR, turned profitable by Q2 2023, and has since racked up more than 31 million signups. HeyGen raised $60M in a Series A led by Benchmark at a $500M+ valuation and was named one of Fast Company's Most Innovative Companies in 2026. Before HeyGen, Andrew spent over a decade at Turo, the peer-to-peer car-sharing marketplace, joining as one of the earliest employees in 2012 and becoming CMO at 29. Under his watch, Turo grew from 54,000 users in two cities to over seven million users across 10,000 cities, with revenue growing more than 250x.

Andrew.

Here's what I find interesting about Andrew. He's a self-described left-brained person who never planned to end up in marketing. He came up through analytics and growth, and it shows in how he thinks about the role. At HeyGen, he's wrestling with a problem that most AI companies wave their hands at: the vast majority of people still can't make video, and the AI companies supposedly fixing that have spent more energy going viral with deepfake stunts than actually building trust. He's also the guy building HeyGen's marketing strategy around Aristotle's principles of persuasion, which sounds like a stretch until you hear him explain why a 2,300-year-old rhetorical framework maps almost perfectly onto a modern demand generation funnel. It kind of does.

What problem does HeyGen solve?

We're a platform that allows people to turn their ideas and their expertise into videos very easily. We live in a video-dominated world. The vast majority of our traffic and user engagement is on video content.

Source: HeyGen.

Yet, the vast majority of people are actually locked out of video production and the content game, either because it's too expensive, they don't have the video editing and production skills, or they don't like how they look and sound on camera. All these people are locked out of this format, which is the predominant way for people to communicate and tell stories with one another. Our platform exists to solve that problem.

Who is your ICP today?

Right now, we're seeing that our target audience is actually solo and small business operators. If you think about it, it makes a lot of sense because these are the types of people who don't have huge budgets. They don't have huge marketing teams, can't hire a big video production agency or a big video team, but they want to get in on the content game. They want to be able to build their brand, sell their business, and share their stories and expertise with the world via social media. HeyGen is a very affordable tool that allows them to do that. I think our lowest plan is only $29 a month.

Source.

I actually do 3-5 customer interviews every week, and it's pretty remarkable that the people I chat with are really not what you would expect. It's a lot of people who are either starting their own business or already running their own business. Real estate agents, travel agencies, a gentleman who's helping his wife sell custom jewelry that she makes. It's people who have been watching social media and all this video content grow very rapidly, and they've essentially just been sitting on the sidelines because they don't know how to do it themselves or they can't afford it. HeyGen is the entryway that allows them to do that.

Source: HeyGen.

What does video as a medium look like right now?

We've defined our core customer problem as ‘we live in a video-first world’. Video accounts for 82% of all consumer internet traffic. The average person consumes 17 hours of video content per week. YouTube, obviously the main format for video, sees one billion hours of video watched daily, which is insane. And then there’s social engagement. Social posts with video see up to 10x more engagement.

I think we also know from reading about algorithm updates on X and LinkedIn that posts with videos get surfaced more in the algorithm, and they get more engagement. The trend is really only accelerating in that direction, not slowing down, at least not yet.

Was there friction getting into a category like AI video?

With all new technology, there is built-in skepticism, especially at the beginning. But with AI, it's probably a hundred times more than any other technology curve, and it's very understandable. People are very scared and hesitant about the technology. A lot of the other cinematic AI video models haven't really done the industry justice because the things that have gone viral are like Will Smith eating pasta or a deepfake of Brad Pitt doing whatever. Those types of things really erode credibility because they make it seem like this is a tool that's not practical (it's used essentially just for fun or for social media viral clips) and not safe. Brad Pitt or Will Smith can get a fake video made of them. It just showcases the worst of what this technology can be used for.

Source: HeyGen.

That's been a real problem for us, rebuilding trust with everybody to show that HeyGen does not operate in that territory. We have the most robust trust, safety, and verification process. We do not allow deepfakes. We go through rigorous detection, both automated and manual reviews, to make sure that the identity and the avatar you're creating is actually you and not somebody else. We censor for profanity, sexual content, suggestive content, political content, things like that. It takes a lot of education because people just bucket us as ‘AI video’ and assume all these nefarious things are happening or that we have no guardrails. We have to educate and show them that we're trying to build this the right way with guardrails and safety in mind. It's a bit of an uphill battle given everybody's built-in predisposition.

How’s the infrastructure of HeyGen?

The foundation of our entire platform is our proprietary avatar model, which is developed in-house. We have our own best-in-class AI research team. Generally, in the industry, it's recognized as the best, fastest, and cheapest. If you go to ChatGPT or Claude and ask what the best AI avatar model is and who's best at creating AI avatar videos, nine times out of ten the LLM will say it's HeyGen. We're known for having the best model in the industry. That underpins everything.

On top of that, we build a whole bunch of bells and whistles. There's a very comprehensive UI. We have a web app as well as iOS and Android apps. We also integrate with other models if you want to use them. For example, we integrate with Seedance, which is the leading cinematic video model right now. If you want a really exciting intro or outro, or you want some cinematic camera movement in between shots, you can use Seedance to fill that in, and our user interface stitches everything together.

HeyGen’d Andrew.

Source.

We also recently launched something really interesting called ‘HyperFrames’. It's an open-source motion graphics engine. Right now, it actually lives outside of the HeyGen product, so it's this free open-source thing we've built for the community, and a lot of people have built workflows into it. But obviously, the long-term plan is to integrate that back into the core product. You'll get this amazing motion graphics engine, the avatar video; you could combine it with other cinematic video models, and then we become the best place for you to stitch, end to end, these videos with the highest quality and the best experience.

What does a typical week look like for you as CMO?

I'll start by saying that working at HeyGen is unlike any other company I've ever worked at. At prior companies, I could answer that question very well to a T, but that's not what it's like here, and I think a lot of people working in AI companies would say the same. It's not that there's no routine and there's no process, but things are moving and changing so quickly that you have to be ready to adapt at any moment.

Source: HeyGen.

The number one thing I try to build around on a week-to-week basis is being grounded in the customer. I mentioned I talk to three to five customers a week. Most of those are onboarding calls where I'm learning about them but also actually helping them solve issues with the product. A lot of ideas and action items typically come out of those, whether it's revisiting our audience strategy, thinking about a marketing campaign, a new channel we should test, or working with the product team to share feedback about something a user surfaced. That core workstream spins off a lot of additional workstreams that can take up a lot of the week.

On the marketing-driven side, I focus on demand generation: things we're doing to build the brand, build awareness, build engagement, and build brand affinity. And then demand capture: performance marketing, growth marketing, converting attention and eyeballs into signups and subscriptions. There are lots of activities underneath each of those, and there's a lot of day-to-day, week-to-week work executing on all of it.

What is working right now for HeyGen?

Most of HeyGen's growth to date has been driven by launching cool and interesting new capabilities or products. The marketing team's job is to amplify those launches and amplify those capabilities, and then people come because the technology is really exciting. There's this insatiable appetite from a large group of people to get ahead of the curve, learn about new technologies, and figure out how to integrate them into their workflow. That has powered a lot of HeyGen's growth to date.

Can’t stop, won’t stop.

As we look ahead to building a long-lasting, durable business, there's actually a lot more we need to do to educate and convince the mainstream user who is not actively looking for AI tools to try HeyGen and give us a shot. A lot of what it takes to do that goes back to building trust in AI video as a category and in HeyGen specifically.

What I think about when it comes to building trust is that I've read and heard about a lot of marketing frameworks, and the one I always go back to is the one from thousands of years ago, stated by Aristotle on the core principles of persuasion. ‘Ethos, pathos, and logos.’ Appeal to the heart, appeal to the mind, and build up the credibility of the speaker. I think a lot about organizing our marketing streams along those lines: building heart, appealing with logic, and building up the credibility of the company and our stewards.

What does the amplification of a product launch look like?

For the most part, it's a social media-first launch, and we're a video company, so the core centerpiece is a great video. With the video, we try to strike the right balance between marketing sizzle and practical utility, which I think is actually a bit of a different strategy than a lot of other companies. A lot of AI company product launches that I see tend to really over-index on one or the other. It's all sizzle, crazy flash, colors, all these characters, but when you go to the actual product, it couldn't be further from the truth. Or it's all the way on the other end of the spectrum: very practical, a step-by-step walkthrough, but kind of boring and not capturing people's imagination. We really try to strike a balance and get the best of both worlds.

Beyond that, we post on our handles on X, LinkedIn, Instagram, TikTok, and YouTube, where we have big followings. We also work with creators to make their own content using our tools. For example, when we launched the latest version of our avatar model, Avatar 5, we posted on our channels and also worked with notable creators, giving them early access to the feature so they could create avatar videos with their own likeness and share them with their followers. That's a big part of our playbook.

What does your team look like?

HeyGen is unlike any other company I've worked at before. At old companies, you'd think about building it with your C-suite people, your VPs who own specific domains, their directors, their ICs, and so on.

We are purposefully building a very flat structure where there aren't all these layers of management and where everybody is expected to be a contributor. Everybody's an IC, including myself. I would estimate I probably spend about half of my time or more on what would have previously been considered IC work. Obviously, a lot of it now is with agents and working with agents through the IC work, so I can be a lot more effective than I was in the past.

Source.

The majority of the team actually just reports directly to me, and it's been working really well. It allows me to be a lot closer to the work, and I think it allows each team member to get the coaching and guidance they need to do their best work. I still do one-on-ones, but I do them shorter and more frequently. I used to do one per week with each team member. Now I do one to two 30-minute sessions per week per person, and sometimes we skip it if there's nothing to cover or we can just handle it async.

How are you using AI agents day-to-day in your work?

We have an agent that lives in Slack. We developed a forked version of OpenClaw so that it's safer and more secure, and our security team has been the pioneer in setting it up. It basically exists as another team member in Slack. We have a bunch of these agents on Slack, and if you look at our Slack instance, it's kind of equal parts conversations from real people and equal parts conversations from AI agents, and everybody's interacting. It's pretty amazing, actually. The one dedicated to the marketing team is called Mara. She does a lot.

I found that as a leader and an executive, you spend a lot of your time doing updates and readouts because it's important to give everybody visibility and connect the dots. I have workflows set up with our agent to do all of that for me, and I just review it.

Source: HeyGen.

On a weekly basis, it generates this long report. We put it into our internal knowledge management system, where it creates a little web app. Of course, I use HeyGen, so I generate a talking video of myself. It's 90 seconds, with me explaining what happened, our current priorities and key insights, and it includes motion graphics and visuals. It generates all of this based on input that other team members put together. They also generate it with agents. It essentially creates this whole cascading workflow.

The great thing about it is that it becomes this knowledge base that keeps track of everything happening within the company and the team. If I ever forget something, or the document doesn't have enough detail, and I want to ask, "What was John working on last week?" or "John mentioned something about podcasts, what did he say and what were the numbers?", I can just have a conversation with our agent, and it usually gives me the right answer. I almost think of it as this brain that knows everything. There's no bottleneck anymore in terms of information flow because the agent is all-knowing and connected to everything. You should be able to have context about pretty much everything going on just by having a conversation with the agent.

How do you get the best out of yourself?

For me, it all starts with having a really great morning routine and getting your day off to a running start. I used to do this thing where I would just wake up and immediately get to work, and I would find that it just tires you out over time. I would save anything personal or family-related for later in the day when you had no energy, and you just don't feel that great about yourself.

So I actually wake up early now, and that's when I do the me-time stuff. That's when I exercise, I spend time with my three-year-old daughter and make her breakfast before dropping her off at school every day. By the time I've started my workday, I'm ready to commit the rest of the day.

I feel great because I've done all the personal stuff, and I feel like I've been able to enjoy my daughter. I feel very ready to commit myself to work. And at work, it's really very simple. I just feel so incredibly intellectually curious about everything that we're doing that it doesn't necessarily feel like I'm just looking forward to the next meeting or counting down the hours. It's actually just fun and interesting. Especially right now with everything going on in AI, I actually feel very lucky. It's like I'm getting paid to play with all these tools and learn about all this stuff that I would have done on my own anyway. I feel very energized by the intellectual curiosity I have about everything that's going on.

Extra reading

• Why we build HeyGen - May, 2026

• HeyGen’s New Avatar V Lets You Clone Your Face in 15 Seconds and Generate Unlimited AI Videos - May, 2026

• HeyGen’s new enterprise stack: API, CLI, and MCP for the agentic era - April, 2026

• Revolutionizing AI Video Creation with Prompt-Based Magic - February, 2026

And that’s it! You can follow Andrew on LinkedIn or check out HeyGen on their website to keep up with what they’re building!

BRAIN FOOD 🧠 

TWEETS OF THE WEEK 🐣 

Source.

Source.

Source.

TOOLS WE RECOMMEND 🛠️

Every week, we highlight tools we like and those we actually use inside our business and give them an honest review. Today, we are highlighting Vanta*—the security and compliance platform companies rely on to stay audit-ready without losing their sanity.

beehiiv: We use beehiiv to send all of our newsletters. Apollo: We use Apollo to automate a large part of our 1.2M weekly outbound emails. Taplio: We use Taplio to grow and manage my online presence.

See the full set of tools we use inside of Athyna & Open Source CEO here.

HOW I CAN HELP 🥳

P.S. Want to work together?

1. Hiring global talent: If you’re hiring tech, business or ops talent and want to do it 80% less, check out my startup, Athyna. 🌏

2. See my tech stack: Find our suite of tools & resources for both this newsletter and Athyna here. 🧰 

3. Reach an audience of tech leaders: Advertise with us if you want to get in front of founders, investors and leaders in tech. 👀 

That’s it from me. See you next week, Doc 🫡  P.P.S. Let’s connect on LinkedIn and Twitter.

👋 Howdy to the 2,318 new legends who joined since our last edition! You are now part of a 513,456-strong tribe outperforming the competition together.

LATEST POSTS 📚

If you’re new, not yet a subscriber, or just plain missed it, here are some of our recent editions.

🛡️ The Agentic CRM For AI‐Native Companies. An interview with Keith Peiris, Co-Founder & CEO at Lightfield.
📡 The Last Industry Software Hasn't Eaten. An interview with Hermann Frank, Co-Founder & CEO at Gigs.
🐝 Unfiltered: beehiiv’s Build Or Die. Inside the platform that ships faster than anyone its size.

PARTNERS 💫

Framer is the pro website builder for creators, teams, and businesses that care enough to get every detail right. Now with Agents that work alongside you across the full website workflow—all without leaving the tool where the real site lives.

Agents bring speed and scale; you bring taste, judgment, and control. Build and iterate faster, and ship updates in minutes. Scale content and keep metadata consistent across pages.

Your attack surface is bigger than you think. Arctic Wolf analyzed more than 800,000 IT assets and found that one in three was missing a critical security control, while 19% had reached end of life.

Explore the hidden visibility, protection, and patching gaps expanding cyber risk, plus practical steps security leaders can take to close them.

_{Interested in sponsoring these emails? See our partnership options here.}

HOUSEKEEPING 📨

If you ever over‑prepare for a big meeting and still find yourself rambling or freezing, you’ll like this. I’ve been using Ultraspeaking to improve my speaking (found it through Lenny), and they’re running a free live class on July 27 where they'll live‑coach three volunteers on how to: explain complex ideas clearly (even on the fly), recover when you get off track, and prepare without sounding stiff or scripted.

You can grab a free spot inside their app here. I love Ultraspeaking. I find it incredibly valuable, and always come away with a number of interesting insights. Anyway, enjoy today’s interview!

INTERVIEW 🎙️

Doug Tallmadge, Co-Founder & CEO at Gradial

Doug Tallmadge is the Co-Founder & CEO of Gradial, a Seattle-based startup building AI agents for enterprise marketing operations. He co-founded the company alongside three fellow Dartmouth graduates: Anish Chadalavada, Deip Kumar, and Anup Chamrajnagar. Before Gradial, Doug worked at SpaceX as a software engineering manager. The company raised a $5.4M seed round in February 2024, a $13M Series A in 2025, a $35M Series B in December 2025, and a $65m Series C in June 2026—led by VMG Partners, with Madrona and Pruven Capital participating—bringing total funding over $120M at a $675M valuation. Customers include AWS, T-Mobile, and Prudential Financial.

Doug.

Doug went from writing software for rockets to building agents that update websites and route campaign approvals, and his argument is that the two are not as far apart as they look. The real story at Gradial is less about the AI and more about what happens when a company has to compress a normal fundraising timeline by two-thirds, grow headcount fourfold in a year, and simultaneously dog-food its own product on the problems that growth creates.

What is Gradial doing for enterprise marketing teams?

Gradial is an AI agent that focuses on solving the operational problems of running a large enterprise marketing team. It's an agent that actually goes and does the work within the workflow: updating the website, sending emails, reviewing copy docs, doing QA analysis on different things you're doing. It’s an operational agent that jumps in and works on all the bottlenecks within enterprise workflows.

Source: Gradial.

You raised your Series B eight months after you raised your Series A. Why so fast?

The AI space has been moving so fast that the growth we've seen (and that many companies in the space have seen too) has really kind of forced us to take the typical company-building playbook and run it even faster than you normally would. We've been lucky to expand to a number of leading enterprises and really grow the team out here in Seattle and across the country to support that customer demand.

What was the hardest part of going from a SpaceX engineer to a founder?

SpaceX likes to describe itself as a startup, and there are a lot of aspects that still are very startup-like about how it operates, even though it's no longer a startup and it has 15,000 or 20,000 employees at this point. I was lucky to be there in 2017 when it was a little smaller than it is today. The similarities are that you've got this need to really move fast and try to be scrappy. At SpaceX, we would think about every dollar we were spending. We were very frugal in how we operated, even though it was already on a multi-billion-dollar scale.

Where it's different is that when you're operating within a company like SpaceX, you take for granted the fact that you can just go work with a recruiter to build out your team, that you have all this infrastructure, like someone literally keeps the lights on in your office. As soon as you start a company, you realize how much additional work there is behind the scenes, doing all those different tasks that you often ignore as an engineer.

The Gradial team.

I think there's actually more similarity between enterprise marketing and building a satellite than people would expect. There's a schedule, there's a rollout, there are people contributing from different places, and there are hard problems to solve along the way. But the thing you really see in a startup is just how many of those smaller tasks that aren't very difficult add up to something that operationally is quite challenging to scale.

What does a typical week look like for you?

Meeting with customers—usually on the road somewhere—and spending as much time as possible thinking about what marketing is going to look like in 2027. Diving into ideas with the team about what's become possible as the models have gotten better.

Source: Gradial.

We spend a lot of time testing and experimenting with the latest generation models to figure out what use cases are becoming possible or on the edge of possible. We've really done that since day one. You'd get these new model releases, and there'd be certain use cases that went from maybe 20-30% to 70-80% accurate, and you start to develop an intuition for what's on the edge of possible. So a big part of every week at Gradial is around how we are going to unlock that as the models continue to improve, and figuring out what tricks we can come up with to unlock use cases and capabilities that aren't possible yet with the current generation.

How has your job as CEO changed since the seed round?

In a startup, your job is going to be different every six months. That's true for the CEO and for everyone at the company. A lot of times, that change is really adding more responsibilities, not subtracting them. Learning to multitask across a larger number of different parts of the business is something all of us are doing, but I definitely feel it very acutely. At the seed, a lot of my time was spent building the product and just writing code. I still do a fair amount of that today, but I've learned there's a need to start spending less time on that and more time thinking about the direction of what code we should even write. The hardest problem is spending enough time thinking about what to delete. It's easy to come up with a scope and ideas. It's harder to figure out what ideas to ignore.

This is the thing I've been most focused on personally: figuring out how I can spend more time deleting stuff across the company that isn't needed. Simplifying the product, the vision, and the roadmap.

Especially when you have AI that can go and build all of your ideas, if you're not careful about eliminating the worst ones first, you could easily crowd out your garden of ideas. It's like throwing weeds in your garden. Those weeds develop, and they take away a lot of space from the things that really matter.

Simplifying in public.

A year or two ago, that was not a problem. At the seed stage, you have the clarity of really one thing you're working on, and you just have to deliver that one thing. It's actually very simple: you just show up and work on it all day. Yet, as you start to take on more customers, more responsibilities, you have to be really purposeful about not spreading things too thin.

Are you using your own agents inside Gradial's operations?

We've been using our own agents since day one. And as you start to add more people, you start to see a lot of the pain points that Gradial is designed to address, which is the operational handoff across people on the team. Something that resonated for me as an engineer coming into the marketing space was that whether it's a marketing project like creating a new website, or an engineering project, a lot of the work actually becomes coordination as the team gets larger. It's been really satisfying to start using Gradial in a new capacity as we've grown, because you get into exactly the problems the workflow has to solve. Like, who's able to approve that update to the website? When we had fewer people, there was just one person handling the website updates; they could approve them themselves. When you get to a certain scale, suddenly you need a creative asset from one person and approval from another person to make the update. And that's exactly what we've built our agent to solve.

Agent harmony. Source: Gradial.

Then, in other parts of the business, whether it's customer experience analytics and optimization on the site, we have agents that can basically optimize your site, the journey, and try to understand where a customer might get stuck across your different digital experiences, whether it's email or web. As our surface area of digital experience has grown, we've been able to really leverage that. We have these always-on brand monitors where the agent is going and looking at your website, looking at all these digital channels, and basically just trying to catch things where there's an opportunity to improve the product or to improve the experience. That's another place where the models have really unlocked capability in the last six months.

The latest versions are becoming smart enough to use computers at a high level, which has opened up some really cool use cases that were not really possible even 12 months ago.

Describe the difference between an AI tool and an AI agent in practice?

A tool is something that takes an input and delivers a specific output. It does one step in a chain of steps. An example would be: go and spell-check this document, and you get back a spell-checked document. An agent is something that takes many different steps across at least one system, and the output is undefined. It can take a range of inputs and deliver a range of outputs. It's not one-to-one, and it's not a single step.

For example, customers use Gradial agents to enter a workflow and execute a campaign in a marketing setting. A campaign can have many different steps. You could have a step where you need to get legal approval from someone in another time zone to review a document you're creating with the agent. One thing a customer can do is take a screenshot or photo of a whiteboard with their workflow and give it to the agent. Then, the agent will execute the workflow according to a diagram they drew on a whiteboard. That's an agentic capability because it wasn't programmed in. It's not hard-coded. I can literally go to the whiteboard, draw the workflow I want, take a photo on my phone, give it to the agent, and it goes and executes.

Working with the best.

In a tool, thinking about the prior generation of SaaS, things like RPA, you would have had to spend many, many hours of configuration just to implement a single workflow. Typically you'd have to bring in an expert to do it because the product is too hard to configure. For example, if you've ever tried to set up even a simple CRM, you end up on the phone trying to learn how the documentation works because you actually can't configure the workflow. In a tool, all the intelligence is in the hands of the person using it. It's like a screwdriver—I have to turn the screwdriver. In an agent, the intelligence is partly in the agent itself. You're not providing all the force. It's like a power tool where the energy comes from somewhere else.

Source: Gradial.

How do you get the best out of yourself?

Source.

I wake up every day excited because I'm working on something I love. That's why I know a startup was the right place to be. I get that energy just from doing something I really enjoy: working on hard problems where there's no clear path. No one's ever built agents at this scale before. No one's ever done what we're doing.

That's been just really gratifying, to wake up every day and say there are many hard problems to solve and also so much value to deliver in a space that is pretty greenfield. There's a lot of opportunity to make the marketing workflow way more fun and way more effective for so many different teams. Having a list of ideas that you could never possibly do all of them. It's like having a travel list you could never fully complete, but you're checking something off every day. To me, that's extremely gratifying.

Anything else you'd like to share?

I think we're going to see the next six months go even faster than the last. That hasn't always been true; there's been this narrative that AI has just been continuously accelerating, and that's partly true, but there have been periods where the relative impact was not accelerating. From roughly January 2024 through December 2024, the impact in the enterprise was not accelerating. It was pretty static. People were still stuck in POC hell across most applications. We were able to break through most of that POC landscape, but there was a period where people were in try mode and not in use-case mode.

That has 100% changed. Everyone I'm spending time talking to in the marketing landscape is accelerating their AI adoption and fundamentally rethinking their entire stack. Everything is on the table right now. The way I see it, and the way I'm hearing it from customers, is that all of the infrastructure built up over 20 years to run these different marketing technology practices has to be reevaluated in the time period of AI.‘ What is actually going to scale with us as we prepare to re-architect our marketing technology for AI agents?’

Gradial co-founders from left: Doug, Anish, Deip, and Anup.

A big part of that is that more of the traffic is coming from AI agents, and that's going to continue to accelerate. So a lot of the tools, techniques, and priorities that made sense three years ago make way less sense now. From a Gradial perspective, that is very much where we're focused: 2027 marketing and what the fundamental data primitives needed to support the marketing agents are.

Extra reading

• Gradial Raises $65M to Build the First System of Work for Enterprise Marketing - June, 2026

• Announcing GEO with execution built in - March, 2026

• From Hours to Minutes: How Agentic AI Gives Marketing Teams Their Time Back - April, 2026

• Agentic Marketing Is Rising. Gradial Is Leading the Way - July, 2025

And that’s it! You can follow Doug on LinkedIn or check out Gradial on their website to keep up with what they’re building!

BRAIN FOOD 🧠 

TWEETS OF THE WEEK 🐣 

Source.

Source.

Source.

TOOLS WE RECOMMEND 🛠️

Every week, we highlight tools we like and those we actually use inside our business and give them an honest review. Today, we are highlighting Lightfield*—an AI-native CRM that assembles itself from your email, calendar, and meetings.

beehiiv: We use beehiiv to send all of our newsletters. Apollo: We use Apollo to automate a large part of our 1.2M weekly outbound emails. Taplio: We use Taplio to grow and manage my online presence.

See the full set of tools we use inside Athyna & Open Source CEO here.

HOW I CAN HELP 🥳

P.S. Want to work together?

1. Hiring global talent: If you’re hiring tech, business or ops talent and want to do it 80% less, check out my startup, Athyna. 🌏

2. See my tech stack: Find our suite of tools & resources for both this newsletter and Athyna here. 🧰 

3. Reach an audience of tech leaders: Advertise with us if you want to get in front of founders, investors and leaders in tech. 👀 

That’s it from me. See you next week, Doc 🫡  P.P.S. Let’s connect on LinkedIn and Twitter.

👋 Howdy to the 1,995 new legends who joined since our last edition! You are now part of a 511,138-strong tribe outperforming the competition together.

LATEST POSTS 📚

If you’re new, not yet a subscriber, or just plain missed it, here are some of our recent editions.

📡 The Last Industry Software Hasn't Eaten. An interview with Hermann Frank, Co-Founder & CEO at Gigs.
🐝 Unfiltered: beehiiv’s Build Or Die. Inside the platform that ships faster than anyone its size.
🧠 Building A Second Brain With Claude Cowork. The ultimate guide to one of the most elusive tools.

PARTNERS 💫

Founding a company is easy, scaling it is hard. Operational systems break, people problems increase, and your calendar keeps getting fuller. Mochary Method is an executive coaching firm founded by Matt Mochary—coach of Sam Altman and Brian Armstrong, and writer of The Great CEO Within—that specializes in helping growing startups.

The Mochary coaching team, who coaches companies like AngelList, Lovable, and Midjourney, is offering a complimentary coaching session to founders and executives for a limited time.

This newsletter runs on beehiiv. So when we say the referral program works, it's not a guess; it's our actual growth engine. Readers share, hit milestones, earn rewards, and the list grows on autopilot. Pair that with their recommendation network and newsletters on beehiiv grow 2.75x faster. We're one of the data points. Worth a look if you're building something.

Get 30% off your first three months on beehiiv with the code: BILL30

_{Interested in sponsoring these emails? See our partnership options here.}

HOUSEKEEPING 📨

I’ve been set a challenge this weekend. My gym is having a trivia night + talent show, and I am opening the show with one of my oldest party tricks, a rendition of Ice, Ice, Baby by Vanilla Ice. The image on the left is the outfit that just arrived from Amazon; on the right is the man himself (and yes, I plan to shave the speed stripes).

The beautiful thing about it is I’m not the slightest bit nervous. If there is one thing that I was born to do, it was recite Ice, Ice, Baby. The original idea was for me to do it, without the karaoke words, at 2x speed. I’ve since tapered that idea down, but I could probably pull it off. I have backup dancers and one breakdancer lined up for the show. Wish me luck, and enjoy today’s piece!

INTERVIEW 🎙️

Keith Peiris, Co-Founder & CEO at Lightfield

Keith Peiris is the Co-Founder & CEO of Lightfield, an AI-native CRM that assembles itself from a company's email, calendar, and meetings instead of waiting for sales reps to fill in fields. Before Lightfield, Keith and his co-founder Henri Liriani met at Meta, where they worked on products used by billions, and in 2020 left to start Tome, an AI presentation tool incubated at Greylock with Reid Hoffman on the board. Tome became the fastest productivity product ever to reach a million users, grew to around 25 million, made the Forbes AI 50 in 2023 and 2024, and raised roughly $81M from Lightspeed, Coatue, Greylock, and others at a $300M valuation.

Keith.

Then they killed it. After a year rebuilding in stealth, Keith and Liriani shut down a product with tens of millions of users to start over. The bet underneath Lightfield is bigger than software. That CRM, an $80B category with tens of millions of users and almost nobody who likes the product, is ready to be rebuilt for a world where LLMs, not humans, do most of the knowledge work. Dan Rose at Coatue, who led the original investment in Tome, backed the pivot. For founders, the pull here is less the product than the decision behind it. What it actually takes to walk away from a company that was working, sit in the negative momentum, and find conviction again from zero.

What was broken about CRM, and why attack a market most people think is solved?

There's nothing inherently broken about CRMs for the old jobs they were supposed to do. In the old world, a CRM was just a place where you wrote down things about your customers so you could serve them and take care of them better, and Salesforce and HubSpot are basically fine at that. But in the past couple of years, the world changed, in two very material ways. The first is that most knowledge work now will be done by LLMs instead of humans. And it turns out most knowledge work is people running businesses, trying to take care of their customers, sell to them, and serve them. There's a need in the world for context—a need to understand people, customers, and businesses better, so that LLMs can serve them.

The other thing, and I'll go all the way back for a moment, is that in pre-industrialization the people who made products were the same people who sold them. If I owned a shoe store, I'd make the shoes and then try to sell them, and if I couldn't sell them, I'd make something else. You had this tight feedback loop between maker and seller. Post-industrialization, in the hopes of getting to scale, we split it up into who markets, who sells, who makes things, R&D versus commercial.

Source.

That's been true for pretty much the entirety of software building. You have these big teams, engineering and sales, using different systems. One uses Linear or Jira, the other uses Salesforce, and they don't really talk to each other except when a manager shares a deck back and forth.

Something changed in the past few years that shook this. All of these new software companies are blurring the lines between those orgs, because the software is all kind of the same now: agents and harnesses and APIs. The value creation is actually in the last mile of integration, of tuning, of being there with the customer to make sure you're solving their problem. That has created an org in the middle. Some people call them ‘forward-deployed engineers’. It's actually the whole company working with customers. We’re changing the way we sell the product. Suddenly everyone needs to be tapped into the same customer truth. That's the opportunity. Salesforce certainly wasn't designed for forward-deployed engineers, scientists, marketers, computer scientists, and so on. We saw an opportunity to reimagine what CRM should be, both for LLMs to do work and for companies to be fully customer-deployed.

How is Lightfield different from the typical CRM on the market?

The first big difference is the latest-generation data model—CRM 3.0. It was all about rigid structures and fields. You had accounts, opportunities, and contacts, and those objects had fields on them, such as the location of the customer or the last thing they said. It was designed for humans to manually fill out those fields. We're designed to absorb everything about a customer. And when I say everything, I mean what they said, what they're doing in your product, the commercial value, what you're selling to them and how. We ingest it all by default, and the reps don't have to update them. You're building a corpus of data about each customer. That's the first foundational difference.

Source: Lightfield.

Source: Lightfield.

The second difference is that we give you nice tools to use all of that customer information, so you can do anything with it. You can use it to make a proposal, to write your email follow-ups, to find more customers like this one and reach out to them. We're giving you your customer truth and helping you play with it so you can actually do things with it.

What was the hardest part of starting Lightfield?

With Tome, we had a lot of momentum, but I learned that it’s very hard to overcome having the wrong idea. We had all the growth and all the capital and this brilliant team, but we just couldn't make good presentations for people doing serious work. That's what made us realize we actually needed to move to a different opportunity. Lightfield was the product the Tome team needed in order to be able to make great presentations, so it felt like a natural chapter two.

There were a bunch of things that were hard about pivoting a company that was already working. The first is that we went from 70 to six people, and those people who are left are all wondering if this thing is going to work out. The only way through that is to put blinders on, work as hard as you can, try to get some customers, and stay focused on the future, not the past. I actually talked to the CEO of Slack about this, one of the more famous pivot stories in history. I don't even think he remembers this, but one day I spoke to him on the phone and he said, “You need to have a team that's small enough to fit around a conference room table. Every day you need to get customers and talk about those customers, and just create momentum until people stop talking about the old thing.” That was the overcoming-the-negative-momentum part.

Then, for Lightfield in particular, there are some strange things. CRM is very big. It does a lot of things. I'd argue we're going through it really fast with AI coding and the way we built the system, but it's very hard to get someone to use your three-month-old CRM. As a founder, you build a better product and company with customers, so you almost have to come up with a trick to get people to use your early product so you can get to the next stage. We had a few tricks. When we were three months in, the first thing we used was our giant office space. We told people that if they were willing to use our CRM and give us training data, they could work out of our office. That's how we got our first five customers. We knew we were onto something when they stopped complaining about using us.

The next hardest part was getting our first hundred. I basically cold-emailed all of these Y Combinator founders, hosted events, and went zero to one there. And then, maybe the hardest thing about CRM is that it's very easy to build a simple CRM that looks like it's going to work. But it's a very painful engineering challenge to make a CRM work for a million records, with all of this unstructured data, with clean APIs. We needed to hire good folks and afford ourselves the time to build all of that.

What's your North Star metric, and how do you improve it?

I think there are two types of metrics. There's breadth, and there's depth. I'm old enough in my founder journey to know that as long as you have depth, you can get more breadth. To put it another way, if our customers use our product all the time, they love the work it produces, they reference it, they're really happy, then I know I can always pay to get more customers. We can always hire more salespeople, and we can always do more marketing.

Source: Lightfield.

Source: Lightfield.

So the first thing I look at every morning is how much time our customers are spending on the platform, how many agent actions they’re running, how often they’re accepting the work we create, and how far we’re spreading within the company. The best version of Lightfield is when we're wall to wall—the engineers are using it, support is using it, sales is using it. So I'm always watching the spread of Lightfield inside a company to see whether we're working toward that vision, or whether we're just a thing a salesperson uses a couple of times a day. Assuming all of that looks good, then I'm obviously looking at how we get bigger companies to be very happy on the platform.

How would an engineer use Lightfield?

There are a couple of ways. The most obvious one is the handoff from sales to engineering. We just got this big customer, so we're going to have to do a bunch of work for them. It's the whole ‘let me ramp up on this customer’. Understand what they need, make sure you show up to their next meeting prepared, and influence the roadmap.

The second one is that most of the companies we serve don't actually have product managers. It's the engineers who run the products and build. To do that, you need a sense of the priority of the roadmap, and you want some scientific measurement of how many people have actually asked for this feature. I know the sales guy was very loud about us having to build feature X, but how many people have actually asked for it, what's the commercial value, and what's the why behind the ask is really helpful for the roadmap prioritization exercise.

Related to that, when you build a feature, it's hard to know whether it's actually going to solve the problem for the customer, or whether you just built an unserviceable version of it.

A lot of engineers like using Lightfield to get to the root of the problem a customer is asking about. It ends up being the caricature of the customer in every product meeting.

The customer, allegedly.

What does your week look like as CEO?

It's funny, I have a very reductive view of what a CEO should do, which is that it's the CEO's job to work on the rate-limiting step of the company until it's no longer the rate-limiting step, and then move to the new one. Every Sunday night I'll write down the three biggest problems of the company. Sometimes it's engineering hiring, sometimes it's ramping up sales, sometimes it's positioning in the market, sometimes it's pricing. Then I'll send it to my core team and let them know that I'm going to pick off these things this week and go really deep, and ask if they need help with anything else. Then I'll basically be an individual contributor on that thing until it's good.

Lightfield of dreams.

Right now my biggest focus is on sales and customer success. The product and engineering team has a ton of velocity, and we actually have way more demo bookings and pipeline than AEs. So my job is to help us find a sales leader, and take what’s worked well for us and help it scale.

Even though we’ve started to build out a commercial team, I still like to take a set of customers and work with them through the entire customer journey, from the first call to post-sales implementation. It helps me stay close to what’s actually happening in the field and gives me firsthand experience that supplements the context I get from Lightfield on what’s happening in our market. I get a lot of satisfaction watching customers go from their first ‘aha’ moment with Lightfield during the sales process, to transforming how their company operates after we stand them up on our platform.

So that’s where 80% of my time goes. I'll spend 20% of my time making sure the other trains are running. I've found that the process meetings, where I go in and watch people do standup, don't really work for me. I prefer to do stochastic sampling, where I'll just Slack a couple of people on each team and ask how things are going and whether they think something should change, and then react off of that.

How do you approach goal setting and product cadences?

We plan about six months out in terms of revenue. We're trying to 10x revenue in the second half of this year. To work backward from that, some things need to be true. The obvious one is that now we need to be able to support a Series C company being very happy in Lightfield, and we need some amount of commercial capacity. So in my head, I know we need to ramp up to 10 reps and have a certain level of product functionality by the end of six months. That part is somewhat rigid. But on a day-to-day and week-to-week basis, we operate with continuous planning. We have one engineering and product roadmap and then a separate go-to-market roadmap, all in Linear. Every day we'll look at it and ask ourselves if it needs to change, and every week we'll have a conscious conversation.

The roadmap, mid-week.

We have a Friday afternoon meeting called ‘roadmapping’, where we check if we need to make any radical changes for next week. We've found that works really well. Be somewhat rigid on a six-month-to-one-year time horizon of where you want to be, but be continuously shifting and adapting day to day, week to week.

What do you look for when hiring? Do you have any unique go-to’s?

As a CEO, you constantly need to hire people into roles where you have low task maturity. I would hope I'm worse than the marketer or the infrastructure leader I'm about to hire. There are some things I do that are pretty conventional. I like hiring people who have intentionality in their career, because if you've made intentional career steps, that suggests to me you'll be intentional in your planning and your autonomy with us. So usually I'll ask someone to tell me their life story, and look for intentionality and good reason from one step to the next.

To the moon.

The other thing I've found is that right now I'm not in a place where I can create lots of structure for my team, because I'm jumping from one problem to the other each week. So I'm looking for folks who can bring structure and order to chaos. The best way to test for that is having a working session in the interview. I do a 90-minute version of reviewing the current state of this domain, the problems we face, the things I don't like, and the threats and opportunities. Then I work with the candidate on two things. First, can we steer this toward truth? The worst thing at a startup is when you have people managing process instead of outcomes, so I want to see us get to ground truth and actually problem-solve. Second, can we steer it toward removing entropy and land on a reasonable plan by the end? That deep situational problem-solving interview has been the best one for me. I don't try to do any posturing. I actually try to be the worst version of me, chaotic and all over the place, and see if this person can turn it into something good.

At our stage, we're a little under 40 people now, and we don't hire anyone who acts like an exec. We try to hire people who are still in the work. My favorite people are the ones who have been really good at their function and are now looking for a chance to run the whole function. Great at craft versus professional manager. For the most part, I hire people for whom Lightfield could be the pinnacle of their career, where they're going to be irrationally bought into making the company work, because making the company work could mean everything to them, as opposed to just being another notch on their belt.

How are you thinking about AI operationally across the company?

The forcing function for AI inside Lightfield is that we're not trying to build a 10,000-person company. We have the aspiration of having more revenue and more growth than Salesforce, but we don't want to be Salesforce-sized. I want this company to be 500 people, not 10k. If you impose those headcount limitations on the team but you enable technology, you start to see whether the incentives can create the outcomes you're looking for. We're almost 40 people, and we still don't have a recruiter. That's one of the sources of pressure. If it's really easy to hire people, you'll just hire people to solve your problems. If it's hard to hire people, you end up coming up with creative solutions.

Inside product and engineering, it's fairly clear. We've moved everything to a mono repository. We put product specs into it so Claude can understand them, we deploy agents everywhere, and our team burns lots of tokens. Outside of product and engineering, it's been more challenging, and really interesting. It turns out we can get pretty far with AI. For example, maybe I don’t really need a head of finance if my chief of staff with a finance background can get really far using spreadsheets and context.

Source.

In addition, we record all of our internal meetings and dump them into Lightfield, so it has a knowledge base of everything we've ever discussed and everything we've ever decided. It turns out that if Lightfield has all of that context, it can do some good work for you. We fed a lot of that context into Paraform, and we're getting good candidates from it. Because of all that, we have about five people outside of engineering and product, and we've been able to get really far without any more.

How do you use AI personally?

I think this is true for a lot of CEOs. The thing that hurts me most is asking an employee for something and being told it'll take weeks when you know it could take hours. I'm realizing I can have a hero complex because of AI, in that we'll all ship features to the codebase instead of begging engineers to do them, and that's been a big unlock. I'm not super technical anymore, but we still push things to prod using Cursor.

The other thing I've found is a strange one. One of the hardest parts of my role is making sure I'm operating from reality, because everyone tells you a different version of what you want to hear. Employees sometimes overrepresent how good things are, so they look good. Investors have a different view that isn't entirely aligned with you. I've found it really useful to have a mirror to double-check if stuff is sensible or believable. Planning different scenarios has been a big unlock for me. Being able to accurately forecast where your business goes is one of the hardest problems, because it's not just the numbers. It's about having the right customers, building the right thing, considering the risks of competition, and more. I'm essentially trying to build a world model of our company so I can forecast better and act differently.

How do you get the best out of yourself?

I find there's just too much noise in tech right now. I used to read everything on X and listen to every founder podcast, and now I never log into that stuff, and I don't listen to podcasts. I just read books, and I find I have a clearer, simpler mind as a result. You can either be in a mode of consuming content or creating content, and I find it's hard to be creating when you're consuming all of this noise. So right now I'm in a state of creation, where I've shut that stuff off and I just keep going.

Related to that, I try to work out every other day. I have a rule with my wife where we go to sleep at the same time every night, no matter what we're doing, to force the eight hours. When I was fundraising, I was taking a magnesium supplement, because it turns out that when you're stressed out, that stuff helps. And because I'm in California, I'll go on long bike rides to try to gain some perspective. Another strange thing is that I used to play competitive chess in high school, and I find that when my mind is running, I just get my butt kicked on chess.com. It's a good barometer. If I open it up and I just get stepped on, that's a good sign I need to focus on peace.

Extra reading

• Building the CRM that works for you - May, 2026

• The founder's guide to evaluating an AI CRM - January, 2026

• Why we built Lightfield - November, 2025

And that’s it! You can follow Keith on LinkedIn or check out Lightfield on their website to keep up with what they’re building!

BRAIN FOOD 🧠 

TWEETS OF THE WEEK 🐣 

Source.

Source.

Source.

TOOLS WE RECOMMEND 🛠️

Every week, we highlight tools we like and those we actually use inside our business and give them an honest review. Today, we are highlighting Granola*—an AI-powered notepad that takes, summarizes, and organizes meeting notes without using intrusive recording bots.

See the full set of tools we use inside Athyna & Open Source CEO here.

HOW I CAN HELP 🥳

P.S. Want to work together?

1. Hiring global talent: If you’re hiring tech, business or ops talent and want to do it 80% less, check out my startup, Athyna. 🌏

2. See my tech stack: Find our suite of tools & resources for both this newsletter and Athyna here. 🧰 

3. Reach an audience of tech leaders: Advertise with us if you want to get in front of founders, investors and leaders in tech. 👀 

That’s it from me. See you next week, Doc 🫡  P.P.S. Let’s connect on LinkedIn and Twitter.

👋 Howdy to the 5,741 new legends who joined since our last edition! You are now part of a 509,143-strong tribe outperforming the competition together.

LATEST POSTS 📚

If you’re new, not yet a subscriber, or just plain missed it, here are some of our recent editions.

🐝 Unfiltered: beehiiv’s Build Or Die. Inside the platform that ships faster than anyone its size.
🧠 Building A Second Brain With Claude Cowork. The ultimate guide to one of the most elusive tools.
✍🏻 Building The Incorruptible Companies Of The Future. An interview with Eric Ries, Author of Incorruptible.

PARTNERS 💫

Smart hiring for teams that need strong talent now.

• Fill key roles faster, without waiting weeks or months for the pipeline to behave.

• AI-assisted matching and structured vetting help you find people who fit the role and the team.

• Cut time and cost with a process built for scale, not endless back-and-forth.

Meetings aren’t the problem. The challenge is everything that comes after them. With Granola, the AI Notepad for people with back-to-back meetings, it's no longer about scrambling to keep up.

Take notes the way you always have. Granola works in the background, turning conversations into clear summaries, action items, and next steps.

Before, during, or after a meeting, you can chat with your notes to quickly understand what needs to happen, write follow-ups, or share context with others.

_{Interested in sponsoring these emails? See our partnership options here.}

HOUSEKEEPING 📨

I have a recommendation for you. One of my go-to moves whenever I need a quick break or am a bit stressed is to watch these animal rescue videos from The Dodo. They are as nearly as heartbreaking as they are uplifting, but they always end in a happy ending. Highly recommended.

It actually makes me think about my life at times, and what I do with my days. I love writing this newsletter; it gives me great joy, and I love my startup, Athyna, as well, but something tells me throwing it all away to open an animal shelter might be the most rewarding thing I could do. Maybe not the most impactful, but I do think the most rewarding. Anyway, that’s all for now. Enjoy today’s great interview!

INTERVIEW 🎙️

Hermann Frank, Co-Founder & CEO at Gigs

Hermann Frank is the Co-Founder and CEO of Gigs, a San Francisco-based company building what it calls the world's largest telecom carrier delivered as software. Before Gigs, Hermann co-founded Hypermesh with Dennis Bauer, a venture the pair treated as a deliberate hunt for the one industry most starved of modern infrastructure. Telecom won. Gigs came out of the 2021 Y Combinator batch and have since raised $97M total, including a $73M Series B in December 2024 led by Ribbit Capital, with backing from Google's Gradient Ventures, among others. The platform now powers mobile connectivity for Revolut, Klarna, Nubank, and a growing list of Fortune 500 enterprise customers.

Hermann.

What makes Gigs unusual is the scale of the problem and the company Hermann keeps while building it. His cap table includes Dara Khosrowshahi, Tony Xu, and Fidji Simo, public-company CEOs who don't typically write angel checks for fun. Gigs has a user base of nearly 950 million people across its platform partners, provisions connectivity to more than 100,000 employees of one of the Magnificent Seven across 40 markets, and didn't hire a single person from telecom until it crossed 130 employees. The thesis: the last industry software hasn't eaten is also the most defensible thing left to build.

What does Gigs do, and why does it matter?

Gigs is building the world's largest telecom carrier, delivered as software. We're the first company to rebuild the entire telecom stack from the ground up—from the radio access—and build it as one end-to-end operating system, so it's programmable, 24/7, global by nature, and composable to technology platforms on the consumer and enterprise side. The reason this needs to exist is that telecom carriers are the last major industry in the world that has not yet been eaten by software. They're still very local. They're usually national entities that have been privatized or partially privatized, dating all the way back to when the railway was built, so you have a lot of legacy, government legacy, the politics, all the slowness of this incumbency. As a result, they haven't been able to modernize over the last 30 years.

The world has moved forward a great deal. It's mostly driven by software now. Enterprises operate on software, consumers operate on software and apps, and telecom has become the largest rent-payer in the internet economy, because they don't have the rails to natively embed and connect themselves. You have this really large, messy industry without a common interface. It's not global, and it incurs insane customer acquisition costs.

Source: Gigs.

AT&T, for example, in one of its recent quarters, paid $900M to acquire 300,000 subscribers. That's $3,000 per subscriber to sell you an unlimited plan. It takes many years to amortize a customer, so there's no more money left. On the other hand, given that they run on a stack of legacy vendors that are mostly outsourced, operational efficiency is not improving. They have to pay a ton of people to do tasks that other industries have automated long ago.

These are the two problems we're tackling. First, we're eliminating the customer acquisition cost by embedding telecom carriers directly into the world's largest consumer and enterprise apps. Second, we're automating the entire telco operation by making everything programmable and agentic—customer support, back office, and so on. The result is a 75% lower cost base for operating a carrier, because that's how much carriers are currently spending on customer acquisition, operations, customer support, and back-office. Only 25 to 30% in some periods goes to actually building and maintaining the network. That's someone who has to climb poles to install, maintain, and operate the antenna. The rest is just inefficiency at this point, and that's what we're solving from the ground up.

Who is your ideal customer, and why are they building on top of Gigs?

Our ICP is tech in a broader sense. These are the largest consumer and enterprise platforms. Revolut, Klarna, and Nubank are just some of our consumer customers, but we also power enterprise connectivity for global companies. For example, we're provisioning connectivity to over 100,000 of the Mag 7 employees through Gigs, and we are abstracting carrier relationships and operations in over 40 markets to make this happen. Now you have one API call that can provision a phone plan for an employee instantaneously and deliver it to their email, Slack, or directly to their device. While previously this was a process of two-plus weeks, over 40 steps, most of them requiring humans, now it's a real-time service that you can call and provision globally.

Source: Gigs.

Currently, we have aggregated a total addressable user base of about 950M people across platforms. It doesn't mean we have 950M lines or that we're visible to that many people, but we are embedded in platforms with this user base, and we're progressing well in selling the service and making mobile a true product. This is the largest distribution surface that anyone in the world has ever built for telecom.

How did you come up with the idea for Gigs?

It started with Hypermesh, a company Dennis and I founded before Gigs. Hypermesh was created to find the biggest problem to solve in telecom. We'd evaluated different industries, and it came out that telecom was the one. There's so much entropy: nothing has happened for 20-30 years, VC hated telecom for a long time, there are no great software teams working on anything, and we'd been around long enough to have seen multiple industries go through the cycle.

Hermann and Dennis.

We built a product that worked really well, and by the end we needed a service that would let us programmatically access an LTE network. It was kind of a product expansion we were experimenting with. Basically, we wanted to buy an eSIM through an API. For us, it was the most obvious thing that this would exist, because you have Stripe, AWS, you can get hosting, you can get payments in an instant, and we found that it doesn't exist. You cannot just programmatically provision yourself onto a network. That was mind-blowing to us, especially in the age of eSIM.

We looked into the industry, tried to understand why, and broke it down to ‘what if we can solve this, and you can just provision yourself or your devices onto a network, not just for IoT but for the actual real phone plan’. Then, basically, what would come next is what follows in any other comparable industry: you create a completely new interface for it. We had the pain. It was mind-boggling to us that it didn't exist. We understood why, and then we just set out to create the tool we hoped would exist.

What was the hardest part of going from zero to one?

When we started talking to people about what we wanted to do, people were really hesitant, believing that no carrier in the world would let you do this. So we needed to find a way to actually integrate ourselves into a carrier, and no carrier would talk to us initially, unsurprisingly.

We found a back door by integrating into someone who already had the right to resell on one of the networks that we're now close partners with. That allowed us to build the first version of the product and prove our point.

Source: Gigs.

This party was an aggregator that had been relatively unsuccessful on its own, so when the network started to see where this growth was coming from (those guys haven't grown in a long time), it kind of piqued their interest. We also got the right people on board to advise us on how to interact with that carrier, and that had pull in there. By proving our point, building a product through the back door, and securing the right internal allies, we secured our first carrier deal. That made it easier to get other deals. The other thing is, no one has created a consistent API for telecom that you can use across multiple countries. That was the next thing that people said was impossible. Some people in the competitive field are still running around saying it doesn't work, but obviously, there's more than enough proof now. So the other hard lift up front was abstracting differently regulated, functioning carriers into a single cohesive layer, then repeating that over and over and over again.

Source.

Then we started with smartwatches because we felt they were relatively forgiving. If you're down on your iPhone for even a second, you're going to hate us, but these smartwatches were not like an Apple Watch. They were like kids' smartwatches, elderly smartwatches. Basically, it's there so someone can open an app and see where the person is. It was an amazing niche that we stumbled upon through one of our advisors. That allowed us to build the initial platform on which we could raise our seed and Series A. Then it was another two years of buildout after that, so it was a total of over three years until we could serve actual smartphone use cases, and people would not shut us down for that.

How did Y Combinator shape your early thinking and go-to-market?

YC is these great rails: they put you on rails, make you fast, help you develop a cadence as a company, and make you hard on yourself and iterate. They also take away a lot of anxieties that founders walk around with, like, you know, ‘it's fine to pivot’. Luckily, we didn't need to pivot during YC, but they bring in the most successful YC founders over the past decade and a half to talk about what they're best at. So you have the Stripe founders telling you how to build a hard company and scale it, and a lot of cultural stuff. You have the Brex founders who came in and talked about pivoting mid-batch, because they pivoted from VR to fintech. You have Airbnb founders who tell you that you need to push through hard times more than you think is reasonable.

Source: Gigs.

You get this intimate view into how some of the best companies in the world are built, while being on this fast-paced curriculum and having access to what is probably the largest brain trust on building companies. If you look at the role of an investor, if they're really helpful, they've seen hundreds or, ideally, in the case of YC, thousands of businesses being built. They don't know how to build your product, and that's why they invest in you, because you have this unique insight about your particular thing. But they are the largest machine on earth when it comes to building an actual business, making decisions as a founder as the company grows. You have the people who have seen the most of it in the world.

Then you end up being one of the best-performing YC companies. You used to have YC Growth, which was basically a graduate school of YC. For us, that was also immensely helpful, because for three consecutive years, we had a YC program tailored to the stage we were in.

So you had YC, then post-A, then Growth. It kept you in this ecosystem and gave you peers who were in a similar or more progressive stage than you. Wouldn't miss it for the world.

What does your day-to-day look like as CEO?

I spend a lot of time on hiring. Dennis and I still interview every candidate who joins Gigs. We used to be first, now we go last, because otherwise it would not be manageable. That occupies quite a lot of time. Then I work on more strategic topics. I'm heavily involved if we do a first-of-its-kind deal, like the first very large fintech or the first enterprise deal for enterprise connectivity. I work with the team on these types of structures so that they can be repeatable afterward. Usually, I disengage after that, but I'm still partially with the team on very important deals, even though they are of a similar nature.

…It’s my nickname.

I'm still quite heavily involved in the connectivity org, because we have the supply side—dealing with carriers, pitching them, selling them on the vision. While there's a growing number of carriers that love us now, there is still a lot of skepticism or even anxiety about what we do. It's an ever-evolving theme of how to address that. How to say “hey, we're actually here to help” because we're truly building this company to make carriers more efficient at the end of the day as well. The TAM for us is not just launching the services we launch right now. It's then deploying this operating system that we build into the carriers themselves. So I spend a lot of time there.

Then, the strategic product roadmap is used whenever we decide to build a product. Because it's such a complex business, I also spend a lot of time trying to see around the corner, what potential pitfalls we might be running into. And finally, fundraising whenever the time is right.

What do you think about hiring in deep tech?

I'm a massive believer in lateral innovation. It's a concept that's been popularized by Nintendo. The Game Boy was a prime example of that. If you look at the Game Boy, none of the technology in it was new. They didn't engineer something from the ground up. The displays were there, the computation was there. What they did was put the pieces together in a new format. That's, in short, lateral innovation.

One thing we've been driving at Gigs since day one is ‘we are building something that others have built in another industry before’. Stripe is, of course, the closest example. So we hire a lot of people from companies that have built these infrastructure companies in other industries. That works tremendously well, especially if you hire from Stripe or Adyen. The people who were there early, they come in, they know intuitively what to do, even though they have zero knowledge about telecom. Until we were about 130 people, we hadn't had a single person from telecom at Gigs. Very purposefully. So that, I would say,’s probably our number one go-to: just hire people who have done this before in a different industry, because they can lay out the map in front of them and then fill it with the telecom roadmap.

A happy bunch of Gigsters.

We hire heavily for stage fit. One of the biggest predictors we've seen of people not succeeding is that you hire someone from a really cool company that we love to have, but they were late. They were employee number 5,000 or 10,000, and they just sometimes don't know how to operate in an environment with this much ambiguity this early. Or what you get is that they were there when everything was already magical. All the problems had been solved; there was a rocket ship, and they lived through this fever dream of building it, but they don't appreciate how hard it was for the first 150 people to get there. How much of a shit show some of these companies were at these early stages. They only got the good part, so stage fit is a top priority.

Then we look for self-guided people. My go-to is, ‘You're the expert. I'm hiring you to be the expert in your field. I most likely don't know as much as you do and don't have as good an intuition for what to do.

Basically, you tell me. I can tell you where to go, but you tell me how to do your thing.’ That's something we prioritize a lot. High autonomy, self-guided missiles, great communicators. You need to be able to navigate the org. When you have a lot of ambiguity and not much structure, you just need to find a way and not be discouraged if someone doesn't reply or you don't have everything you need in front of you.

How are you using AI operationally at Gigs today, and how are you thinking about it?

We're currently building what I think many companies are building. This kind of ‘company brain’ where we record almost everything we do in meetings and funnel it into one central plane to eliminate information entropy and asynchronicity, and create one promptable source of truth about everything that's going on in the company.

Source: Gigs.

I've started creating an operating system for myself, too. It started very simply, ‘I run a 150-person company, and I'm struggling with this particularly difficult stage for companies—because at 150, things start to break.’ So I need to upgrade the company's operating system. I need to get better at this, and we need to get better at this as a company. I'm a bottleneck for people because, yes, on some subject matters, I need to review and approve. Now I have an operating system that makes it easier for me to run things and create a cadence and a communication system. We're also trying to implement some of the things many companies go through on their journey as AI companies, which we found resonated strongly with what we've been trying to achieve and the problems we're seeing, so we'll see where that goes.

Cash App unveiled mobile powered by Gigs. What does this represent for the industry?

Cash App is the largest personal finance app in the U.S., and now it's enabling millions of modern earners to have a mobile native product inside their banking app. Distribution of telecom and customer relationships are migrating into the apps people use and love. Cash App Mobile runs on AT&T's network, through Gigs—a trusted carrier, a software operating system that makes it embeddable and operates the subscriber base, and a fintech with tens of millions of users. There’s a population-scale migration to redefine what connectivity is and how it's embedded into the apps people already rely on.

Cash App's new mobile offering, running on Gigs.

We've already seen this play out globally. Revolut, Klarna, and Nubank all launched mobile offerings, and none of them treated connectivity as an experiment. Once they owned the primary customer relationship, adding mobile was a natural extension of what they already provided. Every major consumer platform will eventually offer connectivity because the economics make sense. If an app already manages your direct deposit, debit card, savings, and daily spending, it's a more natural place to manage a phone plan than a carrier store.

How do you get the best out of yourself?

When I'm in San Francisco, I try to go for extended hikes every weekend. That takes me out. The only downside is that I don't do it enough. I should be exercising way more, and I'm trying to get into these routines more, but I'm never anywhere more than three weeks. You build a habit, and now you're switching continents, the habit breaks again, and then you need to start over. So I'm getting better at this slowly. I have an exercise routine. You wouldn't believe how much even 30 minutes of light-to-moderate exercise a few times a week can change your energy, mood, inflammation, and all of that. That's been helpful. The good thing is, every time I actually do it, I really, really feel differently. I feel much, much better.

Real ingredients, no virtual stack.

I love cooking. That's something I picked up during college and then tried to drive to the highest level that I could for myself. It's relaxing for me because we're building these virtual things, and that allows me to be creative in the real world with real ingredients and stuff. That's something I enjoy doing. I read a lot too. I think you just need to. In a world that just bombards you with all kinds of bullshit, books and history are great guides to how things work.

Extra reading

• Embedded connectivity: A better economic model for phone retailers - January, 2026

• Gigs & Google Cloud: Inside the Reinvention of Telecom - December, 2025

• Mobile: A new frontier for B2B fintech - August, 2025

• Mobile services are unleashing new banking frontiers with Gigs - July, 2024

And that’s it! You can follow Hermann on LinkedIn or check out Gigs on their website to keep up with what they’re building!

BRAIN FOOD 🧠 

TWEETS OF THE WEEK 🐣 

Source.

Source.

Source.

TOOLS WE RECOMMEND 🛠️

Every week, we highlight tools we like and those we actually use inside our business and give them an honest review. Today, we are highlighting Framer*—the site builder trusted by startups to Fortune 500.

beehiiv: We use beehiiv to send all of our newsletters. Apollo: We use Apollo to automate a large part of our 1.2M weekly outbound emails. Taplio: We use Taplio to grow and manage my online presence.

See the full set of tools we use inside Athyna & Open Source CEO here.

HOW I CAN HELP 🥳

P.S. Want to work together?

1. Hiring global talent: If you’re hiring tech, business or ops talent and want to do it 80% less, check out my startup, Athyna. 🌏

2. See my tech stack: Find our suite of tools & resources for both this newsletter and Athyna here. 🧰 

3. Reach an audience of tech leaders: Advertise with us if you want to get in front of founders, investors and leaders in tech. 👀 

That’s it from me. See you next week, Doc 🫡  P.P.S. Let’s connect on LinkedIn and Twitter.

👋 Howdy to the 5,687 new legends who joined since our last edition! You are now part of a 503,402-strong tribe outperforming the competition together.

LATEST POSTS 📚

If you’re new, not yet a subscriber, or just plain missed it, here are some of our recent editions.

🧠 Building A Second Brain With Claude Cowork. The ultimate guide to one of the most elusive tools.
✍🏻 Building The Incorruptible Companies Of The Future. An interview with Eric Ries, Author of Incorruptible.
🐝 When David Ate Goliath: beehiiv’s Zero To One. The scaleup with attitude that’s killing the competition.

PARTNERS 💫

Meetings aren’t the problem. The challenge is everything that comes after them. With Granola, the AI Notepad for people with back-to-back meetings, it's no longer about scrambling to keep up.

Take notes the way you always have. Granola works in the background, turning conversations into clear summaries, action items, and next steps. Before, during, or after a meeting, you can chat with your notes to quickly understand what needs to happen, write follow-ups, or share context with others.

You're paying for a CRM, a sequencer, an enrichment tool, a call recorder, and an agent builder. Lightfield replaces all five.

One agentic CRM with built-in agents that build your pipeline, prep your meetings, send follow-ups, and keep records current. Connect email and calendar, and it starts working in your first meeting.

_{Interested in sponsoring these emails? See our partnership options here.}

HOUSEKEEPING 📨

Not sure if you noticed, but I tried something new in our most recent edition. Firstly, the content type was new, but also the Housekeeping section was replaced by a Heat Check section. Heat check, meaning a poll on a trending topic; in this case, the budding Apple v OpenAI court case. Here were the results.

I read a newsletter called The Tangle, which does a number of pulse check-type surveys on the U.S. political ecosystem, and I find it fascinating. We have such a large audience now, I figured it could be an interesting way for us to ‘read the room’ in tech, so to speak. Did you like it? Would love to get some feedback. Thanks!

ZERO TO ONE 🌱

Tyler Denk, Co-Founder & CEO at beehiiv

Tyler Denk is the Co-Founder & CEO at beehiiv. He was Morning Brew’s second employee, and in four years, he led product, engineering, and growth, building the referral program, content management system, and ad stack that helped Morning Brew scale to 4M readers before selling to Business Insider for $75M in 2020. After a brief stint at Google, Tyler co-founded beehiiv in 2021 with two former Morning Brew engineers, Ben Hargett and Jake Hurd. The company launched in October that year and has since raised roughly $50M, including a $33M Series B at a $225M valuation in April 2024. Revenue has tracked from near zero to an estimated $30M in annualized revenue by mid-2025, with clients now including Time, TechCrunch, Newsweek, and Sinclair.

beehiiv wasn't built by someone who spotted a market gap from the outside. It was built by someone who had lived inside the most successful newsletter of the past decade, was frustrated that nothing on the market could do what they'd already built internally, and decided to do it. He builds in public, ships product every week, and runs a personal newsletter (Big Desk Energy) with over 100,000 subscribers. The result is a company that has taken direct aim at Substack's business model, treats product velocity as a genuine operating philosophy, and has climbed from scrappy challenger to platform of record for some of the biggest media brands in the world.

Why are you suited to build this product?

I was the second employee at Morning Brew, back in 2017, and did everything from product and engineering to growth. I was fortunate enough to hire a bunch of really talented people, but two of the most talented were the engineers who've since become my co-founders, Ben and Jake.

Tyler, Benjamin, and Jacob.

Where a lot of people have built companies in this space purely from the angle of ‘I want to build a software platform,’ we came at it differently. We'd already built this internally at Morning Brew as a case study, and it worked extremely well. We scaled to four million readers and a $75M exit. We sat next to the content team (the people actually creating the newsletter day-to-day) and really understood what they wanted from a product to build an incredible newsletter. On the other side, we sat next to the sales team and all the brand partnerships people, and understood what they were doing daily to sell Morning Brew, to monetize it, and to do the tracking. We're still very true to the newsletter ecosystem today. We've expanded to a bunch of different products, but the newsletter is definitely at the core, and I think Morning Brew is one of the best case studies of a successful newsletter in the past decade.

What made you confident you could take market share?

The way I look at that question, most of the incumbents are typical ESPs. Dozens and dozens of email platforms that, fundamentally, send emails. If you were creating content on the internet and wanted to do what Morning Brew did—website, newsletter, growth tools, monetization—those are all different pieces. When I was there, we used MailChimp, then Campaign Monitor, then Sailthru. All three only did one of the fifteen things we needed as a scaling media company, and that was sending emails. As someone who'd used all of those ESPs, I knew how frustrating they were and how little they'd invested in the product experience. They were also expensive. A lot of them had gone through acquisitions and weren't innovating the way hungry startups do when they're chasing market share. There was also a consolidation opportunity. The product itself only did one of many things, and we wanted something that did a lot of things together to help us build and scale.

Source.

And then there was Substack, which came to market about three or four years before beehiiv. At the time, the product was genuinely poor, and a lot of people on Twitter were very critical of it. They'd clearly struck a nerve with the independent media, newsletter-first, VC-backed angle, but I was pretty confident that what we'd built internally at Morning Brew was already more advanced and more capable than the core Substack product. If we could just productize what we'd built internally, there was a huge market, and all of the signals pointed in that direction.

Who were the first people you brought onto the team?

The initial team was me, plus Ben and Jake, my two engineering co-founders. They were still employed at Morning Brew, had good jobs, and it took a lot of convincing to get them to make the leap. But they bought in during the summer leading up to launch—summer 2021. We also got Andrew Platkin to join as our initial CTO, so before we'd even launched, it was essentially four engineers. The thesis on those first hires was simple: we are so far behind. All of our competitors had been around for years. There was a massive surface area to catch up on. It took five engineers and four years to build what we built at Morning Brew, and that was for one customer. To build it for thousands of customers meant an incredibly heavy engineering roadmap, so the first five hires were all engineers.

Source.

Eventually, we got to a point where we had a few hundred paying customers, growing a couple of dozen sign-ups a day, and while the engineers were building, someone had to handle everything else. Customer support, onboarding, emails, customer success. It was all falling to me while I was also trying to build, so we made what I called a ‘community hire’ (really, everything that wasn't engineering). Around employee seven or eight, we realized that person wasn't quite the right fit for the role. That's what got me talking to Preeya. She had been my Account Manager at Sailthru, where Morning Brew was a customer, and she was incredible. Any time I had problems or was creating headaches, she was super quick to address everything. Professional, knew her stuff better than any other account manager or vendor I'd ever worked with. I went to her to get a read on our support situation. She confirmed what I suspected about the existing hire, and then came in to replace them, not in some entry-level role but as Chief of Staff, taking over everything that wasn't engineering. So the first ten people were roughly seven engineers, Preeya running everything else, and EJ as our first growth lead.

What has Preeya's impact been since she joined?

Preeya is incredible at what she does. She spent six or seven years at Campaign Monitor and Sailthru, so she knows the email industry like the back of her hand. Sailthru is a very enterprise business with around five hundred customers, all high-paying media companies. She's handled customer comms at that level and is deeply familiar with how the biggest players in our industry operate.

Preeya, replying at speed.

She's also just exceptionally well-spoken and organized. Any time a user needed to complain about something, that used to fall entirely on me. She became someone who handles it better than I can, and brings all of the soft skills around communication and customer success that I don't always have the bandwidth for. And she never drops the ball. Whether it's 2 am her time, she'll fire off a message with an update, nothing's slipped through. The more you scale a company, the more you realize how rare that is. Someone you can hand any task to and it just gets done, and done well. A lot of people need constant follow-up. She's the opposite of that. She takes ownership and pulls things off your plate. At an early-stage startup, you need those jack-of-all-trades who can just execute.

How did you win noteworthy publishers in the early days?

Even before Milk Road, our first big customer was Liquidity, the meme account. The way we got them was straightforward, as he'd invested in the company. That was the shortcut in the early days. I went through all of our initial angel investors, and a number of them had newsletters. There was a guy named Gannon Breslin who had a newsletter of about 25k people. Not a huge name, but we needed traction.

Source: beehiiv.

We raised from about 25 angels, and I'd say 20 of them moved their newsletters over within a month of us launching. Forever indebted to them for that. Liquidity was definitely the biggest one. He had 100k subscribers on MailerLite, and we couldn't do more than 10% of what it could do at that point. He just totally trusted us, and we figured it out. And then it's a game of snowballing success stories. Now that we had Liquidity on the platform, we could turn that into a case study for all the other publishers and newsletters who looked up to him. The narrative became, ‘This is the founding team behind Morning Brew's internal tech. We scaled that to 4M readers. We already have Liquidity with 100k subscribers, and his open rates are better.’

The other big early success story that jumps out is Matthew Berry, the ESPN fantasy football personality who left ESPN to build his own media company. We had a mutual connection, hit it off, and he moved off PlayVio onto us. That was 300k subscribers. So within the first month, we had Liquidity, a few months later, Matthew Berry, and Milk Road were right around that time too. We recently signed Time as a large enterprise client, TechCrunch, Newsweek, and parlayed that into closing a deal with Sinclair, which is a massive media conglomerate. You keep doing that, climbing based on credibility, logos, and case studies.

What was the lead domino in the zero-to-one stage?

Product velocity and product-led growth were our MO and what we're best known for. But it truly came out of necessity, almost out of insecurity. When Liquidity moved his newsletter over, we could probably do 10% of what MailerLite did, and I knew it. Every single day, users reached out to complain that we weren't up to par with other platforms. I'd wake up to 15 angry emails from people asking for things they could easily do on their old platform, threatening to leave if we didn't build it. And Liquidity alone represented about a tenth of our revenue. So I literally had a tenth of our revenue threatening to churn in my inbox every single morning.

That anxiety and urgency are what pushed us to move quickly and hire more engineers. Our entire product launch strategy came from three buckets. One is educating existing users. When you launch a new product announcement, you're basically saying, ‘I heard you, we built it, it's live today.’ You're preventing churn from the people who were about to leave. Two is marketing. When you launch new features publicly, you're checking boxes for people who were watching but hadn't switched yet because you were missing XYZ. And three is the combination of both. When you do it every week, you change the narrative from ‘they're missing a lot of features’ to ‘they ship so quickly that even if they only have 70% of what I want today, I'm confident they'll have 100% in a few weeks.’ That was the biggest lever. It lets us punch well above our weight. Layered in there is the building in public. I'm showing that I genuinely care, that we do listen to customer feedback, and that I'm personally in DMs answering questions and prioritizing tickets. Those things combined drove a lot of our early growth.

How have you built and maintained that product velocity?

Part of it is what I just described. It was build-or-die for the first two years. Only now I feel calm enough that if something we planned to launch tomorrow isn't ready, I can push it to next week. In the early days, that would have killed us. Every day we waited was potentially costing us churn. More tactically, of our first ten hires, seven were engineers. We made it very clear we were an engineering-first company. We also hired full-stack engineers who skewed heavily toward product and design, and I think that's the real secret.

One of my co-founders, Ben, is the most creative, design-forward engineer I've ever come across. Whatever he touches is beautiful, and he doesn't need mockups at all; he just has taste. So we cut out the need for designers entirely. No design handoff. No front-end to back-end handoff. No separate QA process. The engineers QA'd themselves, and the product was less complex in those early days, so that was manageable. Once you build that muscle and create that DNA, once the pace is what it is, it's really hard to slow it down. People would feel bored if we stopped. We've established that this is our operating cadence, and that's just how it runs.

What were the most difficult moments in the zero-to-one stage?

Every day felt like a roller coaster, honestly, and it kind of still does. On the operational side, raising our seed extension was a real education. We'd raised $2.6M at seed, and as a first-time founder at twenty-seven, you naively think the money will last forever because you can't even comprehend how you'd spend it. Then you start hiring good engineers who have well-paying jobs and want a 401(k) and health insurance, and you hire a growth person who wants to run paid acquisition. So you realize $2.6M isn't a lot.

Source.

We had to go out and raise a seed extension about six months later. I naively thought we were at 15-20k MRR and growing fast, given clear product-market fit. I walked into that process thinking we'd close in a week. It took three months, tons of passes, and we ended up closing with a couple of UK-based funds. Every minute of that process was painful because we were fighting fires and trying to scale revenue with not much runway left. A lot of those early-stage funds took me for a ride. Email deliverability was another one. There's a lot of infrastructure behind the scenes (IP addresses, sender reputation), and we got burned a few times early on as a learning experience. We'd tried to isolate our best senders from unproven senders to protect deliverability, which is pretty standard for any ESP. We just moved too quickly, and our best senders ended up in spam for two weeks.

That was brutal because we'd essentially taken our highest-paying customers and screwed up the one thing we were supposed to do well: deliver their email. The fix wasn't a quick bug patch. The only way to correct email reputation is to wait it out while it course-corrects over three weeks. So every day for three weeks, I was fielding threats from users wanting to leave, taking it to social, furious. I actually sat up one night with one of our investors and drafted a ten-tweet apology. Then we decided against it. It would have made it worse, and email deliverability is too nuanced to explain on Twitter anyway. And then when Nathan Barry and ConvertKit started emailing our users to tell them about the mistake—essentially saying they would never do that at ConvertKit because they actually know what they're doing—that was infuriating. But we were also genuinely vulnerable. Not much cash, best users are all experiencing a problem we've caused. Not a position of strength.

Does going into battle against rivals energize you?

Totally. I think every company benefits from having some kind of enemy or rival. Even outside of companies, like sports teams, having a rival you actually dislike, even a manufactured one, drives competition and urgency. It's why cost-plus contractors and government work have no incentive. No competition is driving anyone to be better. Michael Jordan would just manufacture rivalries out of thin air, make them completely fictitious, and it worked.

Source.

I think doing it the right way, where other people in the company also feel motivated by the mission, matters. The business model we've built, and what we're empowering our users to do, is more creator-friendly than most of what's out there. When you actually believe the work you're doing is the right call, and you know other companies are trying to take your market share and your customers, and you care about it that much, that's a genuine emotion. The more emotionally invested and passionate I can be about this company, the more I can build and accomplish without burning out, and the more I can rally a team around it.

What’s your approach to personal wellness? And how do you transfer that to beehiiv?

On the personal side, I've always been active. Workouts, physical activity, and eating well. I've always prioritized my physical health. Mentally, I was firmly in the ‘I'll fix mental health later’ camp when I was younger. I don't need to sleep, I don't need to meditate, I'll just push through the burnout because the exit will solve all of this eventually.

Travel wears on me a lot. As the company has grown, I've been trying to get better at saying no, but I still end up flying from South America to London for a three-day trip to meet with top media executives because the comms team insists I have to be in the room. And then there's SXSW, Cannes, all of that. Last year, I hit a real wall of burnout trying to manage a 60-hour work week while also jumping between time zones. I realized I had to actually change things, or I wouldn't be able to keep doing this. A lot of small habits have added up, like meditating and cutting back on coffee. Even after I'd finished everything on my to-do list for the day, I'd still feel this anxious sense that I wasn't doing enough. When I cut coffee, that went away immediately.

Being on the West Coast with most of the team on the East Coast or in Europe, I used to wake up at 5:30 am, already drowning in Slack messages about what was broken, angry customers, and fires to put out. That's the least healthy way to start the day. Now I do a full morning routine before I even touch my phone. I want space to actually think for myself before I'm reacting to everything I'll have to deal with anyway. I can wait thirty or forty minutes.

Today, I'm working more than ever, and I feel the least stressed I've ever been. Which I find genuinely interesting. For the team, the ‘monthly wellness day’ came from my time at Google during COVID. Sundar did it once, and the whole company loved it. A three-day weekend, out of genuine compassion for the team during a hard period. I thought it was such a brilliant idea that I said we'd do it before we'd even launched beehiiv. We've done it ever since.

Source.

I'm careful about talking about a four-day work week right now—we're burning a million dollars a month, and there's a lot of runway to cover before that makes sense. But we're at a point in the company's maturity where employee satisfaction and retention are genuinely one of the most important things we can focus on. If a three-year engineer leaves, it takes four months to find someone equivalent, two months to onboard them, and a year to match their context in the business. It's far better to pay people well and create a place they actually want to be.

Preeya Goenka, Chief Customer Officer at beehiiv

Preeya Goenka is Chief Customer Officer at beehiiv. She joined the team in late 2022 as employee #7, first as Chief of Staff and then Chief Operating Officer, before being promoted to Chief Customer Officer in January 2026. Before beehiiv, she spent six years at CM Group (now Marigold), rising through customer success roles across Sailthru and Campaign Monitor, then ran go-to-market strategy at Noogata, a Series A AI company.

Preeya first met Tyler Denk nearly a decade before joining beehiiv, when she was his vendor at Sailthru, and he was building Morning Brew. When he eventually pitched her on joining, the company had almost no features and was entering what looked like one of the most crowded markets around, yet another email platform. She joined anyway. For the first two years, beehiiv spent zero dollars on acquisition. Everything ran on personal onboarding, white-glove support, and a roadmap built almost entirely from what customers actually asked for. By the time legacy publishers like Hearst were doing due diligence on the platform, that early customer obsession was already baked in.

How did you end up at beehiiv?

Almost four years ago, I was employee number seven. I had met Tyler about ten years before joining, when I was working at Sailthru and Campaign Monitor. That was a portfolio of email companies I was at for about six years, while Tyler was at Morning Brew. I was one of his vendors. We worked together pretty closely for a couple of years, kept in touch, and then he launched beehiiv. He was very active on social even back then, so I was following along from the sidelines, watching him post about the launch, the announcement, and how the early days were going.

One day, when we were catching up, and he was filling me in on what he was building, he was picking my brain on how to think about hiring someone to own pretty much everything else besides engineering.

Someone to focus on the customer experience, work with higher-touch enterprise customers, and manage the broader customer base. The more he talked about what he was building, the more interesting it seemed. I had a background in email, I knew Tyler really well, and I had a lot of confidence in what he was building. So I took a bit of a risk and ended up joining as Chief of Staff.

Preeya and Tyler.

I remember in one of those early conversations, I was semi-teasing him, but also kind of probing why he decided to start an email company. From my perspective, it was the most crowded space. But he was telling me about his vision, and yes, it was an email platform, but there were so many other components he was layering on top. He had a very clear vision for the direction he was building in and a long-term plan for what the product would be. The way he envisioned building beehiiv was very much about filling a real gap in the market. Very different from any other email platform I was familiar with, so it felt like a really strong proposition and a good opportunity.

What did the role actually look like in those early days?

I was definitely a jack-of-all-trades. The company was almost all engineering. Tyler was pretty much doing everything around product and growth. Everything else was kind of under my scope. Mostly all customer-facing work. For the bigger customers already on the platform, I was their high-touch point of contact at beehiiv. It was a combination of support tickets, managing the entire sales, product demos, post-sales, onboarding, an ongoing success flywheel, and then layering on operations. I can't really call it operations because there aren't really operations at an eight-person company. It's just about doing the work while establishing some kind of process as you go. That was probably about 50% of the work.

The other 50% was things that popped up every day that Tyler didn't have bandwidth for. We'd tag-team whatever had to get done. Every day felt like a list of 50 things. Tyler would say, I'll do these 25, you do these 25, rinse and repeat. A lot of putting out fires, a lot of small things that had to get done, a lot of chipping away at the bigger items. We were just bouncing around different areas of the company, trying to get things done as fast as possible.

How did you win customers when the product barely existed?

A few different things. Tyler was a newsletter person through and through, and I had some of that email experience myself. So we led with that in all these conversations. The message was that we understand your problems and challenges, and we're building a platform specifically to address them. I felt like that touched a very raw pain point for many of the customers we were working with at the time. They all probably agreed that the email platform they were using wasn't the perfect fit for them. beehiiv wasn't the perfect fit yet either, but we were building in the direction that actually addressed their problems, versus other platforms that were building e-commerce solutions that had nothing to do with what these customers were trying to do.

The hive.

We were also extremely user-focused. If a customer said, "I like beehiiv, but I can't use it until you have these five things," we'd say, "Let me get back to you in a month." And we would actually get back to them in a month with those features built. Even now, a large portion of our roadmap is driven directly by customer requests. Back then, it was probably 50/50. Half was what we knew we had to build to be a competitive email platform, and the other half was just what prospects told us they needed or what customers already on the platform said they'd churn without. I don't think a lot of these newsletter operators and publishers were used to that kind of high-touch support. They definitely weren't used to companies actually following through on feature request promises. That went a very long way.

What was the main pain point you were solving that others weren't?

No other email platform, or maybe none at all, really understood the workflow of a newsletter-first operator, a publisher, or a media company. What it actually takes to create an email and get it out to your audience.

A lot of the people we talked to had very clunky workflows, or they used their email platform only for email and had eight different integrations, plugins, and add-on tools to do everything else. The people building those platforms didn't really understand the end goal.

beehiiv landing page at launch.

beehiiv had one very singular use case: we're built for media, for publishers, for people sharing their content. Every step in the email creation workflow is intuitive and built for that purpose. Newsletter operators care most about two things—growth and monetization. We're an all-in-one solution with features built into the core platform for both. If you're using MailChimp, you probably have six different integrations to grow and monetize your list, since MailChimp doesn't do it natively. We understood the pain points of what these people were trying to do and built a platform to address them directly.

What are the early wins that stick out?

There was a phase when we were probably post-Series A but still very, very young, both as a company and as a product. We were talking to some of the world's biggest publishers. A lot of massive global media companies were in our pipeline and actively doing due diligence on beehiiv. In that moment, win or lose, the fact that we were sitting at the table alongside enterprise solutions as a two-year-old company was actually insane. It was a great learning experience. You understand how these companies actually buy, and it's a very different world when you're selling to organizations like that. But it was such a good feeling, because I came from an enterprise email solutions background where those companies had been around for ten or fifteen years. Just having a shot at that kind of business was a nice win.

Source.

The other moment that sticks out is around the ad network. One of our big value props is that whatever you're paying for beehiiv, you should ideally be earning a lot more through the platform's native monetization tools, whether that's paid subscriptions, the ad network, or something else. We don't take a cut of any of it. When we first rolled out the ad network, we didn't really lead with that value prop because it was still new. But I remember one of those early calls, when someone had this aha moment, realizing they could earn this much money from running one ad, which was more than what they were paying for beehiiv for the entire month. And that's exactly what we're trying to do here. Having them arrive at that conclusion before we'd even prompted it was a really nice feeling.

What was the most important thing beehiiv did to drive early success?

It comes down to how much we focus on the customer. Every SaaS company loves to say they're user-first. I've worked at several other tech companies, bigger and smaller, and none of them came close to how much beehiiv focused on the user experience. That's definitely true now, but in the early days, when every customer was kind of make-or-break for the business, it was even more the focus.

Being only seven people, we all had to do it. We personally onboarded every customer, listened to their feature requests, and hopped on calls with them. It was ‘Let me personally import your list. Let me personally help you build your template. Let me keep an eye on your analytics. Let me talk to you a few times a week to make sure things are going well.’ I can almost guarantee they'd never had that kind of high-touch, white-glove experience anywhere else. And that kind of message travels very fast through word of mouth.

Source.

As people started using beehiiv, they knew it might not have these ten features they needed, but they had our entire exec team on speed dial, and the way we were paying attention to them is so invaluable. So they'd tell their newsletter friends, post about it, and word traveled fast. Which also connects to the other part: we didn't spend a single dollar on acquisition for the first eighteen to twenty-four months. Everything was entirely through word of mouth and social media. Tyler was building in public. We'd share customer stories, case studies, and wins. When customers shared their newsletters, we'd repost them, engage with them, and help amplify their messages. That energy on social is what propelled us in those early days.

Why are you still at beehiiv?

It doesn't feel like four years. Every six months, beehiiv feels like a totally different company. Going from seven people to where we are now, there's just no comparison. It's night and day. So it always feels like a fresh, new chapter.

Hive minds.

And even four years in, I feel like we're still so early in terms of what we're building. We start every year with a twelve-month roadmap that feels, in January, like we have all the time in the world to build all these incredible things. Then you start slotting things into the quarter and realize the ten things you thought you'd get to this year won't even fit into the next year. There's just so much left to do.

What makes Tyler good at what he does?

He's extremely plugged into the ecosystem, the market, and the customers. He truly understands their pain points and how they operate. He understands the competitive landscape, who the other players are, and why someone might use a different platform. He talks to customers all the time and does lots of sales calls. He's deeply ingrained in the space, and that makes it a lot easier to build the right product.

Source.

He's also very good at staying focused. Many founders and CEOs get distracted. They want to raise this round, chase the next thing, do this. He knows what he's committed to, and he sticks to it. There isn't much space for that kind of distraction with him. He's very focused on what he needs to do and build.

Has the creator economy grown the way you expected?

Way more than I expected. But its definition has also changed a lot. Before, the creator was very much a side hustle that might become the main thing. Someone with a podcast or writing in their free time, starting to monetize it a little. That was largely what the creator economy looked like when we started. Now it's very much a business, and people operate it like one. And what I didn't foresee was the emergence of a hybrid world, bridging legacy media and business with new-age media and the creator economy. Companies like Netflix and Spotify, which were in their own lanes in how they approached media within their specific verticals, are now leaning in more and more with content creators. Retail companies, clothing brands, and companies I never thought would care so much about creators are now investing very heavily in that space.

Source: beehiiv.

Source.

Creators are running themselves more and more like businesses and finding ways to work with the legacy companies. You see independent journalists who are fully independent but also working in some capacity with larger broadcast media companies. There's this really interesting convergence of many different elements that I didn't foresee when we started.

Shreya Srikanth, Product Manager at beehiiv

Shreya Srikanth, Product Manager at beehiiv, joined the team over three years ago as the first product hire in the company's history. When she joined, beehiiv was still at seed stage, the whole team was fewer than twenty people, and Tyler was personally PMing everything himself. Since she teamed up, beehiiv has crossed $30M in annualized revenue by mid-2025 and grown to serve over 55,000 active creators.

Shreya is not a founder, but she has the institutional memory of one. She got the job by cold-reaching out to the CEO with a list of product ideas that matched where he was already thinking. Three years later, she manages the core platform roadmap for a product that ships faster than arguably any SaaS company of its size from a structure of three PMs, two designers, and roughly 50 engineers that by any normal logic shouldn't work, but somehow does. Her perspective is from inside the engine room, from the nights triaging support tickets with Tyler to now.

When did you join beehiiv, and why?

I joined in May 2023. At the time, I was looking for an early-stage company, a startup environment, and I wanted a product that people genuinely loved. While doing my LinkedIn research, what stood out to me were the comments users were leaving on beehiiv's posts. People were raving.

You know what they say about products: ten people who love it are worth more than hundreds who just like it. I could see that clearly from the users and then also from Tyler, EJ, and a lot of the early team. They were so active and engaged.

There was this real human touch, almost a craze around the product, that made me feel like there's something genuinely cool here.

How has your role evolved since then?

I was actually the first product hire at beehiiv. Until then, it was just Tyler PMing everything. I remember being super excited. I reached out to him with a list of ideas and things we could build, and it resonated with him because it was roughly where he thought the product was headed anyway. I joined, and it was, as it is with any early startup, kind of no man's land. You do fifteen different things, but my main role in those early days was to help make sure the team was always building the most important thing next. Keeping that organized across workflows and helping execute. I was also leaning heavily on user feedback. We didn't have a fleshed-out support team, so I was the last barrier. I was triaging that every night and looping in engineers the next morning. Tyler and I did that together every single night for a while. It was hectic; I honestly don't know how we survived it. At that point, Tyler was still at the helm of the high-level vision, and I was his right hand for executing day-to-day operations. I learned a lot during that process.

The last barrier, pictured.

What do you think fostered that early craze around beehiiv?

Definitely the fact that people got direct access to the people building beehiiv. Every early user got a DM from Tyler on LinkedIn when they were verified. People would get responses to their tweets and support tickets, and Tyler would personally sign off on them. I think you feel really special when the company's CEO is paying attention to the issue you're facing with their product. It makes you feel genuinely valued.

Yup. Source.

Even today, every PM talks to at least two or three users each week and helps a brand-new user onboard to the platform. We've maintained that 'do things that don't scale' philosophy, getting an actual human in front of these users. That adds so much to how people feel about the product and the company.

We’ve always been so in tune with what users wanted, constantly monitoring everything. I think that's ultimately what helped us build the right next thing every time. The people using it were telling us exactly what they wanted.

How have you built a culture of shipping so fast, at such high quality?

Product velocity is definitely our superpower, and I think it comes down to the people and the way we approach projects. In traditional product management, it's always weeks or months of figuring out what to build, a long document, fifteen stakeholders reviewing it, rounds of back and forth, then it hits engineering for a tech review, then design, and it's this whole process. At beehiiv, the moment we identify something exciting to build, we have a PRD within 24 hours. With Claude Code now, I'll usually generate wireframes or mockups right away, and designers and engineers jump in immediately. We're starting development and design work before the vision is even 100% set in stone. We tackle each problem as it comes up, as long as we know the core opportunity and the core functionality we need. We have really quick feedback loops.

Sometimes we'll be launching something the day after tomorrow, and someone comes in with 'it'd be really cool if we did this.' Nobody asks why they're bringing it in so late. Everyone wonders how we can make something work in such a short timeframe. That attitude is pretty rare, and it comes from the people's mindset and everyone being willing to roll with the punches. Everyone is a kind of equal stakeholder in solving the problem. No rigid lanes. All the work happens in parallel.

Source: beehiiv.

Our engineers are genuinely exceptional. You can give them a ticket that's one sentence with a Slack thread linked, and they'll come back with something 80% of the way there. PMs and designers come in at the end and make a couple of tweaks to improve user-friendliness. The caliber of talent means you can give someone a vague prompt and they'll just run with it. Everyone who joined early has that level of ownership. The last piece is that everyone is working on multiple things at once. An engineer is typically working on a feature, while also building a one-off request, and probably tackling bugs from a previous feature, too. All at the same time. Every day, people are context-switching. We try to protect people from that, but honestly, the fact that everyone's been doing it constantly is what lets us launch fifteen things in a quarter instead of three.

What does the PM-designer-engineer structure look like?

We have three PMs plus a Head of Product, two designers, and somewhere around forty to fifty engineers. So the product team is very lean. That works because our PMs—especially with the current AI tools—generate mockups and wireframes tied to the design system our designers have built. By the time a concept hits a designer, they're not building it from scratch because they can already see the PM's vision and jump from there. The same logic applies to engineers. Everyone's empowered to run, and that makes all of our roles pretty scalable.

What was something that bombed in the early days?

One thing that didn't go well was that we'd launch a feature and immediately move on to the next one. There's a feature called ‘Boost’ that, in theory, could have become a multi-million-dollar revenue arm for the business. Sparkloop is an entire company built on that same mechanic. But we launched it and left it as is. It just became a feature people complained about constantly. It became a real time and energy sink for a long time. We couldn't even invest in fixing it because we were always moving forward to the next thing. I have many examples of that same story. We launched at maybe 30% of what the feature needed to be, then moved on. We never got the full benefit for users or for the business.

Source: beehiiv.

That said, I think shipping fast and shipping early was actually the right call for that stage. Back then, people would have left the platform if we didn't have even a basic version of certain features. There's a time when you just need to say “we have this”. The key is that the team has to evolve its approach as the company grows. In the early days, a bug wouldn't have impacted anyone meaningfully. Now, a bug is affecting Time magazine's newsletter, which has 400k subscribers. The stakes are completely different. A big part of my roadmap today is going back, simplifying features, making them actually useful, and tracking proper success metrics. It’s not launching new functionality, but going deeper on what's already there.

How did you think about competitors?

There were two categories. The old incumbents just confused us. They never seemed to ship anything. We really wondered what everyone was doing over there. That said, we respected what they'd built since they clearly figured out the constituent parts—the foundation for newsletters, email campaigns, and traditional email use cases. I'd always have a FigJam open with screenshots of every flow and believed we could do that ten times better. Not easily, but we could. They were inspirational in the sense that they showed us the users’ needs. Our job was to figure out how to do it better.

What shipping fast looks like.

The newer competitors were more serious because users would actually jump between them and us. They were real competition. But with Substack, we could clearly see they were moving in a different direction (notes, social, discoverability, and algorithm). That was actually really helpful for us, because it forced us to find clarity. We realized we weren't actually a Substack competitor. Internally, we don't even really think of them as direct competition. Watching them go social helped us sharpen our own message—own your audience, direct relationship, no algorithms. With Kit, the relationship is a bit more combative, at least from Tyler's side. But that also came from the fact that they ship very slowly, and they spent months just talking about their rebrand. That made us lean harder into our own strength: we're actually shipping functionality. Every time a competitor did something, it helped us discover that we have our own lane. Now, when I scope a new feature, I'll still look at what competitors are doing, but we're increasingly building on our own instincts.

What cadences does the product and engineering team live by?

We have quarterly roadmaps, and within that, we run three-week sprints followed by one-week cooldowns. Each quarter is three sprints and three cooldowns. The sprints are where the bulk of feature work happens. Each engineer typically has one to two major features per quarter, and each sprint usually covers one project. Cooldowns are for bugs and quality-of-life improvements. We always set extremely ambitious timelines, and as we approach those deadlines, the decision to launch or hold a bit longer kicks in.

Source.

Each week, we have two company-wide syncs where everyone shares what they're working on. There's also a weekly tech sync for engineering, product, and design. Then the individual project syncs, plus people are always ready to hop on a huddle. Every day, we try to list the 10 things they want to get done. Tuesdays and Thursdays are no-meetings focus days, so those are the days people really grind through their lists. And there's always a rough announcement date from Tyler for new features, which is the North Star. We do our best not to waste a single day.

What's changed inside product and engineering since you joined?

The big mindset shift is that back then, the question was around the bare minimum we could ship, while today the question is what's the best possible version of this feature.

We're growing into that, actually asking ourselves how we can make features delightful. We never used to talk about delight in the early days—it was just functionality. Now we have so many more conversations about aha moments and about what makes this fun for users. That's been the big shift.

Source.

The second thing is data. We never looked at data back then. Someone was commenting on a LinkedIn post, so we figured people were probably using the feature. Now we actually pull the numbers. We look at how people are using it, how much they're earning from it, and use that to drive decisions. More data-driven thinking, plus quality over quantity and functionality. Those two things are the biggest changes.

What's the vision for the product going forward?

The original vision was simple: nail newsletters. The natural evolution from there was thinking about what else these creators use. We realized that most serious beehiiv users are also using WordPress, Discord, and something else on top. So the next step was clear: build an insanely good website builder because that's where your content lives. And we built that.

Then we looked at podcasting. We tested the waters by adding an audio feature to newsletters that let you convert an issue into audio, and we saw a big uptick. And doing market analysis, we found that people with podcasts were actually trying to get into newsletters, and vice versa. So we built podcasting. The next step, coming in July, is community. You've already got all these ways to reach your audience with content—newsletters, podcasts, your website—but how do you actually turn that into a meaningful two-way stream? A lot of our users' current stack includes Discord, Reddit, or Slack. We want to replace that.

Source: beehiiv.

Source: beehiiv.

The goal is to be the one-stop shop for everything a creator needs to run their business. On top of newsletter, website, podcast, and community, you have subscriptions and the ad network. And we don’t take a cut. Eventually, the goal is to let you run ads across your podcast, newsletter, and website. Right now, that's only live for newsletters because that's where we started. Every way you can reach and engage your audience directly, and every way you can turn that into revenue. That's what we're building toward.

Why do you still work at beehiiv today?

Honestly, so much of it is the people. One of my teammates described it well: it feels like being on a high-powered sports team. Everyone does their part, everyone's a go-getter who's really good at their job, but everyone's also fun and goofy. Our all-hands calls and our Slack channels are genuinely silly. People can be themselves. It's not a cutthroat startup environment. People here actually care. We get a Friday off every month—we call it a Monthly Wellness Day—because the founders are committed to people not burning out. That balance, where everyone works really hard and strives for their goals but also shows up as a human being, is rare. Everyone at this company also has some side project or passion project they're absolutely killing at, on top of their job. It's an inspiring place to be.

The company keeps getting better, too. Leadership is genuinely receptive to feedback. There've been times when things were more stressful, and I've shared that directly and actually seen things change.

That matters. You want to work somewhere you feel valued and where your feedback means something.

And I actually love the job itself. There's no bureaucracy, no politics. You can drop an idea in Slack, people jump on it, debate it with you, and the next thing you know, it's something being worked on. That doesn't happen in most places. Most jobs get mundane, repetitive, and bureaucratic. Here, it's a creative space where everyone can pitch in, and everyone cares.

What are Tyler's strengths?

Setting an ambitious pace. He'll come in with the important thing to solve, boost us, and set a timeline. It's almost like a challenge. Instead of sitting and debating it for three weeks, let's actually put something together. You get people excited, rally them, give them a timeline, and nudge them toward it. That ability to create momentum on a dime is genuinely one of his strengths.

Shreya and Tyler behind the scenes.

He also works harder than anyone I know. He'll personally DM and respond to every employee. Even at the scale we're at now, he knows what every single person is working on. He'll jump into a random thread, give feedback on something, and immediately go back to whatever he was doing that's ten times more important. That no-ego thing—I will jump into anything and unblock whoever needs it—he genuinely embodies that, and it carries through the whole company.

Extra reading

• The Acquisition Mechanics: beehiiv + Swapstack - January, 2025

• How beehiiv Raised $50M (Fast) - April, 2025

• beehiiv’s 5-Year Sprint to $20M ARR - August, 2025

• When David Ate Goliath: beehiiv’s Zero To One - July, 2026

And that’s it! You can keep up with Tyler, Preeya, and Shreya on LinkedIn or check out beehiiv on their website.

BRAIN FOOD 🧠 

TWEETS OF THE WEEK 🐣 

Source.

Source.

Source.

TOOLS WE RECOMMEND 🛠️

Every week, we highlight tools we like and those we actually use inside our business and give them an honest review. Today, we are highlighting BILL—Your AI-powered financial operations platform.

beehiiv: We use beehiiv to send all of our newsletters. Apollo: We use Apollo to automate a large part of our 1.2M weekly outbound emails. Taplio: We use Taplio to grow and manage my online presence.

See the full set of tools we use inside Athyna & Open Source CEO here.

HOW I CAN HELP 🥳

P.S. Want to work together?

1. Hiring global talent: If you’re hiring tech, business or ops talent and want to do it 80% less, check out my startup, Athyna. 🌏

2. See my tech stack: Find our suite of tools & resources for both this newsletter and Athyna here. 🧰 

3. Reach an audience of tech leaders: Advertise with us if you want to get in front of founders, investors and leaders in tech. 👀 

That’s it from me. See you next week, Doc 🫡  P.P.S. Let’s connect on LinkedIn and Twitter.

👋 Howdy to the 5,451 new legends who joined since our last edition! You are now part of a 497,715-strong tribe outperforming the competition together.

LATEST POSTS 📚

If you’re new, not yet a subscriber, or just plain missed it, here are some of our recent editions.

✍🏻 Building The Incorruptible Companies Of The Future. An interview with Eric Ries, Author of Incorruptible.
🐝 When David Ate Goliath: beehiiv’s Zero To One. The scaleup with attitude that’s killing the competition.
🌀 Content Is The Funnel, Community Is The Business. An interview with Sid Yadav, Co-Founder & CEO at Circle.

PARTNERS 💫

79% of enterprise buyers say generic AI voices damage brand perception—and that risk starts long before recording begins. It's in how the voice is sourced, licensed, and built.

Voices' free guide breaks down the 5 AI voice myths costing CX platforms their biggest deals. Get it before your next enterprise conversation. 

Framer is the pro website builder for creators, teams, and businesses that care enough to get every detail right. Now with Agents that work alongside you across the full website workflow—all without leaving the tool where the real site lives.

Agents bring speed and scale; you bring taste, judgment, and control.

_{Interested in sponsoring these emails? See our partnership options here.}

HEAT CHECK 🌡️

Apple just sued OpenAI in federal court, claiming a coordinated scheme to nick its hardware secrets. The suit says one ex-Apple engineer walked out with a laptop full of confidential docs, and that OpenAI's hardware boss (ex-Apple himself) coached candidates to spill secrets in interviews. OpenAI denies the lot.

Apple's camp says this is textbook trade secret theft: 400+ former Apple folks now sit in OpenAI's hardware team, and the alleged behavior went well past normal job-hopping. The other camp believes a company that missed AI is using lawyers to slow a rival's device push. Big precedent either way.

GUEST POST 💄

Building A Second Brain With Claude Cowork

I’m not sure if you feel the same way, but over the last 12-18 months, AI went through a real transition: from something that was fun, to something you saw potential in, to something useful, and finally, to something that, on the right occasion, will totally blow your mind.

One of the things white collar workers and technology advocates across the globe have yearned for for years is the idea of a true ‘second brain.’ A place with all of your context that allows you to be the best version of yourself. From a personal dump file, to your health and wellness coach, to the calendar for your dog’s next vet checkup. But it’s always been a very underwhelming process because the right tool for the job hasn’t really existed, until now.

Today, I am enlisting my buddy Eliot Prince to walk us through what the ultimate second brain looks like in the age of AI. Eliot runs one of the best YouTube channels for AI upskilling online and posts his thoughts in his newsletter, too. He’s the perfect person to upskill us all today. Without further ado, here is my guy, Eliot.

It was all a fugazi

Hey, ladies and gents, Eliot here! If the Guinness Book of World Records had a ‘how fast can you forget an idea’ category, I would own that title by some distance. Even a goldfish couldn’t wrestle the record away from me. I have an awful memory. Not the cute kind where I forget to pick up tomatoes for my fiancée (even when it’s the only reason I’m going to the store.) It’s the kind that’s frustrated me for a decade+ as an entrepreneur, creator, and writer. The lightning strike idea in the shower, a quote in a book that stops me dead, or a podcast line that reframed my whole mindset. Poof, Fugazi, it’s fairy dust, gone in seconds.

I’ve tried to do the diligent thing, too. Highlight the book quote, snip the podcast, and save the URLs. But here’s the cruel thing about my brain… You can only go digging for any of it if you remember it exists. I don’t just forget the idea or quote; I forget it ever existed in the first place. So in 2019, I set out on the big fix. I’d heard about something called ‘Second Brain.’ A digital mecca that catches every fleeting idea and hands it back to you exactly when you need it. That led me down a rabbit hole of 100+ hours of Notion. I bought templates, courses, and built the PERFECT architecture. It lasted two weeks.

The maintenance tax was brutal

I’d spend two hours a day keeping things tidy before I even got anything done. It felt like cooking a gourmet meal every evening, but being too tired to eat it. Everything collapsed. Over the next few years, I tried Obsidian, ClickUp, Claude connectors, and even vibe coded my own software. All sinking ships. I was just rearranging deck chairs on the Titanic.

But over the last two months, I’ve realized something: I’ve stopped yearning for a second brain. And when I thought about it, it’s because I’d quietly built the whole thing underneath Claude Cowork. All those false starts and lessons… I’d built them into how I use Cowork daily. Now I open my laptop, and it’s all there, ready to work. Every idea, every quote, running commentary of my days, transcripts from calls. And more importantly, the right information is presented to me exactly when I need it.

So if you’re here with a head full of forgotten ideas and a graveyard of note-taking apps, then I want to hand you my Claude Cowork Second Brain setup. This is your chance to wipe the slate clean and reboot your digital life. You can set the whole thing up in 1 hour. And enjoy an information dumping ground, trusting that it’ll come back to you when you truly need it. Let’s go.

1/ Open Cowork > to create your ‘Second Brain Project’

First thing, download and install the Claude desktop app. You can't do this in Claude Chat in the browser; there isn't enough power under the hood. Open the app, click into Projects, and create a New Project. Start from scratch. Call it ‘Second Brain’. Leave the custom instructions blank; we’ll build that later. This is the only project you'll ever need to build, because your whole world is going to live inside it. And here's the bit I love. You're not going to build the structure by hand (we’re AI architects after all). Claude is going to build it for you.

2/ Understand the PARA System

Normal filing systems are cruddy. You end up with giant department folders, then folders inside folders inside folders. It’s folderception. You spend more time hunting for info than doing the work. So today we flip it. We use an organizational system called PARA, from a book by Tiago Forte. Here’s the TLDR: PARA uses four folders.

• Projects: Live work that's time-bound and has a finish line. A product launch, a fundraise, a client deliverable, a talk you're writing for next month.

• Areas: The ongoing parts of your work and life that never really ‘finish’. Think hiring, finances, house management, marketing. They don't have a deadline; they just keep rolling, and you're responsible for them month after month.

• Resources: Knowledge you know you'll want one day that doesn't belong to any current project or area. For me, that's a folder of business growth advice I'm slowly collecting. For you, it might be business frameworks, leadership resources, fundraising notes, anything.

• Archive: Anything that's gone quiet. Finished projects, dormant areas. Out of the way but not deleted.

That's it. Four buckets. Everything you own has one obvious home.

Now imagine you give that power to Claude Cowork too!?

3/ Give Cowork the power of PARA

Next, hand Claude the job of setting up your PARA folder structure. Drop this prompt into your Second Brain project:

Prompt:

You're setting up my second brain in this project. Create a PARA folder structure for me. Make four top-level folders: Projects (live, time-bound work with a deadline), Areas (ongoing responsibilities with no end date), Resources (reference material I'll want later but that doesn't fit a current project or area), and Archive (anything finished or dormant). Leave them empty for now. When you're done, show me the structure so I can check it.

Give it a few seconds, and your Second Brain folder structure on your computer will look like this. Remember, Cowork runs using files and folders on your computer!

4/ Add an Inbox so capturing never breaks your flow

The single biggest reason my old systems died was capture. An idea hits while you're mid-task, and to save it properly, you have to stop, decide which folder, which note, which database, which Drive. By the time you've filed it, you've lost the thread of what you were doing. So you stop filing, the system rots, and you end up with a million different browser tabs open in your brain. The fix is an Inbox. One folder that sits right next to Projects, Areas, Resources, and Archive, and acts as your dumping ground. Capture first, organize later. You don't decide anything when you capture. You just dump. Sorting happens later, either by you or, as you'll see in a minute, by Claude.

Prompt:

Add one more folder at the same level as Projects, Areas, Resources and Archive. Call it Inbox. This is my capture folder. Anything I drop in here is unsorted and temporary, and gets filed into the right PARA folder later. Don't put anything in it yet.

Editor's note: A deserved shout-out

Hey, it’s Doc again here. I just wanted to quickly make a note of Eliot’s channel in a bit of a broader way. I stumbled upon Eliot, and we started chatting because his channel was the main channel I would share during our fortnight AI Watch Party at Athyna, my startup.

Eliot is too humble to try to shoehorn his own videos into this step-by-step style piece, so I thought I’d better go ahead and do it for him. Here is but a small sliver of his best stuff. Onwards!

5/ Write the behavior rules with CLAUDE.md files

You've got the architecture. Now you tell Cowork how to behave inside it. This is where CLAUDE.md files come in, and they're the part most people have never heard of.

A CLAUDE.md is a plain-text instructions file you put inside a folder, and Claude reads it every time it runs in that folder. Think of it as a system prompt that lives with the folder. Every folder in your second brain can have its own, so each part of your world behaves exactly how you want it to. You can ask Cowork to write these from scratch, or use the framework I've set out in the prompt below, which gives you a sensible setup straight away.

Prompt:

Write CLAUDE.md behaviour files for my second brain. The root CLAUDE.md should explain PARA (every file lives in one of Projects, Areas, Resources or Archive, with Inbox as the staging area for anything unsorted), tell you to start each session by asking me as a multiple-choice question whether it's Inbox capture, a Project, or an Area (for a Project or Area, ask which, scope in, and stay there unless I say otherwise; for Inbox capture, save everything I send to the Inbox as dated notes and do nothing else), make captures always hit the Inbox first rather than a deep folder mid-task, and put anything you create in an Outputs folder inside the current Project or Area, creating it if needed. Then add a short CLAUDE.md in Projects, Areas and Resources saying what belongs there and how to name files, and if one's empty, help me decide its starting knowledge base. During long sessions or after critical outputs you should write/update a memory.md file to keep a project or area upto date. 

When Claude finishes, you'll have CLAUDE.md files sitting through your folder structure.

The root file teaches Claude the whole filing system and tells it to ask, every time you log in, what you're working on and where to scope in. Plus any other house rules. The folder-level files give each area its own rules, so your Finances folder can behave differently from your Marketing one if you want.

Two ways to do this, and the right one depends on what you've already got organized. If you've already got knowledge sitting in other places, old notes, a swipe file, or docs from past projects, seed it. For example, my projects are filled with client and internal projects that are live. Each has its own folder holding the key knowledge and outputs. Either add these folders and files yourself. Or ask Cowork to help!

If you're starting cold, build as you go. Every call transcript, every idea, every useful article goes into the Inbox, and over a few weeks, the thing fills itself.

7/ Automate the filing with an inbox-processing skill

Last build step, and it's the one that makes the whole thing run itself. We're going to give Claude a repeatable routine for sorting your Inbox and filing everything into the right place.

Prompt:

Using the /skill-creator, build me an inbox-processing skill I can run any time I say "process my inbox". Go through every file in the Inbox folder one at a time. For each item, read it, work out which Projects, Areas or Resources folder it belongs in based on what it's about, then suggest where it should go and a clean filename. Once I approve, move it there. If something has no obvious home, leave it in the Inbox and tell me why. Never delete anything. Give me a short summary at the end of what moved where.

Now your loop is closed. Capture into the Inbox whenever, however messily you like, and let Claude file it later. You can run it on demand by typing ‘process my inbox’, or set it to run on a scheduled task.

💡 Note: 65,000+ professionals and founders every week read and watch Eliot every week. Get access to 200+ prompts at his AI Recipe Vault.

Wrapping things up (and daily use)

Once it's built, the daily rhythm is boringly simple. You open Cowork, say ‘Let's work,’ and Claude asks one thing: are you here to capture, or to work on a Project or Area? Pick capture and just dump a quote, a link, a half-formed idea between meetings. It all lands in your Inbox without breaking your stride. Pick a Project or Area and Claude scopes into that folder, pulling you in with all the context already loaded. No hunting.

When the Inbox has built up over a day or week, you say "process my inbox" and let Claude file everything away, or leave the scheduled run to do it for you. Once you've lived in it for a few weeks, you start extending it. You can wire in your call transcripts so every meeting lands in the inbox automatically. Point it at your email to pull out anything worth keeping. You can connect it to your CRM, so the client context is always at hand. The architecture stays the same, you're just feeding it more and more of your world.

Extra reading

• We Raised $2.5M To Build Athyna AI - May, 2024

• If You Ain’t AI First, You’re Last - November, 2025

• Treating Candidates Well, Post-Training & AI Recruitment - June, 2026

And that's it! You can follow Eliot on LinkedIn, and also don’t forget to check out his website while you’re at it.

BRAIN FOOD 🧠 

TWEETS OF THE WEEK 🐣 

Source.

Source.

Source.

TOOLS WE RECOMMEND 🛠️

Every week, we highlight tools we like and those we actually use inside our business and give them an honest review. Today, we are highlighting Granola*—an AI-powered notepad that takes, summarizes, and organizes meeting notes without using intrusive recording bots.

See the full set of tools we use inside Athyna & Open Source CEO here.

HOW I CAN HELP 🥳

P.S. Want to work together?

1. Hiring global talent: If you’re hiring tech, business or ops talent and want to do it 80% less, check out my startup, Athyna. 🌏

2. See my tech stack: Find our suite of tools & resources for both this newsletter and Athyna here. 🧰 

3. Reach an audience of tech leaders: Advertise with us if you want to get in front of founders, investors and leaders in tech. 👀 

That’s it from me. See you next week, Doc 🫡  P.P.S. Let’s connect on LinkedIn and Twitter.

👋 Howdy to the 5,127 new legends who joined since our last edition! You are now part of a 492,264-strong tribe outperforming the competition together.

LATEST POSTS 📚

If you’re new, not yet a subscriber, or just plain missed it, here are some of our recent editions.

🐝 When David Ate Goliath: beehiiv’s Zero To One. The scaleup with attitude that’s killing the competition.
🌀 Content Is The Funnel, Community Is The Business. An interview with Sid Yadav, Co-Founder & CEO at Circle.
📐 The Lab Builders Are Leaving. An interview with Andrew Dai, Co-Founder & CEO at Elorian AI.

PARTNERS 💫

This newsletter runs on beehiiv. So when we say the referral program works, it's not a guess, it's our actual growth engine. Readers share, hit milestones, earn rewards, and the list grows on autopilot. Pair that with their recommendation network and newsletters on beehiiv grow 2.75x faster. We're one of the data points. Worth a look if you're building something.

Get 30% off your first three months on beehiiv with the code: BILL30

Smart hiring for teams that need strong talent now.

• Fill key roles faster, without waiting weeks or months for the pipeline to behave.

• AI-assisted matching and structured vetting help you find people who fit the role and the team.

• Cut time and cost with a process built for scale, not endless back-and-forth.

_{Interested in sponsoring these emails? See our partnership options here.}

HOUSEKEEPING 📨

It’s pretty crazy that Australia has three of the top ten cities in the world, according to The Economist. Nevertheless, in Australia, all you hear in certain circles is how poorly the country is doing today. Granted, in the 2013 version of The Economist’s poll, there were four cities on the list, but I still feel so incredibly lucky to be an Australian, and in particular, a Melburnian.

I feel like there is a general sense of dissatisfaction across the globe today, stemming from the cost of living driven by COVID-era money printing, but even so, I think we are in a pretty good place in 2026. I’m bullish on closing out the decade in a blaze of innovation, growth, and, hopefully, better vibes.

INTERVIEW 🎙️

Eric Ries, Author of Incorruptible

Eric Ries is the creator of the Lean Startup method and the author of The Lean Startup, the 2011 book that sold more than a million copies, was translated into 30-plus languages, and gave a generation of founders its core vocabulary—the MVP, the build-measure-learn loop, the pivot. He later built the Long-Term Stock Exchange (LTSE), an idea he first floated in that book's epilogue and then spent close to a decade turning into a fully SEC-approved national securities exchange, the first new one cleared to list companies in over 50 years (Twilio and Asana were the first to sign on). His latest book, Incorruptible: Why Good Companies Go Bad... and How Great Companies Stay Great, arrived as an instant New York Times bestseller.

Eric.

What makes Eric worth reading right now is where his thinking has gone. The man who spent two decades teaching founders how to build is now fixated on why so many of them lose what they built, ejected, bought out, or slowly drained of everything that made the company worth starting. He calls the culprit ‘financial gravity’, the pull that bends even great companies toward short-term thinking and quiet mission abandonment, and he argues the fix is structural rather than ethical. Governance you design early and on purpose to defend the things that score as ROI-negative but matter most. He's not theorizing from the sidelines, either. He advised Anthropic's founders when they left OpenAI, pointing to their Long-Term Benefit Trust as the idea working in the wild, and keeps a file of founders whose own names are stuck on products they no longer control. His through-line lands at exactly the wrong moment for anyone mid-build. It's always too early to protect your company, right up until it's suddenly too late.

What is the core of the book? And why do good companies go bad in the name of profit?

The core idea is really simple. Most founders lose control of their company involuntarily. Either they get kicked out, or the company's culture and structure drift from what they had in mind, and eventually it becomes an out-of-control monster. The book is about how to stop that from happening.

The book is structured like a double mystery. The first mystery is why companies are so often destroyed in the name of profit. Think about your favorite brand that got taken over by private equity, a company that went public and is now a shadow of itself, or a place where the founder was ejected.

In the short term, they made more money, but in the long run, the whole thing collapsed. This has been going on in our so-called capitalist system for centuries. In the book, I document story after story like those. And it's so strange. I thought capitalism was about perfect competition, about competing to create the most value for customers. Yet we've built a modern economy that's full of ways to make money without creating any value at all, or even by destroying it.

The second mystery is that if such a decline is inevitable, why are there exceptions? Because there are. The same person who'd tell you the process is inevitable in one conversation will be raving about how much they love Patagonia, Costco, or some other remarkable company in another. Start noticing which names come up again and again, like Vanguard, Costco, Patagonia, Novo Nordisk, the John Lewis Partnership, Hershey, Tony's Chocolonely, and lately Anthropic. As a set, these companies violate almost every one of our modern best practices for building, structuring, and governing businesses.

Source.

That's a fascinating clue that something's gone rotten. Why are we preaching best practices that the very best companies—the ones with real longevity, the ones that stay true to their mission and still make an awful lot of money for investors—don't actually follow? That magic combination is what we all claim to be chasing. So I started tugging on the thread, and in researching the book, I found plenty of evidence that many of our modern best practices are, in fact, value-destroying. We've materialized a set of practices that leaves companies weak and vulnerable to the financial gravity that drags them down into mediocrity or worse.

What does Patagonia understand that most companies don't?

Patagonia's such a great story because it shows a lot of the concepts in miniature, and it's so many people's favorite company. They're worth billions now, so it must be a result of incredible management and flawless execution, but not really. For many years, the company was extremely poorly run. Yvon Chouinard used to disappear on long mountaineering and climbing expeditions and return after months of total neglect. There's one story where he returns to the office and sees, in the bin next to his Head of Sales, a giant invoice for a huge order they'd sold. He fishes it out, all crumpled up, and asks if they really got an order for this and just threw it in the trash. And the guy says, “Yeah, you weren't here. It seemed like a lot of hassle to fulfill such a big order, so we threw it away.” There are loads of stories like that. The company famously overexpanded, outsourced too much production, had to lay off people, and nearly died. Plenty of reasons it shouldn't exist.

The reason it does is a testament to the power of mission. The mission is so strong it's become a magnet for talent and for customers. They once ran a catalog whose front page said, 'Everything we make pollutes.' Rather than dodge the environmental cost of what they do, they put it right up front. Before they were into environmental activism, they were obsessed with product quality. They told people not to buy extras, but to buy one thing and have it last. They'll still repair any of their gear for free, because they believe things should last forever.

Chouinard used to teach his employees classes on things like the nature of quality itself. He'd put an object on the table, like a bag of coffee beans, and ask whether it was a quality product or not. The poor students would say it depends on whether you like coffee, but he would correct them, saying quality is an objective function. Something either is or isn't the highest quality it could be. This is a man who hand-forged his own pitons and carabiners because he couldn't find ones good enough. I know this all sounds very la-di-da. Great, they have values, but what about the business? The thing is that this is the secret superpower of the best companies in the world. When you magnetize people, everything costs less. The best people want to work for you, customers are supremely loyal and far more forgiving of your mistakes, they'll try your new products, and your brand is worth more. I call the principle 'harder is easier'. You do the harder work up front to commit to your values, and you reap the benefits that make business easier down the road.

Source.

The story that captures it is when the company hit a rough patch and had to restructure a pile of debt with its banks after Chouinard had stepped back as CEO. They agreed on terms, but at the very last moment, the lead bank asked Patagonia to suspend its nonprofit donations program. Honestly, maybe the most reasonable request a bank has ever made. However, the CEO just says, “Sorry, no deal.” The point of this exercise was to save Patagonia, not to destroy it. It was such a power move.

They do what they do, and they’ll never compromise. If that means they go out of business, fine. That's a level of internal consistency most companies can only dream of.

How do you make something like mission and culture actually stick?

This is one of the book's biggest preoccupations, because I didn't want to write another manifesto about how great trust and mission are. We have enough of those. What we need is a blueprint. A step-by-step guide for building organizations that can be trusted, that can make and keep promises. It took me a whole book to answer this properly, so forgive the compression. Take OKRs, for example. Almost every company runs on them now. You take a grand goal, break it into sub-goals, break those into sub-metrics, and everyone ends up with a metric they have to move. Sounds logical. But ask people how they'll pursue their goal, and they'll say, “I'll take every possible thing I could do, stack-rank by ROI, and do the highest-ROI things that hit my number.”

Here's the problem. The things that make us most trustworthy—to partners, customers, employees, investors—almost always score as ROI-negative, because the benefits are intangible and the costs are tangible. Andrew Mason told me a great story about Groupon. They went public on a dead-simple model; one email a day with a great deal. After the IPO, people kept coming into his office, insisting that two emails a day were better than one and that they’d make more money.

Source.

He’d refuse, but they wore him down with the idea of experimenting, so he relents. They run two a day, and metrics and OKRs go up. Before long, it's three emails a day, and on it goes. You see that slow compromise everywhere.

Trustworthiness is the most underrated and valuable asset in all of business. It's a genuine financial asset, in some cases worth more than money, because money is something you spend and is gone, but trustworthiness, if you protect it, endures. People stay loyal to brands for decades, even across generations, even in a cynical age. So once you see trustworthiness as an asset, you realize that when you hand an individual an OKR, they can quietly spend that asset to boost their personal number, at the expense of the company's balance sheet. Imagine a business with no financial controls, where anyone could spend money without anyone knowing. You'd be bankrupt in no time. Most companies end up morally bankrupt for exactly that reason. They have no mechanism to preserve this asset. The book is about how you build the vaults to measure, protect, and defend them against the people who'd squander them.

Why did you want to write this book?

Honestly, I'm sick and tired of this phenomenon. I've helped too many people make too much money only to end up miserable. If the company becomes a disaster, I don't see who it's for. I know so many people whose kids and grandkids will be ashamed of how the family money was made. They've become the villains of their own story, and in so many cases, they have to watch from the sidelines as the thing they created turns into a malignant monster.

I've started keeping a file on founders whose own names are also the name of their product or company, who then get ousted and have to watch someone else degrade the quality of the thing that bears their name. I even know of a case where the founder is legally barred from starting a competing company to fix the problems, because his name is trademarked by the original company and is too similar to theirs. It's absurd. I've watched this up close. I've seen so many founders lose control and get absolutely massacred by the buzzsaw of the financial system, because they come in naive. They don't take enough steps to protect themselves or their vision, and they reach out for help when it's far too late. One of the key ideas in the book is that it's always too early until it's too late. So I'm trying to get people to take this seriously from as early as possible.

What's the good news story about Anthropic?

The situation with AI is complicated, and I think we're simultaneously too optimistic and too pessimistic about the technology. Meanwhile, there are people actively trying to burn the world to the ground, and I think we can save some of the criticism for them. People always ask me what the problem is with this or that. Even during the GameStop mania, journalists kept calling to ask why people were so upset. People get upset whenever someone finds a way to make money without creating value. Turn the economy into a casino, and yes, you can make money, but people get angry. We have an intense, intuitive aversion to it, and yet we lack the modern language to talk about it. One of the breakthroughs for me was realizing our grandparents had no trouble talking about it. They'd have called these corrupt ways of making money. That's what the word corruption used to mean. We've forgotten it.

I happened to meet Anthropic's founders when they left OpenAI. They came to me for advice on exactly these topics, on an investor's suggestion. I want to be super clear that I'm not taking credit for their success. I played a tiny bit of a role. But ask people why Anthropic is winning, and they'll say lower inference costs, more focus, the best of this or that. Let’s play five whys.

Source.

Why the lowest inference costs? Best technical architecture. Why? Best talent. Why? The best talent wants to work there. Why? Because people want to feel like they're the good guys. Why do they believe that? Because the culture is genuinely aligned with that ethos. Keep going, and you’ll land where many of the companies we've discussed land. A magic combination of two things: they have the ethos, the culture, the mission that the best people find intensely attractive, and they also have the integrity and structure to resist being pushed off their values.

There's been very high-profile bullying of Anthropic by enormously powerful people. Part of why they've kept their nerve under that pressure is that they saw it coming and built a structure to protect themselves. The Long-Term Benefit Trust is an external set of trustees with the power to appoint directors to the PBC's for-profit board. They've used some of the techniques I recommend in the book, and I think that's part of why they seem a bit more courageous than their competitors.

Source.

I'll never forget the night Anthropic said no to the $200M contract and OpenAI said yes. Someone sent me a video of the sidewalk around Anthropic's headquarters, which had been chalked up. Normally, if I told you activists chalked a San Francisco tech company's offices, you'd brace for something grim, but people were thanking them for doing the right thing. Claude went to number one on the App Store the next day. They reaped so many counterintuitive benefits for doing the 'right thing' that some people started criticizing them for it, saying they're only virtue-signaling to get the benefits.

That's the funny thing about mission-driven companies. The magnetism is so powerful that people can't quite believe it. When I started the book, Anthropic wasn't well known; I had to revise that chapter a hundred times as they kept having stratospheric success. But the pattern is common. Show people a young company, and they’ll say, “Never heard of them, how successful can they really be?” Show them a wildly successful one, and they’ll say, “That's not realistic.” Is there any company whose success you'd accept?

With these magnetic powers, you're either nobody or you're insanely successful. That's the point. So the skepticism is the tribute vice pays to virtue. In trying to stay skeptical, people are quietly admitting something powerful is going on. We keep pretending mission is some soft-hearted thing you trade off against profit, when it's one of the most powerful sources of competitive advantage in the world.

Has your thinking changed since The Lean Startup?

I know a lot more than I did then. Some of the original book was a bit simplistic, and tactics are aged. The tools we have now for building an MVP and measuring what's happening are so much better than before, but the principles hold. The whole point of that book was that if we reason from principles about what can and can't work—if we think about entrepreneurship more rigorously—we can work out what to do in each era of technological disruption.

I made a claim at the time about the semiconductor revolution that its long-run effect would be two things. First, more and more democratization of the means of production and faster cycle times, letting you enter more markets more quickly. So acceleration is one major effect. Second, increased uncertainty and unpredictability about what will work, and a decreased ability to forecast the future in any detail. Those two pillars are really what The Lean Startup is for. When you find yourself in a fast-moving, highly uncertain situation, The Lean Startup is what you want. I think that prediction has held up pretty well.

What is the Long-Term Stock Exchange, and why does it exist?

My first attempt to dent this problem wasn't the book; it was the many years I spent building the Long-Term Stock Exchange, the LTSE. I actually floated the idea in the epilogue of The Lean Startup. I try to end all my books with a final chapter that's a bit different from what came before. What struck me as a management theorist is that we now have a real body of evidence on which management techniques work, and one of the strongest findings is that a philosophy of long-term thinking leads companies to outperform over a generational timescale.

People used to ask me why I was telling them to build venture-backed companies and take them public, when the public markets are the most short-term environment in the world. I didn't know anything about public markets at the time—this was 15, 20 years ago—but it just seemed off. Why would investors want companies to be less valuable? You'd think they'd want them more valuable. So I had this idea of building an institution that brings together managers and investors who share a long-term philosophy, so everyone makes more money. That was just a stock exchange, and someone should build a long-term one. When nobody would, I decided to do it myself.

I don't run it day-to-day anymore, so I have to break myself of saying 'we.' But the company got SEC approval and has listed and traded multiple stocks (the first to do that since the NASDAQ was created more than 50 years ago). You may have seen recent headlines suggesting the SEC looks set to move companies from quarterly to semi-annual reporting. That's the result of a petition LTSE filed last September. We don't get much credit for it, the way the press is covering it, because it's become a political food fight. Still, the company's mission is to make public markets more conducive to long-term thinking. A sanctuary for companies that want access to public markets without all the associated drama and nonsense. It's been a very difficult project, but if you have the patience, you can carve the Grand Canyon out of water and gravity.

How do you get the best out of yourself?

Honestly, I struggle with this because I'm always doing so much, and I tend to say yes to things. The upside is you get to do a lot of interesting stuff; the downside is you run yourself ragged. I have a robust family life, and I'm incredibly blessed. I draw so much energy and inspiration from my wife and kids, from that family support. People these days like to mock the idea of work-life balance, but to me, it's the foundation for achieving anything worthwhile. I'd call it work-life integration, and I've strived for it for a long time.

On techniques, it’s meditation for me. If you can't achieve equanimity, you're probably not going to succeed at much else, and entrepreneurship is inherently anxiety-provoking. You're constantly confronting failure and the fear of it, getting publicly pilloried, and you don't get the dopamine hit of success the way you do in other kinds of work. So a spiritual practice, the ability to seek equanimity and curate your inner life, is critical.

A lot of people get into entrepreneurship with demons chasing them, thinking worldly success will solve an inner problem. Most of the time, it really doesn't. I raise it because entrepreneurship is one of the great truth-seeking disciplines, right up there with being an artist or a scientist. At the end of the day, reality either will or won't accommodate your vision. To do truly great work, to create real impact, to build something with generational returns, to see your vision realized, you need the freedom to choose. Imagine trying to do a startup in a country where someone else tells you what you're allowed to build. We're drifting in that direction right now, which is genuinely a little scary, but we're not there yet. Externally, we still have the freedom to do what we think is right.

My real question is whether we have inner freedom. I think a lot of people don't. They don't have the self-mastery or the equanimity to take in the truth of their situation and do what it requires. Often, the reason they can't find success is an inner belief that stops them from doing the right thing. That's part of the power of this book. Readers have already written to tell me it changed how they see business, and what it means to be a for-profit company, and that as a result, they're pursuing an idea they'd never have had the courage to chase before. It was staring them in the face, and only the inner shift let them see it. I'm pretty excited about that.

Extra reading

• Why Companies Go Bad - June, 2026

• Why Corporate Corruption Is so Common - May, 2026

• Incorruptible: Why Good Companies Go Bad... and How Great Companies Stay Great

And that's it! You can follow Eric on LinkedIn and X, and check out his The Lean Startup to keep up with what he’s doing.

BRAIN FOOD 🧠 

TWEETS OF THE WEEK 🐣 

Source.

Source.

Source.

TOOLS WE RECOMMEND 🛠️

Every week, we highlight tools we like and those we actually use inside our business and give them an honest review. Today, we are highlighting Mochary Method—an executive coaching firm that specializes in helping growing startups.

beehiiv: We use beehiiv to send all of our newsletters. Apollo: We use Apollo to automate a large part of our 1.2M weekly outbound emails. Taplio: We use Taplio to grow and manage my online presence.

See the full set of tools we use inside Athyna & Open Source CEO here.

HOW I CAN HELP 🥳

P.S. Want to work together?

1. Hiring global talent: If you’re hiring tech, business or ops talent and want to do it 80% less, check out my startup, Athyna. 🌏

2. See my tech stack: Find our suite of tools & resources for both this newsletter and Athyna here. 🧰 

3. Reach an audience of tech leaders: Advertise with us if you want to get in front of founders, investors and leaders in tech. 👀 

That’s it from me. See you next week, Doc 🫡  P.P.S. Let’s connect on LinkedIn and Twitter.

👋 Howdy to the 4,952 new legends who joined since our last edition! You are now part of a 487,137-strong tribe outperforming the competition together.

LATEST POSTS 📚

If you’re new, not yet a subscriber, or just plain missed it, here are some of our recent editions.

🌀 Content Is The Funnel, Community Is The Business. An interview with Sid Yadav, Co-Founder & CEO at Circle.
📐 The Lab Builders Are Leaving. An interview with Andrew Dai, Co-Founder & CEO at Elorian AI.
🧗🏻 Unfiltered: The System Vanta Built To Grade Its Own AI. How they build language models as the judge, human as the jury, and more.

PARTNERS 💫

Summer weekends stack up fast: BBQs, lake days, long dinners, and late nights. ZBiotics Pre-Alcohol is the world’s first genetically engineered probiotic, developed by PhD scientists to help break down a byproduct of alcohol linked to next-day sluggishness.

Drink one bottle before your first drink and go about your night as usual. A simple habit to help protect your next morning. Get 15% off with code OPENSOURCECEO.

Founding a company is easy; scaling it is hard. Operational systems break, people problems increase, and your calendar keeps getting fuller. 

Mochary Method is an executive coaching firm founded by Matt Mochary—coach of Sam Altman and Brian Armstrong, and writer of The Great CEO Within—that specializes in helping growing startups.

The Mochary coaching team, who coaches companies like AngelList, Lovable, and Midjourney, is offering a complimentary coaching session to founders and executives for a limited time.

_{Interested in sponsoring these emails? See our partnership options here.}

HOUSEKEEPING 📨

My partner won a dog photoshoot contest a few weeks ago. It was a $2,000 voucher from an agency that shoots your images for free, then only gives them to you in a framed package for $5-20k. I think the voucher was actually a sneaky lead magnet trick, to be fair. But we did manage to sneak this epic photo of the two of us with our little man, Ziggy.

Nothing more to add here, really. Just sharing a little snippet of my life with you all. I love this photo. Anywho, today’s piece is a great one, with beehiiv, a company that I really love. I hope you enjoy reading it!

ZERO TO ONE 🌱

When David Ate Goliath: beehiiv’s Zero To One

The word ‘velocity’ /vɪˈlɒsɪti/ originates from the Latin noun vēlōcitās, which means swiftness or speed, itself derived from the Latin adjective vēlōx, meaning fast or rapid. Although I have enough Italian heritage to claim I am a son of Ancient Rome, I don’t speak Latin and couldn’t point out the Tiber River on a map. Being a tech nerd, when I think of velocity, I think of product velocity. And when I think of this particular flavor of startup speed, I think of one company: beehiiv.

Along with a dash of customer obsession and a sprinkle of can-do attitude, product velocity has allowed beehiiv to enter a wholly owned market, disrupt it, and take a whole whack of market share along the way. As a user of the product, it’s been a total joy to watch.

Latin hivelords.

While they have crushed it in their short existence, it’s not been easy for beehiiv. They’ve battled incumbents, waged war on their upstart competition, and dealt with shocking loss and heartache along the way.

Today, I’m here to tell their story. In researching this piece, I sat down with beehiiv’s CEO, Tyler Denk, along with Chief Customer Officer Preeya Goenka, and senior product OG Shreya Srikanth. What I unearthed was very cool, so let’s jump right into it!

The Morning Brew cheat code

Some startup teams are Harvard grads, megamind types, that are just itching to find the right problem. The types that have so much intellectual horsepower they could power the streetlights of a small city. Strong teams in search of a worthy problem can win, but it’s often a better play to solve a pain you yourself personally feel. This is the story of beehiiv’s founding.

Tyler, along with co-founders Benjamin Hargett and Jake Hurd, had worked inside the successful business news for college grads-cum grown-up people newsletter, Morning Brew. You might know it now as a multimedia network of newsletters, podcasts, and YouTube channels, but it started as a humble email rag. A humble rag that would grow to 4M subscribers and subsequently be acquired by Business Insider (owned by Axel Springer) for a cool $75 millioneys.

Source (ish).

As you can imagine, this experience paved the way for beehiiv, because they had felt (and built) things from the inside at The Brew: "We were so deep in the weeds of what we needed to build, as we'd already built this all internally at Morning Brew, as a case study, and it worked extremely well."

Tyler explained to me that he and his co-founders sat right at the intersection of content and sales. To their left, among the creative-content hippies, they saw firsthand what it took to be a world-class publisher. To their right, they sat close enough to the uber-capitalist sales team that they also knew the language they spoke, and really understood what it took to close high-value advertisers.

Most of the incumbents were typical ESPs—email service providers—such as MailChimp, Constant Contact, and MailerLite. All these tools really did for the most part was send emails. Sure, it’s the most important part of the job, but what about the website design, newsletter growth tools, and monetization?

Full circle moment (Source).

Some of this existed on other platforms; the rest, Tyler and the team coded up manually themselves. They quite literally built beehiiv before building beehiiv.

And, Tyler told me, the experience was altogether just frustrating: “They were also expensive, and a lot of them had gone through acquisitions and weren't innovating the way hungry startups do when they're chasing market share.” To them, there seemed to be a market screaming for a consolidated, creator-first company to be born.

Ship or die

One of the most fascinating things to me about companies like beehiiv is the general chutzpah it takes to launch in a market dominated by huge incumbents. I can’t imagine going into sales conversations knowing that what we can do is likely only 10% of what the platform they are currently using can do.

_{*Note: Let’s switch to verbatim for a minute as I think it will play better.}

Bill (me): “How did you win over noteworthy publishers in the early days, given how feature-deficient you were compared to the incumbents?”

Tyler: ”The first big customer we had was Litquidity, the meme account. Getting him as a user was actually pretty straightforward, as he'd invested in the company.

That was the shortcut in the early days. I went through all of our initial angel investors, and a number of them had newsletters. There was a guy named Gannon Breslin who had a newsletter of about 25,000 people. Not a huge name, but we just needed traction; we needed people to use it and test it, and then we could build case studies off it.”

Bill: ”And what came after that?”

Tyler: ”Well, we raised about 25 angels, and I'd say 20 of them moved their newsletters over within a month of us launching. Litquidity was definitely the biggest one; he had 100,000 subscribers on MailerLite. We couldn't do more than 10% of what MailerLite does at that point. He just totally trusted us, and we figured it out. And then it's a game of snowballing success stories. Now that we had Litquidity on the platform, we could turn that into a case study for all the other publishers and newsletters who looked up to him. The narrative became: this is the founding team behind Morning Brew's internal tech. We scaled that to four million readers. We already have Litquidity with 100,000 subscribers, and its open rates had improved since moving to us.”

Me again. Tyler told me that although they had the small win with the investor cohort, these times weren’t exactly easy: "I literally had a tenth of our revenue threatening to churn in my inbox every single morning." The only way for Tyler and team to combat this was to move fast: ship or die mode.

Bill (me again): “If you had to put your finger on one thing, what was the lead domino in the zero-to-one stage?”

Tyler: ”Product velocity and product-led growth. That was our MO and what I think we're best known for. But it truly came out of necessity—almost out of insecurity. Users reached out to me every single day, complaining that we weren't up to par with the other platforms. I'd wake up to fifteen angry emails from people asking for things they could easily do on their old platform, threatening to leave if we didn't build it.

That anxiety and urgency pushed us to move quickly and hire more engineers.”

As far as I remember, anyway.

Bill: ”So, I guess you needed to ship and ship fast?”

Tyler: “Yep, our whole product launch strategy came out of three buckets. One is educating existing users; when you launch a new product announcement, you're basically saying, ‘I heard you, we built it, it's live today.’ You're preventing churn from the people who were about to leave. Two is marketing: when you launch new features publicly, you're checking boxes for people who were watching but hadn't switched yet because you were missing XYZ. And three is the combination of both. When you do it every week, you change the narrative from ‘they're missing a lot of features’ to ‘they ship so quickly that even if they only have 70% of what I want today, I'm confident they'll have 100% in a few weeks.’ That was the biggest lever. It let us punch well above our weight.”

Something else that was layered into those early days, and still rings true today, was building in public: Tyler showing that he genuinely cares, that they listen to customer feedback, and that he can be personally DM’d if the proverbial excrement hits the fan.

This final point is actually something I witnessed firsthand. I clearly remember a time when my buddy Tom Alder, who writes Strategy Breakdowns, and I had a problem with our CTR on the beehiiv platform and ended up chatting directly with Tyler on Slack for an hour to solve it on the spot.

How the sausage gets shipped (the velocity machine)

Product velocity, you see, is what makes beehiiv tick. While their early speed can be reverse-engineered from a place of fear and desperation, the company is just very much built to move fast. Let’s rip into a few of their key concepts.

1/ Improved handoff-ectomy

For one, all three founders are technical, which really helped in the early days. What also helped was that seven of the first ten hires were also engineers. What also, also helped was that beehiiv have never really hired much by way of product management and design. “When you're constantly coordinating between a designer finishing mockups, a front-end engineer building it, passing it to back-end, and then QA, there are too many cooks in the kitchen,” Tyler would tell me.

It wasn’t that the team didn’t believe in hiring for these functions; they just wanted them in one role. ‘Engineers with taste’ is the high-level way to think about it. Product engineers is another. This way, one person carries the feature from idea to shipped. Every handoff you remove is a queue, a meeting, and a misunderstanding you never have to pay for. Co-founder Ben is, in Tyler's words, the most creative and design-forward engineer he has ever come across, too. Ben doesn't need mockups. He just has taste. So in the early days, they cut the design function entirely, to make way for lightning-fast handoff-ectomy (new word).

_{*Note: This is not to say absolutely everything went smoothly; features were launched and then abandoned; things broke, etcetera, etcetera, but beehiiv have remained very lean on the PM side still to today with 40-50 engineers and only 5-6 people in product and design.}

2/ The 24-hour PRD

While Tyler helped flesh out the hiring philosophy, Shreya filled me in on how nimble their PRD (product requirements document, for you non-product readers) is. In her previous PM roles, Shreya saw feature progress over weeks or months: figure out what to build, write a long document, get fifteen stakeholders to review it, rounds of back-and-forth, then a tech review, then design, then finally engineering. The thinking happened entirely before the building.

At beehiiv, the building starts before the thinking is finished.

The unlock here is psychological, not procedural. Most orgs treat an unfinished vision as a reason to wait. beehiiv treats it as a reason to start, because building the thing is how they finish the thinking. They only need two things locked before they move: the core opportunity and the core functionality. The rest gets solved as it shows up. "Sometimes we'll be launching something the day after tomorrow, and someone comes in with 'it'd be really cool if we did this.' Nobody asks why they're bringing it in so late. Everyone just thinks about how we can make this work in this timeframe.”

3/ The ambitious launch date

And finally, this would all be for naught if it weren’t for Tyler’s ambition around a day and time in the calendar. But first, let’s talk about how they structure their cadences. The shape of a quarter looks like the following: a quarterly roadmap, then three-week sprints, each followed by a one-week cooldown. Three sprints and three cooldowns per quarter. Sprints are where the bulk of feature work happens, usually one project per sprint, with each engineer carrying one or two major features a quarter. Cooldowns are for bugs and quality-of-life fixes.

The deadline philosophy is the interesting bit, and it is genuinely counterintuitive. They set deadlines they fully expect to miss.

An ambitious deadline is not a promise; it is a forcing function. It pulls the work forward even when it slips. A realistic deadline gives you a realistic pace; an ambitious one you are allowed to miss gives you beehiiv's pace.

beehiiv’s pace does not mean Roman’s pace.

Something else I found refreshing was that beehiiv protects their deep work time like Ghost protected Jon Snow. They carve Tuesdays and Thursdays out as ‘meeting-free’ days to go heads down, bum up, and get work done. And it’s one of Tyler’s go-tos, as per Shreya: “It's almost like a challenge. That ability to create momentum on a dime is genuinely one of his strengths.”

Do things that don't scale → customer love

In 2013, Paul Graham wrote the most quoted essay in startup history, telling founders to recruit users by hand and do the unscalable stuff early. Think the Airbnb founders going door-to-door manually taking photos that looked good so their listings looked appealing, or Reddit’s Steve Huffman and Alexis Ohanian keeping the site busy with fake user accounts. beehiiv read Graham’s essay and ignored the part about stopping once you scale.

This mindset also began with CCO Preeya Goenka, who, in the early days, would make sure users were happy and onboarded well by holding their hands all the way through the process: "It wasn't 'here are ten things you have to do.’ It was me personally importing your list. Let me personally help you build your template."

Four years and 55,000 customers later, every PM still onboards real users by hand, on purpose. This type of service, and more importantly, care, isn’t something that just happens on its own. It starts with Tyler, who famously answered support emails at the Hour of the Wolf (let’s call it two in the morning) years into the journey, then through to Preeya, and on to the product team.

Source.

This type of devotion to the customer means you end up with loyal users who are passionate enough to shout it from the rooftops. These are the NPS nine and ten folks; those who tell their friends and their entire social audience about their experience. Priceless advocates. I consider myself one of them. In Shreya’s words, "Ten people who love it are worth more than hundreds who just like it." beehiiv have tens of thousands of said people.

Source. Source.

Source.

_{*Note: For context, Darragh above runs Half Baked, which boasts 130,000 subscribers, and Matt Paulson runs MarketBeat, a $50 million dollar financial publication, housed on beehiiv.}

And it paid massive dividends. "We didn't spend a single dollar on acquisition for the first eighteen to twenty-four months,” Preeya would tell me, “Everything was entirely through word of mouth and social.” This means the money machine definitely goes brrrrrrr, but you also see countless people shitting on the competition. Which leads nicely into the next chapter in our beehiiv story.

Every hero needs a villain

If you were one of those people, such as myself, who devoured the Netflix documentary The Last Dance during COVID, then you know that Michael Jordan has a voracious appetite for an enemy. He hungers for them; feeds on them; it gives him energy. Tyler works in the same way: "Every company benefits from having some kind of enemy or rival... Michael Jordan would just manufacture rivalries out of thin air."

From what I can see, Tyler and the entire beehiiv community at large have three ongoing beefs: 1) Substack, 2) the company formerly known as ConvertKit (Kit), and 3) the incumbency (Mailchimp, MailerLite, et al). Substack is a beef because they don’t share the same creator-first ethos as beehiiv; ConvertKit is more personal; and the incumbents, well, they barely register.

Starting in reverse order, in the earliest days, it was important for beehiiv to counter against the MailChimps of the world, but senior PM Shreya Srikanth said it best, “We were just confused by them. They never seemed to ship anything... I'd always have a FigJam open with screenshots of every flow, and I'd look at it and think, 'this is so bad, we can do this ten times better.'” Rivalry level out of 10 = 3 (afterthoughts).

Next, we have the beehiiv x ConvertKit beef. Or, more specifically, the beef between Tyler and his opposing founder, Nathan Barry. The beef started when beehiiv made a critical mistake that sent their emails into spam for two full weeks. A real down moment for the early-stage startup at the time. “Every day for three weeks, I was fielding threats from users wanting to leave, taking it to social, furious,” Tyler would say.

The beef hit a fever pitch when Nathan and the ConvertKit team began emailing beehiiv users telling them about the mistake, essentially saying they would never do that at ConvertKit because they ‘actually know what they're doing.’

”We were genuinely vulnerable. Not much cash, best users are all experiencing a problem we'd caused. Not a position of strength.” As you can imagine…

What Nathan didn’t account for was who they were going into battle with. If your product is your weapon, then they were effectively bringing a rusted teaspoon to a gun fight. beehiiv have gone on to lap ConvertKit on the brand, product, growth, and overall success front since. Rivalry level out of 10 = 6 (genuine combatants).

The last rivalry, the philosophical one, and likely the most vicious, is that of Substack. Aside from the fact that Substack has a well-documented Nazi problem, what Tyler is referring to here are the moves both companies make: one to empower and protect creators, the other to disempower and restrict. Let’s take a look at some of the most compelling differences.

"It's the classic social network playbook. Gradually eroding the ownership, identity, and independence of writers on the platform," Tyler would quote in his newsletter, "If you're not a breadwinner on Substack, you're the yeast."

Read more here.

I got the feeling that throughout this whole rivalry thing, Tyler thought of it as a real strength.

Bill: ”Does going into battle against rivals energize you?”

Tyler: “Totally. Having a rival you actually dislike, even a manufactured one, drives competition and urgency. It's why cost-plus contractors and government work have no incentive. There is no competition driving anyone to be better.

Doing it the right way, where other people also feel motivated by the mission, matters. What I've done with Substack is rooted in something real. I genuinely believe we're on the right side of history. The business model we've built and what we're empowering our users to do is more creator-friendly than most of what's out there. When you believe the work you're doing is the right call, and you know other companies are trying to take your market share, that brings genuine emotion. The more invested and passionate I can be about this company, the more I can rally a team around it.”

Tougher times

Before we continue the piece, I want to make sure I paint a quick picture showing it hasn't been all sunshine, lollipops, and rainbows for beehiiv. As per the nature of startups, they’ve certainly encountered their fair share of slow-motion knife fights, along with some genuine heartache.

• The great deliverability meltdown: beehiiv moved too fast, isolating their senders and dumping their best customers into spam for three weeks. Email reputation has no hotfix, so Tyler just had to wait and eat the churn threats. It's where the ConvertKit rivalry got personal.

• The seed extension from hell: 27-year-old Tyler raised $2.6 million, assumed it was infinite, then blew through it on engineers and growth. The extension he thought would close in a week took three months of passes and ghosting VCs, all on fumes.

• The burn scare: beehiiv torched nearly 20 months of runway in 16 weeks, watching it crater from 35 months in January to 16 by May. A hard reset cut the burn rate more than 60% and clawed runway back to 2.5 years. But things were trending downward, fast.

Source: Big Desk Energy.

• The growth reset: The early growth team was great at zero-to-one, but couldn't crack $20 million in ARR. Tyler cut a third of the team and paid spend from $700k a month to $50k. It cost them their momentum and eight scary months to win it back.

• The Boost that fizzled: Boost could be a multi-million dollar revenue arm (Sparkloop is a whole company built on the same mechanic). But beehiiv shipped it, then moved on. The classic tax of shipping fast and never circling back.

The most difficult period in beehiiv’s history, though, was the passing of an early team member, Andrew Platkin. Andrew was beehiiv's first CTO and co-founder. Per Tyler, he took a pay cut to lengthen the company’s runway, worked hard and passionately with the team, and then passed away suddenly six months after launch. There's no neat takeaway here. The team taught themselves his code, vested his equity to his mum, and kept shipping as tribute.

Climbing the logo ladder + creator economy 3.0

Another reason the team at beehiiv have been successful is that “Tyler is extremely plugged into the ecosystem, the market, and the customers,” says Preeya. “He truly understands their pain points and how they operate.” If you’ve followed Tyler and beehiiv for any period of time, you’ll see this. Countless events, podcast appearances, and other bits and pieces of PR really add sheen to the beehiiv brand.

This constant drive on the brand side has made it much easier for the sales team to work behind the scenes to win business from other platforms. Remember that snowball from earlier? This is it picking up mass. As Tyler puts it, "You keep climbing based on credibility, logos, and case studies."

Aura farming.

Product team.

Preeya, Tyler and a man named Lachlan.

If you remember back, this all began with ~20 of 25 strategic angels moving their newsletters over within a month of launch. The first big logo was Litquidity (100k subs, an angel), then Matthew Berry (300k), then Milk Road, all within months. Each logo became a case study that brought the next, bigger logo aboard. They redlined the credibility hill climb until they signed Time, TechCrunch, Newsweek, and Sinclair. Fun fact: Arnold Schwarzenegger’s media arm calls beehiiv home.

Damn straight, homie (Source).

When beehiiv started, to be a ‘creator’ meant someone with a podcast and a day job. A side hustle with delusions of grandeur. (Guilty, your honor.) That’s not the world we live in anymore. The creator economy is now a $250B+ industry, bigger than global gaming, and the people in it run themselves like actual businesses. Legacy media and creators have stopped living in separate lanes. Netflix, Spotify, retailers, clothing brands, all piling in. Independent journos who are fully solo but also cutting deals with broadcast giants: from Ezra Klein to Bari Weiss. The whole thing converged while nobody was looking.

beehiiv's bet is that these creators-cum-businesses don't want fifteen tools. They want one. The newsletter is still the beating heart, but stacked on top now: a website builder gunning for Webflow, podcast hosting that pipes you out to Apple and Spotify, subscriber segmentation, a referral program, polls, surveys, automations, and an Ad Network they’re betting on being the best thing since sliced bread. There's a Link in Bio hub, digital products, even webinars. Oh, and an AI layer that builds your site, helps you write, and lets you literally chat with your podcast analytics like it's a mate at a bar.

Source: Big Desk Energy.

They're also not done. I know for a fact that the Summer Release 26 launch will have some major community, Ad Network, and content updates, but unfortunately, I’m sworn to secrecy and cannot divulge (you can register to see it for yourself here).

The final big bet worth mentioning is the aforementioned AI layer, particularly the beehiiv MCP. The MCP, which was released in March, is the scion of a family of legendary features, but it may be their greatest one yet, crowning beehiiv the first newsletter platform you can run straight through Claude, ChatGPT, Gemini, or Perplexity. You can now not only read, but write, using beehiiv’s MCP, all from the comfort of Claude Cowork.

Source: Big Desk Energy.

Tyler believes we're barreling toward a world where people stop clicking around inside apps and just tell an AI what they want done. The winners will be those who build for both agents and humans simultaneously. Turns out, if you ain't AI-first, you're last.

💡 Note: If you are looking for the best place to publish your content, you can use the code BILL30 for a 30% off 3 months on beehiiv.

Playbook / how you can apply this

• Ship to change the narrative, not just the product: A weekly launch isn't only a feature, it's a message that, ‘We listen, and we're catching up faster than you can leave.’ beehiiv turned ‘they're missing features’ into ‘they'll have it in a fortnight.’

• Velocity is an org design choice: Full-stack engineers with taste, no handoffs, no separate QA in the early days. Every coordination layer you remove is a speed you gain.

• Do things that don't scale, then refuse to stop: White-glove onboarding got beehiiv its first 24 months for $0 in acquisition. Three years on, every PM still onboards real users by hand.

• Build your roadmap from your inbox: ~50% of the roadmap was direct customer asks. When someone says, ‘I'll switch when you have X,’ build X and call them back in a month.

• Pick a real enemy: Not for spite, for energy and clarity. Watching Substack go social is what told beehiiv who it wasn't.

• Different strokes for different folks (and that's not a failure): Their perfect team for zero-to-one was the wrong fit for $20M→$100M. Knowing the difference and acting on it is the job.

Future

So, where does this end up? Who knows, really? What I do know is that the creator space is in a much better place than it was in the past. And with their chin up, and chest out beehiiv will continue to sprint into this future, with the same exuberance they’ve always had.

Source.

Here's the thing, though. Every all-in-one platform on Earth promises this exact roadmap. The graveyard is full of them. The reason to believe beehiiv actually ships it is the same reason you've read this far: it's the one thing they've never once failed to do. I ship, therefore I am. That's the whole bet.

Extra reading / learning

• The Acquisition Mechanics: beehiiv + Swapstack - January, 2025

• How beehiiv Raised $50M (Fast) - April, 2025

• beehiiv’s 5-Year Sprint to $20M ARR - August, 2025

And that’s it! You can follow Tyler on LinkedIn and Twitter, and also check out beehiiv on their website to learn more.

BRAIN FOOD 🧠 

TWEETS OF THE WEEK 🐣 

Source.

Source.

Source.

TOOLS WE RECOMMEND 🛠️

Every week, we highlight tools we like and those we actually use inside our business and give them an honest review. Today, we are highlighting Framer—the site builder trusted by startups to Fortune 500.

beehiiv: We use beehiiv to send all of our newsletters. Apollo: We use Apollo to automate a large part of our 1.2M weekly outbound emails. Taplio: We use Taplio to grow and manage my online presence.

See the full set of tools we use inside Athyna & Open Source CEO here.

HOW I CAN HELP 🥳

P.S. Want to work together?

1. Hiring global talent: If you’re hiring tech, business or ops talent and want to do it 80% less, check out my startup, Athyna. 🌏

2. See my tech stack: Find our suite of tools & resources for both this newsletter and Athyna here. 🧰 

3. Reach an audience of tech leaders: Advertise with us if you want to get in front of founders, investors and leaders in tech. 👀 

That’s it from me. See you next week, Doc 🫡  P.P.S. Let’s connect on LinkedIn and Twitter.

👋 Howdy to the 4,782 new legends who joined since our last edition! You are now part of a 482,185-strong tribe outperforming the competition together.

LATEST POSTS 📚

If you’re new, not yet a subscriber, or just plain missed it, here are some of our recent editions.

📐 The Lab Builders Are Leaving. An interview with Andrew Dai, Co-Founder & CEO at Elorian AI.
🧗🏻 Unfiltered: The System Vanta Built To Grade Its Own AI. How they build language models as the judge, human as the jury, and more.
⚡️ The Founder Lightspeed Backed Twice. An interview with Ali Hussain, Co-Founder & CEO at Tabs.

PARTNERS 💫

Meetings aren’t the problem. The challenge is everything that comes after them. With Granola, the AI Notepad for people with back-to-back meetings, it's no longer about scrambling to keep up.

Granola works in the background, turning conversations into clear summaries, action items, and next steps. Before, during, or after a meeting, you can chat with your notes to quickly understand what needs to happen, write follow-ups, or share context with others.

Every newsletter wants to grow. Few actually do. beehiiv is the platform built by ex-Morning Brew folks who got tired of duct-taping ten tools together, and newsletters running on it grow 2.75x faster. Referrals, recommendations, analytics that mean something, all in one spot. It's the closest thing to a cheat code we've seen for getting your words in front of more people.

Get 30% off your first three months on beehiiv with the code: BILL30

_{Interested in sponsoring these emails? See our partnership options here.}

HOUSEKEEPING 📨

It’s honestly so simple to outcompete your competition or simply outlive your neighbor. Gambling sucks, and if you have social on your phone, get rid of it. You will be light-years happier and more successful. I am pretty successful with numbers one and two in Shaan’s lift here, and am slowly working on the third.

Source.

One thing I’d add to this list is reading. If you read a few pages every night before bed, you will also be outlearning and giving yourself the best chance of getting quality sleep. One of life’s great win-wins. I’m currently reading Dracula. I’m loving it.

INTERVIEW 🎙️

Sid Yadav, Co-Founder & CEO at Circle

Sid Yadav is the Co-Founder and CEO of Circle, the all-in-one platform for community-led businesses, covering membership, courses, events, discussions, and, as of this month, a full website builder. Originally from India, Sid grew up in New Zealand, where he spent his teenage years writing Rev2, one of the internet's earliest tech blogs, before moving to the US as a founding engineer at Teachable, where he eventually rose to VP of Product and helped scale the platform to over $1B in creator earnings before its acquisition. He co-founded Circle in 2019 with Andrew Guttormsen and Rudy Santino, both former Teachable colleagues. The company has raised $33M, with Tiger Global leading the Series A at a $200M valuation, is now profitable, and has crossed $50 million in ARR with 280 people and a customer list that includes Jay Shetty, Tim Ferriss, and Ali Abdaal.

What makes Sid's perspective worth paying attention to isn't just the trajectory. He and his co-founders wrote Circle's values document before they signed their Stripe Atlas incorporation papers, having watched culture quietly degrade at Teachable as it scaled. He accidentally got the timing right too, launching in August 2020 into a lockdown that made digital community feel urgent for the first time, and unlike Zoom and Shopify, Circle never gave back the growth. He also still codes new apps every few days to understand the business. At 280 people, the big question now is whether turning Circle into an AI platform, which is exactly what a major recent launch signals, changes what the community model actually is.

What are you building at Circle?

We started six years ago. It's the all-in-one platform for community-led businesses (we call it an ‘ecosystem’ these days). The thesis we have is that there are all these people building content, making videos, writing newsletters, which is a fantastic way to build an audience, but when you look at a lot of these businesses and creators after a few years, there can be a sense of fatigue and burnout. This feeling of being on a content hamster wheel where you have to keep producing and producing. And on the audience side, there's this feeling of ‘there's a lot of content out there, but am I actually doing something with my life?’

The idea with Circle is to build this ecosystem that gives every creator and entrepreneur the platform to build their dream community-led business. These are really any businesses where people are at the center of the value proposition, versus the content or the audience. They typically have aspects of membership, discussions, events, courses, something for people to show up to, to interact with each other, and get value out of. Not every community-led business is the same. They all vary in shape, size, and outcomes. But there's typically an element of belonging, and an element of transformation, where you don't just show up to hang out in a Discord or a Reddit—you actually show up to grow as a person.

Source: Circle.

What are the best case studies you have on the platform?

To start with the known names, we're very grateful that we've built a platform used by some of the top creators, like Tim Ferriss, Ali Abdaal, and Jay Shetty. We've got all the Harrises, too (Sam Harris, Dan Harris, and Johnny Harris). They all follow a similar theme, which is that these people have built audiences, and at some point in their creator-led business, they discover that they want to run events. They want their people to have an area to belong that isn't just the YouTube comments. So they need a platform, and especially when they build in-house teams and get serious about it, we tend to become that default platform.

Source: Circle.

Source: The Brain Docs.

But Circle isn't just for the top creators. It really is for everyone. I have a couple of more concrete examples I find genuinely cool. There's this couple, Dr. Aisha and Dr Dean, who are neuroscientists. An entire career thinking about brain health, operating on patients, doing surgeries, and they had this idea at some point: we can operate on hundreds of people over the course of our lives, but if you actually want to impact the world, you have to do more than that. So they started a community called ‘The Brain Docs’. It takes all of their learnings and builds a community experience around people who want to get better at brain health. They run weekly cooking sessions with ingredients that are best for your brain, and so they've built this audience of people who aren't just passionately watching Andrew Huberman interviews, but actually doing something about it. Here are these people who built a career, who are still doctors, but now also have this business. They've had much greater impact, and all of it has been built on Circle.

Sam and Dan Harris are the ones you'd expect people to know. What do they actually do with their communities?

Dan Harris’ ‘10% Happier’ is a good example. He's written books, been on podcasts, and built an audience around happiness. These creators share ideas, build audiences, get book deals, monetize the ads, but at some point, there's this feeling of ‘we're speaking to these large audiences, and where do they go after that?’ Dan's community membership is really built around just that. There are a lot of people who love Dan Harris but also want to actually live out his ideals with each other. And what a lot of these creators find is that, unlike a Substack or Patreon, where people show up mostly for the creator or the content, with Circle, they show up for the creator, but they stay for the community. Once they engage with each other, there's this feeling of ‘I'm genuinely happy just meeting people who are like me, undergoing the same transformation, and I'll pay for that’. They don't actually need to keep watching the creator's exclusive content to stay engaged.

What was the hardest thing about getting the company off the ground?

Honestly, I don't think I can complain too much, because we had a phenomenal first year. Unlike a lot of other stories I hear, we were starting the company in late 2019, and then COVID happened. The creator economy really entered the mainstream zeitgeist. There was this real need people felt for communities, especially online digital communities. We started the platform at this accidental stroke of luck moment, and we went from zero to a million in ARR in three months, from our August 2020 launch to December. So we don't have the typical struggle of finding product-market fit.

The inner Circle.

With that said, I had a lot of personal challenges that first year. My co-founders and I are all engineers and designers, very hands-on people. The company is 280 people now, and we're still very hands-on. The challenge was selling the product to all these people and onboarding them while building it yourself. We were engineering and designing the product as it went, which was fantastic for the feedback loop, but we were selling, building, and marketing simultaneously. And then, because we'd raised pre-seed money, building out the initial team, delegated support, brought in more engineers, formed the first leadership team, all in one year. TLDR: it was just a lot of work. No weekends off. So my struggles were really personal struggles around time management and delegation.

Did you worry when lockdowns ended?

Yeah, we were paranoid about it in some ways as COVID was ending. I know it affected Zoom, Shopify, and Stripe; all of these companies overhired and then had to correct for it. For context, our ARR went ‘triple, triple, triple, double, double’ journey, not that we were intentionally optimizing for that formula. But I think the reason we didn't struggle is that the creator economy just kept doing this and growing, unlike virtual events, Zoom calls, or working from home. Those trends reversed, while the creator economy didn't.

Also, the definition of community and where these communities lived changed a little bit after COVID. In-person events are returning, but those communities still need Circle. They might not need it as intensely for chat and discussion inside the product, but they still need that home base for their members.

On top of that, the creator economy itself just kept getting bigger. Podcasts are everywhere. Creators have bigger and bigger audiences. That never slowed down.

What changed after the Series A?

So we did $1.5M pre-seed, $4M seed, and $20M Series A. The biggest shift, honestly, was resources. In the seed days, there was always this runway countdown. In the pre-seed days, we paid ourselves as founders almost minimum wage, because we didn't want our salaries affecting the burn rate. You're at sixteen months, then fifteen, fourteen, thirteen. Then the seed tops it up a bit; you're at thirty. But then you build out the team, start spending a bit more, do your first offsite in year two, and even that starts to count back down. Then you raise the A, and suddenly the runway is just massive. In our case, the burn was fairly low, and we'd just raised $20M. You can't even really think about the runway anymore. And then the problem flips. Now the question is how fast you can deploy this money across all the areas of the business that really need the best people. And those people now need teams. About a month after we raised the Series A, we hired our head of sales, head of customer success, and head of finance. So the first job was building out that initial leadership team.

Source.

For me, as someone who'd been doing everyone's job, it was about backing off a little and helping these people establish their roles. We went through this evolution in phases. Phase one, we did all the work. In phase two, we hired generalists who could handle more of the work while we were still very hands-on. Phase three, leaders building their own teams, those teams needing to be really aligned with your vision and values. And then it's just a delegation, trust-but-verify scaling game from there.

What does your actual week look like as CEO right now?

Every day is pretty different, but I try to maximize my time with my co-founders and a very small number of people in the leadership team. I think about it in terms of leverage. If I'm talking to my co-founder Andy for an hour and a half, he runs all of GTM, which is close to eighty to a hundred people under him. Whatever topics we're covering in sales, marketing, customer success, that's now being taken care of. That frees me up to spend the rest of my time on areas where I'm deeply engaged, which tends to be culture and product, and usually one big project. Recently, it’s been a major launch we led in June.

What's changed in the past year is that I used to have just a ton of meetings. My entire work week was 80% meetings at some point. Now I've dialed it way down to the bare minimum, but super high-leverage meetings. Because finding my own personal time as an IC (writing to people in the company, writing memos, thinking through ideas, vibing with the AI) is actually some of my highest-leverage time. I'll write and code a new app every two to three days just to understand or operate on some aspect of the business.

Source.

And then there's being the face of the company internally and externally, which only I can do in my role. So I've completely flipped it from a week full of meetings to trusting the team to run most of that and getting super involved wherever I think I'm the best use.

How did the culture change as you scaled?

One thing we had going for us was that my co-founders and I spent five years on the founding team of Teachable before Circle. I grew up there, went from IC engineer-designer to Product Manager to running Product. The company sold for about a quarter of a billion dollars. But one thing we learned from Teachable is that we never codified our culture or values very well.

In the early years, everyone was just having fun, but as you scale, aspects of the culture get degraded. We all felt that. So with Circle, literally the first thing my co-founders and I did before even signing the Stripe Atlas incorporation documents was sit down with a values doc that I drafted, and they commented on, going line by line on what this company is about.

Source: Circle.

A lot of those top values came from experiences at Teachable, where we'd felt them but never written them down. That document still exists today, and it was there when we onboarded our very first team member. The most challenging period for culture was probably when there were 50 to 100 people.

At that scale, each team begins to define its own culture. You're not just building your leadership team—those leaders are also building their own teams with their own subcultures. And you're negotiating how much of their subculture maps to Circle's defaults versus how much the sales team just does things one way and the design team does it another. That was one of the hardest challenges.

Source.

How is Circle using AI operationally right now, and where is it going?

About 18 months ago, we did this AI mandate offsite in Spain. We're remote and international, so we flew the whole company to incredible locations. We spent an entire day with each leadership team and started wondering what this function would look like with AI. It started with an insane process mapping. The philosophy was that this won't be top-down or bottom-up; it'll be both. It's not going to be the leadership team or me dictating how everything changes with AI, but it's also not going to be a free-for-all. So we mapped what everything looked like, assigned AI leaders in each team, and held those leaders accountable to AI outcomes. And we took the approach of defining the outcomes, defining the leverage points, but not being opinionated about the tech stack. If one team wants to use ChatGPT and another Claude, that's fine, as long as they're achieving outcomes. So Circle has achieved a lot across the board, but in this somewhat fragmented state where we haven't unified the tech yet, and that was intentional.

Circle's 8th offsite in Marbella, Spain.

Now we're entering a phase of unification. If sales, CS, and CX are all doing their own AI thing, but they're all dealing with customers, what does that one true end-to-end system look like? And more importantly, how does the product itself change? We're going to be launching a lot of critical AI features, almost turning Circle into an AI platform, so that a bunch of things that currently require human intervention just won’t anymore. Before a ticket even emerges, before someone in customer success has to talk to the person, their jobs evolve to higher-level problems.

How do you get the best out of yourself?

I cold plunge and sauna every day. That's been a massive personal unlock. You know how it is running a company. By mid-afternoon, you're a bit exhausted and not bringing your best energy; if you're talking to high-leverage people, whatever energy you're bringing is affecting them. So around 3 PM, I head out for a quick sauna plunge, and then I'm back for round two. It’s almost an entirely new day for me after that, an extra four to six hours of good energy with a full reset. Before I did this, at 4 PM, I'd just be exhausted.

The deeper thing is that, as a Founder, you think you have to be working all the time because there are always more problems. But when you do something like this for yourself, you come back with better energy, a clearer mind, your decisions get better, and you feel it. That's the counterintuitive bit. Investing in yourself beyond a certain point is also investing in the company. You don't have to feel guilty about it.

Extra reading

• Circle's Eclipse Is The Biggest Creator Launch Of 2026 - June, 2026

• The best ways to use Circle MCP - June, 2026

And that’s it! You can follow Sid on LinkedIn or check out Circle on their website to keep up with what they’re building!

BRAIN FOOD 🧠 

TWEETS OF THE WEEK 🐣 

Source.

Source.

Source.

TOOLS WE RECOMMEND 🛠️

Every week, we highlight tools we like and those we actually use inside our business and give them an honest review. Today, we are highlighting BILL—Your AI-powered financial operations platform.

See the full set of tools we use inside Athyna & Open Source CEO here.

HOW I CAN HELP 🥳

P.S. Want to work together?

1. Hiring global talent: If you’re hiring tech, business or ops talent and want to do it 80% less, check out my startup, Athyna. 🌏

2. See my tech stack: Find our suite of tools & resources for both this newsletter and Athyna here. 🧰 

3. Reach an audience of tech leaders: Advertise with us if you want to get in front of founders, investors and leaders in tech. 👀 

That’s it from me. See you next week, Doc 🫡  P.P.S. Let’s connect on LinkedIn and Twitter.

👋 Howdy to the 4,316 new legends who joined since our last edition! You are now part of a 477,403-strong tribe outperforming the competition together.

LATEST POSTS 📚

If you’re new, not yet a subscriber, or just plain missed it, here are some of our recent editions.

🧗🏻 Unfiltered: The System Vanta Built To Grade Its Own AI. How they build language models as the judge, human as the jury and more.
⚡️ The Founder Lightspeed Backed Twice. An interview with Ali Hussain, Co-Founder & CEO at Tabs.
🗑️ The Future New Media: a16z. How the famous VC firm has thrown out the old playbook.

PARTNERS 💫

Founding a company is easy, scaling it is hard. Operational systems break, people problems increase, and your calendar keeps getting fuller. 

Mochary Method is an executive coaching firm founded by Matt Mochary—coach of Sam Altman and Brian Armstrong, and writer of The Great CEO Within—that specializes in helping growing startups.

The Mochary coaching team, who coaches companies like AngelList, Lovable, and Midjourney, is offering a complimentary coaching session to founders and executives for a limited time.

Every founder hits the wall where there's more work than people. Viktor is the AI employee that lives in your Slack and Teams.

Brief him like a new hire and he writes the docs, builds the landing page, runs the campaign, and ships the code. Connects to 3,200+ tools and gets to work before you ask.

_{Interested in sponsoring these emails? See our partnership options here.}

HOUSEKEEPING 📨

I’m sure by now everybody has seen the Trump financial disclosure report. But I feel like it’s irresponsible for me not to talk about it here today. What we are watching is brazen, unfettered corruption. The most disgusting part of it isn’t the 14,000 trades he’s on to make this year (Pelosi averages 15); it’s the fact he’s rug-pulled his closest supporters of $1.4 billion dollars personally, much more if you count his family.

Source: Reuters.

This is the leader of the free world and someone whom many young people look up to. The damage that will have been done to our culture by the time these four years are up will be felt for decades to come. What a sad state of affairs. Anyhow, onto today’s interview!

INTERVIEW 🎙️

Andrew Dai, Co-Founder & CEO at Elorian AI

Andrew Dai is the Co-Founder & CEO of Elorian AI, a multimodal reasoning research and product lab based in Palo Alto, announcing a $55M seed round backed by NVIDIA, Menlo Ventures, Altimeter, Striker Venture Partners, and Jeff Dean personally. Andrew spent 12 years at Google Brain and DeepMind, most recently co-leading data for Gemini, a team that grew to 400+ contributors. Before that, he co-authored papers that quietly shaped the trajectory of modern AI, including an early sequence-learning paper with Quoc Le that predated GPT and GLaM, and a foundational paper on mixture-of-experts architectures.

The thesis at Elorian is that today's frontier models (including the ones Andrew helped build) have a fundamental blind spot: they can't reliably count, match shapes, or reason over what they actually see. He calls this the visual AGI gap, and believes it's both underappreciated and foundational. His co-founders are Yinfei Yang, who led multimodal research at Apple, and Seth Neel, a former Harvard assistant professor. The fact that Jeff Dean, the man Andrew reported to at Google, chose to back the company he left to start says something worth paying attention to.

What is Elorian AI building?

We're building models for visual AGI. If you look at the current frontier models, they really fail some basic tasks that elementary school or high school kids can do, such as counting, reasoning over things, matching shapes, similar objects, and objects that aren't similar. The current models are really unable to do that.

I firmly believe you can't have AGI that can't count. If people are running out of a flaming building, you need to be able to count how many people ran out. That's a pretty important thing. We're filling that gap, and we're calling it ‘visual AGI’. It's a big gap that's not really well covered by any of the frontier models, and to do that, we're building a research and product lab. So not only are we pushing the frontier in research, but we also believe in having that research really affect people and improve their lives.

What does your week actually look like as a first-time CEO?

It's a lot more meetings than I expected. Even more than when I was leading the data area in Gemini. Meeting with candidates, meeting with our investors, meeting with potential customers, people in the industry who could really benefit from using these models.

And then, of course, meeting with the team to brainstorm what the next research idea we want to go after is, the next modeling technique or data transformation we want to try to build a better model. I still have some time to do some vibe coding myself, or set up the AV system, user access, and even some random admin stuff.

How is leading Elorian different from running a team at DeepMind?

Very different. The biggest thing is that we feel much more connected to what we're building. DeepMind is a few thousand people—a very big team—and there it's several steps between what you're building and seeing it in the actual Gemini product. That makes things a bit less connected. Things seemed further away.

Here, we have full control end-to-end: from building the model, to knowing exactly what data is going into it, to having full control over the API and whatever safety tuning we build. In Gemini, you could build something great, but then safety tuning could affect the final result in ways you couldn't fully predict. For us, we have complete ownership over the model and the product, and with that comes more responsibility to do it properly.

What's your North Star metric for 2026?

Our most important goal on the research side is getting to state-of-the-art in visual reasoning. We want to be better than any other model out there. The definitive model people use for visual reasoning and multimodal reasoning tasks. But we also want a solid public external API launch, which we're hoping for later this year. We want to get our model into the hands of people, into the hands of enterprises, and really see that it's clicking for them for their use cases. So, in the end, it's both: the research side and the product side.

Source: Elorian AI.

How does Elorian use AI internally to move faster?

We're an AI-native company, of course. We've all had a hand in building these models, so we don't shy away from using AI for coding and developing our infrastructure. But I think what's a bit more unique for us is that we also understand where AI is not working, where it's failing, where you shouldn't use it. These days, that's just as important to know. It's not only about knowing where to use it, but knowing where you shouldn't, and where you need to double-check the code and make sure everything is actually working.

We've seen a lot of issues, especially around visual tasks, where AI gives unreliable answers, so there we have to be more careful.

But it's definitely helping a lot with coding, which means we can build up our infrastructure very quickly. The other benefit is that we're starting from scratch. We're a new lab, and whereas other companies have a lot of old code that needs to be onboarded into coding agents, we don't have anything to onboard. Our codebase is a blank piece of paper, which means fewer issues with models having to learn unfamiliar code.

Source.

How do you get the best out of yourself?

My motivation is the vision I have for making Elorian AI one of the world's leading frontier labs. If you look at a lot of the other labs out there (OpenAI, Anthropic, Thinking Machines Lab, Periodic Labs), a lot of those founders come from Google Brain—the same heritage I have. I spent 12 years at Google Brain, and I know a lot of those founders personally. It's very motivating that there are people who've gone down this path and been able to build great models and awesome companies. It gives me confidence that I can do the same: build really great models and build a company that genuinely pushes the frontier of AI.

Source.

How do you build a great company culture?

Company culture is very important to us. We've seen how cultures can go badly at other places, creating toxic working environments and real problems, especially with researchers. For us, it starts with hiring. We have a very high bar. After our public launch, hundreds of people applied through our website, and we're likely going to hire very few of them. We've decided that even if someone is brilliant, if they can't work in a team, we wouldn't hire them. Maintaining a culture where everyone can work together matters more than any single person's individual output.

The Fellowship.

We also don't want to build five separate research groups that never talk to each other. You see that at big companies because they're so big; that's the only thing they can do. We've built a very close-knit team where everyone knows each other's strengths and weaknesses. Only then you can build the kind of trust and reliability that makes a team actually strong. That comes from both hiring and from team events. We have board games whenever people from our New York side join us. We have fun events, and we like to liven up the office. We even have a fountain on the patio one of our researchers is centering a zen space around.

We want to give people some variety in their days, a supportive team, and a mental space to think about things outside of work. With vibe coding, a lot of the boring stuff can just be done by AI now. We want to emphasize the team's creativity.

Extra reading

• Ex-Google DeepMind researchers’ Elorian raises $55m - April, 2026

• Elorian Founder Spotlight | Andrew Dai - April, 2026

• AI Mastered Language. The Harder Problem Was Vision - April, 2026

And that’s it! You can follow Andrew on LinkedIn or check out Elorian AI on their website to keep up with what they’re building!

BRAIN FOOD 🧠 

TWEETS OF THE WEEK 🐣 

Source.

Source.

Source.

TOOLS WE RECOMMEND 🛠️

Every week, we highlight tools we like and those we actually use inside our business and give them an honest review. Today, we are highlighting Lightfield*—an AI-native CRM that assembles itself from your email, calendar, and meetings.

beehiiv: We use beehiiv to send all of our newsletters. Apollo: We use Apollo to automate a large part of our 1.2M weekly outbound emails. Taplio: We use Taplio to grow and manage my online presence.

See the full set of tools we use inside of Athyna & Open Source CEO here.

HOW I CAN HELP 🥳

P.S. Want to work together?

1. Hiring global talent: If you’re hiring tech, business or ops talent and want to do it 80% less, check out my startup, Athyna. 🌏

2. See my tech stack: Find our suite of tools & resources for both this newsletter and Athyna here. 🧰 

3. Reach an audience of tech leaders: Advertise with us if you want to get in front of founders, investors and leaders in tech. 👀 

That’s it from me. See you next week, Doc 🫡  P.P.S. Let’s connect on LinkedIn and Twitter.

👋 Howdy to the 4,298 new legends who joined since our last edition! You are now part of a 473,087-strong tribe outperforming the competition together.

LATEST POSTS 📚

If you’re new, not yet a subscriber, or just plain missed it, here are some of our recent editions.

⚡️ The Founder Lightspeed Backed Twice. An interview with Ali Hussain, Co-Founder & CEO at Tabs.
🗑️ The Future New Media: a16z. How the famous VC firm has thrown out the old playbook.
🌟 How Vanta Learned To Trust AI. Wiring AI into the one product where AI isn't supposed to be trusted.

PARTNERS 💫

AI is the fastest-moving field most of us have ever worked in, and the hiring at frontier companies looks the same.

The Athyna AI Job Board watches openings at the labs and AI-first companies actually shaping the field, matches them to your profile, and sends the strongest fits straight to you.

Set up a profile once. Let the right roles find you.

Every founder hits the wall where there's more work than people. Viktor is the AI employee that lives in your Slack and Teams.

Brief him like a new hire and he writes the docs, builds the landing page, runs the campaign, and ships the code. Viktor connects to 3,200+ tools and gets to work before you ask.

_{Interested in sponsoring these emails? See our partnership options here.}

HOUSEKEEPING 📨

Not much to report as of today. It’s 8:19pm on Sunday evening in Melbourne, which is technically 19 minutes after I planned to ship this newsletter. It’s cold in Melbourne, kinda miserable, too. The one thing I feel like keeping me going is training and dog walks. But things are okay generally.

Anywho, we have a nice piece for you today. This is another one from the cutting room floor of our recent Vanta piece. These are the raw interviews with two of Vanta’s AI leaders. Super interesting stuff. I am sure you’ll love it. Cheers!

INTERVIEW 🎙️

Iccha Sethi, SVP of Engineering at Vanta

Iccha Sethi is SVP of Engineering at Vanta, where she leads all of engineering across the company's trust management platform. Before Vanta, she was Senior Director of Engineering for Compute Products at GitHub, overseeing Actions, Codespaces, npm, Packages, and Pages, and before that held principal engineering and director roles at InVision, Atlassian, and Rackspace. Vanta, now past $300M in ARR with over 16,000 customers, is well beyond product-market fit and pushing hard into enterprise. Iccha joined to lead all of engineering through that transition, from Series B to what is now a Series D company.

What makes her perspective worth paying attention to is the overlap between her background and the specific problem Vanta is now trying to solve. She's been inside two of the companies most associated with developer infrastructure at scale (Atlassian and GitHub), and she experienced compliance pain firsthand at both, as a practitioner, not as a vendor. She came to Vanta already knowing what it feels like when the process breaks. Now she's on the other side of it, building the AI systems that are supposed to fix it, while managing the harder question that nobody in this space talks about enough: how do you build AI-powered features into a trust product when AI is, by nature, not trustworthy?

What’s actually broken about how companies approach security and compliance today?

There's a whole set of security and compliance companies that aren't even thinking about AI in this space. One group is almost too afraid to touch it because they think security and AI are orthogonal—AI is too risky, it's non-deterministic, it's probabilistic. Then there are the companies that are a bit more AI-forward. They know AI sells, and they're just trying to get it out there. What people truly need to be thinking about is, now that we have a new piece of technology that can analyze information and is really good at coming up with solutions, how can we actually use it? That's what we've been trying to do at Vanta. I've been through a compliance process several times, and it's really manual. We SaaS-ified it, which meant automating some workflows, but even then, it's still do X, then Y, then Z, then ask this person, then go write this document.

What we've done instead is focus on what actually takes a long time in a flow. Writing a policy takes a long time. Analyzing all your information and developing a good set of risks takes a long time. When things change, figuring out which risks need updating and their impact takes a long time. That's still a lot of wading through information for a human, and LLMs are really good at analyzing text and generating suggestions. So that's where Vanta has been investing: identifying what's genuinely painful in that flow, where AI is actually good, and being honest about its non-deterministic nature, figuring out the right place to bring a human back into the loop.

What are the actual benefits of AI in compliance?

Security and compliance are often mistakenly thought of as a cost center, and they're never funded sufficiently. But it's actually mitigating risk for the business, and every new customer who buys from you wants you to prove you're secure before they sign. Historically, if you ask most security teams, they'll tell you they could use more people. What Vanta helps you do is achieve a lot with a smaller team. There's a ton of context in any organization (especially as you grow, a lot is happening in different parts of the company), and it helps you pull all that context together, context that no single person can hold in their head. You don't need to hire a huge team when AI can do a lot of that analysis and workflow streamlining for you. So I'd put efficiency first, speed second, and quality third.

Source: Vanta.

When we say AI improves quality in compliance, what does that actually mean?

Every feature in Vanta will have some form of AI at this point. We just call them agentic features, and the quality we're talking about is the quality of the outcomes they're enabling in the compliance space. For example, say your company was fully in-office and decided to go partly remote, or vice versa. Now you have to update your policies to reflect that your infrastructure situation has changed. Your policies impact your frameworks and your controls. A human is fallible; they won't always remember that this policy maps to this framework, which maps to these controls, and that they need to go update or tweak those, too.

What Vanta does in a case like that is, when you go to update your policy, the AI agent flags that the update affects your compliance posture across certain controls. In that way, your compliance program improves in quality because the agent helps the human do their job better by catching what they might have missed.

Source: Vanta.

We have multiple features like this. We understand what we call your ‘trust graph’—the whole picture of your company: your risks, your frameworks, your vendors, your customer commitments, your legal obligations. The goal is to be your compliance brain, so that even if no single person in your organization holds all of this, we are nudging you and surfacing suggestions to make sure your program is actually sound. That's how AI raises the bar on compliance quality.

The other side of it, how we build these things, matters too. If you give incorrect suggestions, users lose trust in the product. They start ignoring it, like they ignored Clippy. So we put a lot of work into making sure all our AI features are genuinely high quality. Evals, experiments, LLMs as judges, an eval maturity model, tracking metrics. All of that is important to us because the higher the quality of our AI features, the higher the quality of your compliance program.

Where does the human actually sit in the loop? Before the AI, during, or after?

It varies by customer segment, and we've deliberately built it that way. For smaller companies, the approach is: give us your context. We have agents that go and pull information in the background, and then we validate it with you. We take a first pass and present it back to check whether it’s accurate. Then they edit and refine from there. We have something called ‘policy builder’. It's very interactive, walking through what they want to achieve and who's responsible for what, and we build up policies together that the human can then review and edit.

Source: Vanta.

The further upmarket you go, the more that changes. Larger companies typically come in with their own policies already written. They're not looking for help creating a policy as they've already done the work. What they want help with is connecting their existing policies to the rest of their program.

How do you actually build these AI features?

It always starts with the customer problem. Engineering, product, and design work closely together. Then we encourage engineers and product managers to go build a quick prototype. That first pass lets you see what you want the prompt to be, what the flow should look like, where the human should be in the loop versus where the AI should be, and how you should surface the response. Not all AI features need to be a chatbot in your face. Some show up more subtly on a page; more MCP-first.

Once you've nailed the basic flow, you get into what we call the ‘quality hill climb‘. We work closely with GRC subject matter experts (we have an entire organization of GRC experts dedicated to helping us build products). We work with them to build the ‘golden dataset‘: what are the inputs into the system, and what would a GRC expert produce as an output? We have a firm policy of never training on customer data.

That's where the in-house expertise earns its keep. They help us create these synthetic, expert-grounded datasets. We start running the feature through that ‘golden dataset’ and scoring whether the AI output matches what a human GRC expert would produce. Whenever it falls short, we go back until we feel good about it.

When we're ready to move to production, the key is instrumentation. One of our engineering managers, Andy, built our ‘AI evaluation maturity model’. Every agentic feature needs traces, so you can see what tool calls the model is making, what responses it's returning. From that foundation, you build evaluators, and as you scale up traffic and exposure, you move to LLM-as-a-judge. And the most mature state is running experiments. The biggest difference from traditional software is that it doesn't end when you ship. The old ‘I'm done with this feature’ doesn't exist anymore. This is a constantly evolving field, and something we shipped a year ago might be running on a deprecated model. You have to keep evaluating, and that’s something we take very seriously.

How does LLM-as-a-judge actually work in practice?

When we move toward production, we go through release phases. There's private preview, where we work closely with early adopter customers or design partners who've been waiting for a feature, reviewing quality and checking whether the response meets the bar. From there, public preview, and then GA, where it stops being humanly feasible to review every agentic output. We still do some human evaluation, especially for anything a customer has thumbed down or marked as inaccurate, but for the bulk of outputs where customers don't give you that explicit signal, you still need to know whether your AI feature is doing a good job. That's where the LLM-as-a-judge evaluator comes in.

If you ask a GRC expert how they would evaluate whether AI is doing well, they'd say that they look for certain things in the response, certain fields structured this way, and they want the tone to be consistent with Vanta as a brand. So we build that into a prompt—essentially encoding how a human reviewer thinks—and run all the feature outputs through the judge, which scores them across multiple dimensions.

Source.

Some are about the correctness of the output itself, while others are about tone and brand consistency. It produces scores, and then humans use those scores to prioritize which outputs to dig into, rather than reviewing everything blindly.

Why does the model still make output mistakes if the constraints were built into the feature?

The non-deterministic nature of LLMs means that even with the same model and the same prompt, you can get different outputs. I ran the same demo six times yesterday with the same model and prompt, and on the sixth run, it gave me a completely different answer. That can happen because frontier companies are also constantly tweaking their models underneath.

The other reason is coverage. When you build a feature, you design for the customers you know and the use cases you've imagined, but maybe you didn't consider a specific type of company or setup.

So everyone else is having a great experience, but a customer in a particular industry with a non-standard infrastructure configuration is not. You might have done a perfectly reasonable job designing the feature, and just didn't have that case in your dataset. The LLM-as-a-judge workflow catches things like that and flags that you need to go back and adjust how the feature handles that new scenario.

What does your day-to-day look like?

A percentage of my time is forward-thinking—for the company and for the engineering org—and a percentage is running the engineering org operationally. I actually separate those. Forward-thinking for the engineering org has been a big chunk of my time lately, especially with everything happening in AI. I'm constantly experimenting with how we can be more agentic and what frontiers we should be pushing in the compliance space. Another big part is thinking about how engineers at Vanta use AI to improve their own productivity. For greenfield work, it's easy to move fast. For brownfield areas, you have to think harder about architecture and API design. I just created an AI developer experience team to focus on exactly those questions.

On the operational side, I spend time with stakeholders outside engineering—product, design, our CEO, Christina, marketing, and sales. Then a lot of time with my direct reports, and I do a lot of office hours. In the last three months, I've done over 50 of them. It's my style to stay connected to the ground. Right now, I have 28 Slack messages sitting there, and probably 15 of them are from individual contributor engineers with whom I have a direct line on some topic or other. That accessibility matters to me, no matter how large the engineering org gets.

Ignacio Andreu, Director of Engineering at Vanta

Ignacio Andreu is the Director of Engineering for AI at Vanta. He joined the team in late 2024, after four years as a Software Engineering Manager at Google, where he led the internal tooling for generating and annotating data to train the LLMs powering Google Search's AI Overviews. Earlier in his career, he co-founded Masterbranch, a startup that aggregated open-source code contributions to build verifiable developer CVs, raising €470k and landing coverage in TechCrunch before eventually moving to San Francisco and the American tech ecosystem. At Vanta, his AI team grew from 3 to over 20 engineers in 15 months, and now has 45 people.

Ignacio is an interesting voice on AI in enterprise software because of the specific problem he's working on. Vanta sells compliance, which is fundamentally about trust. LLMs, historically, are not trustworthy. Threading that needle, building AI-powered workflows into a product where a wrong answer isn't a minor annoyance but a potential audit failure, puts him in one of the more demanding engineering contexts in the current AI wave. He came from Google Search, where latency is measured in milliseconds and 3B daily users create zero tolerance for sloppiness, and he's now applying that data-first, eval-driven rigor to a company of a very different scale with very different stakes. He has strong opinions about what gets lost when engineers apply 2019 thinking to 2025 models. And he doesn't sound like he's read them in a blog post.

What does ‘agentic trust platform’ actually mean?

I'm going to answer that more in product terms, because I think that's where it actually lands. Vanta has been very useful for a long time, but we can now do much more for our customers with AI. There are many manual processes in the compliance space that LLMs and agents can handle, but one very important distinction we make is that we always want the user to be in control. This is your program; this is your information. You need to know what you're signing off on. So even when our agent is working in the background doing more, you are still in control. You are always the decision-maker.

From an engineering point of view, that means we start by involving LLMs in many of these decisions, using agentic workflows, giving those agents more context, and providing enough information to the LLM to actually reach an outcome. When we talk agentic development, especially with reasoning models, we're talking about outcomes-based design. We tell the LLM what we want to happen, but we don't need to be prescriptive about every single step. We let the LLM decide.

Source.

From an engineering perspective, it also means that the cycle has changed. It's a lot more data reviews, data iterations, evals, and evaluators, and keep improving the quality because AI features are broad by nature. The variety of input you can see is very wide.

Where's the real value of agents in compliance? Speed, accuracy, or something else?

It's both. Think of the Vanta agent as your 24/7 compliance engineer. Someone is constantly looking at your program, aware of what's happening, and recommending what to do based on the shape of your specific setup. In compliance, everything is interconnected. Think of all your objects as a graph. When you change a policy at your company, that usually means you have to change some of your controls. Classically, before something like an agentic platform, you'd change your policy, then manually search through all your controls, find the ones you need to change, maybe think of ones you need to add, review if they don't contradict each other, then move to the next step, and so on.

Now you can start in one corner, let your Vanta agent find everything for you, and do the work. It's both speed and helping companies become more compliant over time through constant monitoring. If you think about Vanta's origins, we monitored your resources, your AWS, your IDP, whatever it was, so you could improve your compliance. With agentic flows, this is a step further. We're monitoring your entire program, and what matters most is that if you miss something, it will catch it for you.

What's the hardest engineering problem in AI compliance automation right now?

One is intrinsic to the time we're living in—everything changes so fast. What we're talking about doing right now was literally not possible six months ago. That is one key element of how we need to think, which may not be a purely technical problem but is a real one. Lately, when we work on features, my go-to question is ‘Are we doing the 2019 version of this feature or the 2024 version?’ The engineers need to understand what is actually possible right now, because it's easy to have an AI intuition that pushes you toward doing something that makes the model less powerful, where you write very detailed instructions and remove a lot of the capabilities that these LLMs have. Getting that balance right is hard.

Source: Vanta.

And then there's the technical side: software engineering is becoming semi-data science. You have to look at the data, understand it, cluster your problems, then look back at your prompt, your harness, your system, and your context, and decide where to add.

How do you actually stay on top of what is possible when a new model drops?

It's evals, primarily, which require datasets and evaluators. When something drops, you want to see how it changes your datasets. You usually end up with two different datasets. One is almost like a regression dataset. Things that can't go wrong, your baseline. The other is more aspirational. Tasks that don't work well today, or a mix. That second one is where you can actually see how a new model improves things, and you can do that without writing a single line of code.

Source.

After that, I believe engineers should use coding as well as general tools that actually expose these models to them. A lot of this is an art. There's no pure science behind prompting. Model providers such as Anthropic and OpenAI usually publish prompting guidance when they release a new model. You can learn a lot from that. And if they open-source things, like OpenAI's Codex, you can study their prompt structure. People should do that more. Then there's Twitter. I opened it today, and the first tweet was from OpenAI announcing o3. I might or might not have an email from OpenAI about it, but I saw it there first. If you use it intentionally, engage with the content you actually want to keep seeing, your timeline changes. It really does pay off.

How do you define ‘good’ when building evals for compliance?

Vanta employs GRC professionals—former auditors, heads of security, CISOs—as part of our workforce. So what ‘good’ looks like is a mix of things. There’s tone and verbiage, which matters for something like the Vanta agent, but also deep technical correctness. We leverage those experts to help us annotate information, build evaluators, and get to quality. We do a lot of that work with what we call in-house GRC subject matter experts.

The other thing is data reviews. Once we think a feature is good enough, we put it in front of our early access testers, watch the data come in, and look at it. When we don't understand the nuances, we bring in the experts. Over time, more and more things become clear before we need to escalate, so we invoke those experts only where needed. And the ‘good’ bar varies widely depending on the task. The risk of getting it wrong is different depending on what you're doing. If you're telling a customer they're ready for an audit, that bar is a lot higher than, say, helping them connect an integration.

Who are the early access testers?

The first thing is our own Vanta instance, used in-house. After that, we have a group of customers who have volunteered. Think of them as AI-forward companies who want to try things as soon as they're ready, and they're comfortable with rough edges. It's slightly different from a design partnership, which is a bigger commitment, more of a formal feedback loop, and more intensive. Early access is a lighter-touch process. Right now, we have around 20 companies in that group. In some cases, we can only release a specific thing to a few of them because they match the target use case, but we try to keep in very close contact. We integrate the account, we ping them directly, they book time with the team, we jump on calls, and go through it together. After we've validated with those early access customers, we move to a broader rollout, from 1% to 25%. And we're not afraid to roll back. If we ship something and the quality isn't what we want, we roll it back fast. Looking at the data is very, very important.

Source: Vanta.

Your eval datasets need to evolve over time. Not just because you find gaps, but because user expectations drift. As a user of any AI product, what felt impressive six months ago is now your baseline. The same happens with your product and, because the input space is so broad, the types of data you see and the intentions behind them shift too. Keeping your evals current isn't about keeping the lights on. It's about keeping your feature actually good.

What carried over from your four years at Google working on LLM tooling before joining Vanta?

Google is a very special company, very different in good and bad ways. What carries forward for me is the importance of quality and its data-driven basis. A very Google way of looking at problems is to look at the data, understand the data, and have evaluators. Google Search has had AI in it for maybe 15 years (a stack of ML models, and now clearly LLMs too). That data and quality orientation are core to how I think. If your feature isn't good quality, if you're not looking at the data, then in my opinion, you shouldn't have the feature. Delete it. That might be a spicy take, but I do believe it.

Some things don't carry over. Google has a lot more resources, so you can throw money at problems in ways that don't translate. The other big thing is latency. I worked in Google Search, which is a very specific part of the company. Latency-first, always. If you've used AI Overviews or AI Mode in Search, it's probably the fastest LLM product for the quality level you can see. That's very deliberate. Google's position is speed and intelligence, not one or the other, and that serves billions of users and billions of searches a day. At Vanta, the trade-offs are different, but I still have great respect for that standard. You adapt it to your problem and your customer.

What does your day-to-day look like?

A lot of meetings. That’s the sad story of my life right now. But turning that around, where I think I'm actually adding value, is in looking at the problems coming down the pipe and spending a lot of time thinking what the right approach is, given the state of technology today, whether it's a novel or a conservative approach. The other big thing is scale. When I joined, my team was three people. Now it's 45, and it's been about a year and a half. There are also more AI people spread across Vanta. So a lot of my time goes into thinking about how I can bring more engineers along.

Source.

I think this might be controversial, but the role of a product engineer is shifting. Every product engineer in the near future will need to do prompting and evals, just as another tool, to build products. It'll be like another framework, another front-end library. Not every problem will call for it, but it will be a common way of solving things. More and more people at Vanta need to get there, so I spend a decent amount of time thinking about how we can make that easier and bring people up to speed. One of my teams is the AI platform team. We build the foundations that everyone else builds AI features on.

What does your week look like in terms of cadences?

The rituals are per-project, so everything I describe applies at that level. We have a team-wide meeting on Mondays for general updates, ideas, and a chance for everyone to check in. Below that, each project or workstream has its own weekly meetings. Workstreams can cross team boundaries. They have the freedom to run things however they need to, but they must produce a specific progress report as output. Every AI feature has a data review. The frequency depends on the phase, but a lot of them start daily, then move to several times a week, and in some cases stay there or drop to once a week.

Source: Vanta.

We also have what we call a more open build session. You're working on something other people should know about, so you come and talk about it. Then bi-weekly demos where anyone can share a piece of code, an eval, a product, a design. Anything people will find interesting. And every other month, we do a hack day. The AI teams pick one or two topics, and before the day itself, we talk through different ideas, and people sign up to work on them. By the time hack day arrives, everyone already knows what they're building. We run a dedicated Slack channel, where each project has its own thread, and people post throughout the day. We usually do it on a Friday. On Monday, we come back and demo. We also create a Slack canvas with all the logs from the day and share it broadly. That format has become the way we validate approaches and ideas that are floating around.

Extra reading

• How Vanta Learned To Trust AI - June, 2026

• ‘Shadow AI’ is real. Vanta wants to help manage it - June, 2026

• Vanta’s Agentic Trust Platform redefines how enterprises earn, prove, and scale trust - November, 2025

• Tying Engineering Metrics to Business Metrics - November, 2024

And that’s it! You can keep up with Iccha and Ignacio on LinkedIn or check out Vanta on their website.

BRAIN FOOD 🧠 

TWEETS OF THE WEEK 🐣 

Source.

Source.

Source.

TOOLS WE RECOMMEND 🛠️

Every week, we highlight tools we like and those we actually use inside our business and give them an honest review. Today, we are highlighting Granola*—an AI-powered notepad that takes, summarizes, and organizes meeting notes without using intrusive recording bots.

beehiiv: We use beehiiv to send all of our newsletters. Apollo: We use Apollo to automate a large part of our 1.2M weekly outbound emails. Taplio: We use Taplio to grow and manage my online presence.

See the full set of tools we use inside of Athyna & Open Source CEO here.

HOW I CAN HELP 🥳

P.S. Want to work together?

1. Hiring global talent: If you’re hiring tech, business or ops talent and want to do it 80% less, check out my startup, Athyna. 🌏

2. See my tech stack: Find our suite of tools & resources for both this newsletter and Athyna here. 🧰 

3. Reach an audience of tech leaders: Advertise with us if you want to get in front of founders, investors and leaders in tech. 👀 

That’s it from me. See you next week, Doc 🫡  P.P.S. Let’s connect on LinkedIn and Twitter.

👋 Howdy to the 3,996 new legends who joined since our last edition! You are now part of a 468,789-strong tribe outperforming the competition together.

LATEST POSTS 📚

If you’re new, not yet a subscriber, or just plain missed it, here are some of our recent editions.

🗑️ The Future New Media: a16z. How the famous VC firm has thrown out the old playbook.
⚡️ How Vanta Learned To Trust AI. Wiring AI into the one product where AI isn't supposed to be trusted.
🦠 The Black Death & How It Birthed Today's Prosperity. How one little flea, on one little rat, changed the course of history.

PARTNERS 💫

This newsletter runs on beehiiv. So when we say the referral program works, it's not a guess, it's our actual growth engine. Readers share, hit milestones, earn rewards, and the list grows on autopilot. Pair that with their recommendation network and newsletters on beehiiv grow 2.75x faster. We're one of the data points. Worth a look if you're building something.

Get 30% off your first three months on beehiiv with the code: BILL30

Around 90% of AI projects fail because you don't have the right people. Aboard pairs powerful proprietary tools with senior engineers who’ve seen it all—and built most of it. We don't put AI in charge, we put it to work as a powerful aid in building reliable custom software for industry leaders in a fraction of the time.

Whether you’re a startup that needs to get to market or an enterprise with complex legacy challenges, Aboard delivers exactly what your business needs.

_{Interested in sponsoring these emails? See our partnership options here.}

HOUSEKEEPING 📨

I faced an interesting predicament today. It’s a brisk 15 degrees Celsius today in Melbourne (59 to your Farenheiters), and I needed to rug up to take our dog, Ziggy, for his morning walk. Once dressed, I threw on a beanie, only to be told that I would either look like a dork or a criminal, depending on how I wore it.

I have never been called a dork, and I have been known to delve into a touch of criminality in the past, so I decide to go full criminal. I bought a gun, just in case the opportunity arises to do an actual crime, but alas, no opportunity availed itself to me. The gun is a lie, of course, but the beanie slurs were not. Enjoy today’s piece!

INTERVIEW 🎙️

Ali Hussain, Co-Founder & CEO at Tabs

Ali Hussain is the Co-Founder and CEO of Tabs, a New York-based AI-native revenue platform that automates the full contract-to-cash lifecycle: billing, collections, revenue recognition, and reporting. The company has raised $91M+ in total funding, including a $55M Series B led by Lightspeed in September 2025 and a 5x increase in invoice volume, crossing $1B+ in annualized billing for over 300 customers, including Rogo, Cursor, and Cortex.

Ali.

Ali is one of those founders who came to the problem through genuine operational pain rather than a market map. He built the product category he needed but couldn't find: one that could read a contract the way a human does, understand what it commits the company to, and automate everything downstream from there. The unlock was AI, but the conviction behind Tabs predates it. What makes him interesting is the combination: the finance operator's instinct for where the real work happens, the builder's patience to wait for the technology to catch up, and an unusually clear view of where the agentic finance stack is headed and why legacy ERPs won't be the ones to take it there.

What's actually broken about how B2B companies manage revenue?

One thing is this fundamental belief that revenue is going to be predictable and consistent. The traditional wisdom was that you have a pricing strategy, a way you do business, a way you sell, and it's going to be consistent. That belief is just fundamentally broken. We are seeing immense amounts of variability and complexity, especially with AI right now. The standardization of revenue is something you've got to throw out the door.

Source: Tabs.

The way people price their products, the velocity of changes, how pricing gets contracted, the move towards usage and consumption models—all of it is shifting. The broken nature of the status quo is that we're still living with systems, processes, and in some cases people who are not ready to acknowledge that. On the other side, we're seeing new companies, especially AI companies, move fast in this new direction, but I've also been amazed by how quickly many legacy companies have adapted.

Why has this problem gone unsolved for so long?

It's really a data problem. When you think about revenue, pricing, or commercial agreements, at the end of the day, it's as creative as the two companies involved want it to be, based on negotiating and signing a contract. The reason it's historically been broken is that when humans codify something, whether in a legal document or just typing it up and signing it, the ability to take that data and administer it, to automate things like billing from it, has been essentially impossible until recently.

The only way you could take that paper and extract what mattered was to have other accountants and humans literally read and type that information. LLMs and AI have created technology that can actually go in, absorb the information, and understand it in a way that was just not plausible until the last 24 months. People always ask how accurate AI is. I say, “Can I guarantee 100% always? No. But I can guarantee it's better in 2026 than many humans are at extracting that information and coming to the right conclusion.”

What was the hardest part of the very first year of Tabs?

A few things. One is just that when you're building, getting really great companies to bet on you is hard. Revenue, accounting, billing, collections, RevRec—these are incredibly intimate, business-critical workflows that matter massively to companies. Convincing buyers and partners to be willing to bet on us was always the hardest part.

The second is that revenue as a category has a very large surface area. You have to extract contracts and read them, handle billing, collections, payments, reconciliations, a bunch of APIs, RevRec. Being able to build wide while knowing where to focus, not diluting the product's value, that level of prioritization while balancing such a big surface area is quite a hard problem.

When did you know Tabs had genuine product-market fit?

One of the cool parts of building a product is that you sometimes get to use it yourself. As we started signing our first 10, then 50 customers, I was the one using and administering the product every day, and I realized how much easier this was than how I'd done it before. People often look to customer signals or willingness to pay. For me, it was the product's actual efficacy. ‘Is this a hell-yes experience? Could I live without it as an actual user?’ That was the tipping point. For Tabs itself, we signed customers, and they wanted to use it. That created the whole value proposition.

Trusted by the best.

How did Lightspeed end up backing you twice?

At our seed, we got to meet a bunch of firms, and it was becoming very clear that Lightspeed was emerging as the top AI investor (they had invested in Anthropic around the same time), but a lot of it was also their FinTech group here in New York. Justin, who led the round, had spent a long time at Stripe and had thought deeply about this problem. He had a point of view on what a winner in this category would need to look like. It was a very clear meeting of minds on vision and philosophy. At the Series B, they had seen the business and watched us do what we said we would.

Source.

We solved a genuinely hard problem from an AI-first standpoint. If you look at Lightspeed's research lab investments, these are companies redefining categories through an AI-first lens. In the 12 months since the seed, we have delivered. I think that created real conviction for them to double down and let us stay heads-down building rather than be distracted by future fundraising.

How has the CEO job changed as Tabs has scaled?

In some ways, I've been really intentional about keeping a lot of the same things, staying as close to the customer as possible. Just because we have more people doesn't mean I've become a people manager. That's the least of what I want to do. I've tried to maintain roughly the same number of direct reports I had at 15 people as I do now at 150.

What hasn't changed much is the amount of time I spend with customers: selling the product, learning pain points, understanding how it's used, and where we need to go next. What has changed is that we have a lot more people, so creating focus and direction matters more. As a company hits this kind of scale, you start working with specialists rather than generalists, people who know their domain deeply. Getting to go deeper on specialization in areas like customer success and security has been a core part of how I spend my time now.

How is AI used inside Tabs beyond the product itself?

Almost everywhere. Everyone in the company has AI tools, whether Claude or Cursor. Engineering has always been the earliest to adopt, but broadly, marketing is using it tremendously for content creation, targeting, and automation. Our AEs and solutions teams are building demos and ROI presentations. It's been an incredible enabler for our post-sales teams, helping them deal with large volumes of data and move customers off legacy systems.

Long life to the Agents.

Personally, aside from using Tabs, most of my other workflows—Slack, email, everything—originate in an LLM. It's how I gather my thoughts, do research, communicate, draft emails and notes to my team, and send structured post-call summaries. Our product team uses Claude to triage Slack and build out roadmap prioritization. AI is integral to almost every part of how we operate. With humans, too: we're a fast-growing company, not a 10-person AI-only outfit, but everyone should be at least 5x what they could do without AI, while still having genuinely talented people here.

Where is the AI-native finance stack heading in the next three years?

A lot of the noise has been around how legacy systems become AI-native for automation, and I actually think that's the wrong frame. Finance teams are a combination of expertise (primarily humans), systems of record, and some type of services. Where I think it's headed over the next few years is that most organizations will still need experts at the top, but a lot of the systems and individuals driving the more tactical parts of finance will move over time to agentic workers.

Source: Tabs.

A big part of what we think about at Tabs is where the inefficiencies of services and human capital are buried. It's less about just revenue and more about the jobs that need doing, teams that are adding bodies to billing and collections, or scrambling to hire revenue accountants when an audit shows up. The ERP will shrink quite a bit. What's replacing it is specialized agents that finance leaders get to administer: one for all things billing, one for revenue accounting, one for payroll, and one for FP&A. Technologies that start to look and feel more like people than something you log into and use your keyboard. That hands-on-keyboard model is slowly starting to shift and will continue to be less relevant in the years ahead.

How do you build culture at a company moving this fast?

Without sugarcoating it, it requires hard work. We are an in-person, five-days-a-week culture. There's no alternative. There is immense personal sacrifice, and I think in 2026, you can't be competitive any other way, given the pace of change.

Proud team.

The second thing is empowerment. The company is only as good as the people who have the agency to run. Talent density is incredibly important. Third is direction: working hard but also smart, towards the right goals. Focusing deeply on what matters and not trying to do too many things at once. One bigger-picture thing we've done well is knowing where AI is going to be better built by others and not over-engineering towards where the models are today, knowing that in 12 to 18 months it'll look completely different. Balancing that is a huge enabler for moving faster and smarter.

How do you get the best out of yourself?

Sleep well; there's no alternative. Keeping that balance is really important. The second is being present in the moment, giving yourself the space with your team or with a customer to really take in what's going on. That means phones down. Living in the moment, you can't multitask.

That's one of the risks when you have an LLM, agents, Slack. You try to do so much because you've become highly optimized, but in the high-value moments—a customer meeting, a whiteboard session—you have to step away from reactive mode and into deep thinking. That's something I've really tried to change about how I spend time with my team, customers, and partners. Really taking those moments in, thinking deeply, and not just creating motion out of it but making sure you're using your time on the highest-value things.

Extra reading

• How AI is transforming company revenue in 2026 - May, 2026

• How Ali Hussain Created and Scaled Tabs - August, 2025

• The Startup Revolutionizing B2B Revenue Automation - February, 2025

• Building with Tabs - November, 2024

And that’s it! You can follow Ali on LinkedIn or check out Tabs on their website to keep up with what they’re building!