Cities are filled with so many interesting one-time food pop-ups—but only if you know where to discover them. These events, often only lasting a couple of hours, can be very hard to find, with only a couple Instagram posts informing people of them. I’m always interested in trying unique food, and set out to find a way to reliably discover such events without too much hassle. While everything could still be a lot easier to discover, I found it interesting to explore the platforms attempting to make people aware of all of the food events going on around them.

What interests me so much about food pop-ups is the uniqueness. Oftentimes, restaurants or other venues will host chefs who are coming from afar, and you most likely won’t be able to try the food item again in your city. Also, if the venue is hosting the chef, they must be pretty good, or else why would the venue bother? The randomness of many pop-ups delight me. Oftentimes, pop-ups are in the most unpredictable locations, which makes it all the more fun.

The first and most obvious way is Instagram’s built-in search. If you put in the date, the word “pop-up,” and the place you’re looking, you’ll probably find one or two good search results, and they’ll probably be towards the top. Unfortunately, Instagram’s search often surfaces events from years prior, or things in a completely different place (in other words, you might find things from 2016, or things that are happening in Australia when you’re looking for New York). It leaves a lot to be desired, and while it may give a couple of results that other platforms don’t, they’re usually either not good ones, or, if they are, they require sifting through lots of irrelevant results to find.

Another helpful method is newsletters. Many neighborhoods/cities have newsletters with events. If you look hard enough, you might be able to find some interesting food events that you wouldn’t be able to find elsewhere. A good place to start is to search Substack for newsletters that cover your area and that publish weekly digests of events. However, these newsletters aren’t necessarily designed for food events, and require filtering out unrelated things that aren’t exactly the target goal.

The most useful of my techniques, by far, has been Mato. Mato is a curated event platform, accessible on the web, and available in many large cities worldwide. It is curated by (presumably) a team, which looks at Instagram events and adds them, along with details, to the platform. Most events it surfaces are at least worth a look, and it has a much higher signal-to-noise ratio. Mato has helped me find some really interesting food events—however, it finds events in all different categories as well, so I highly recommend you try it!

Today, I visited Vietnamese restaurant Ði Ǎn Ði for an onigiri (Japanese rice ball) pop-up collaboration with Amenohi in Greenpoint, Brooklyn. The meal was delicious, and reminded me of how much fun it is to go to food pop-ups, but how they’re so hard to find. If they haven’t been widely publicized, they’re not going to get very many people to show up (that wasn’t the first onigiri pop-up I’d been to). That one I found on Mato.

Onigiri at Ði Ǎn Ði. Source: @diandi.nyc

After lunch, I stopped by vintage furniture store Renewfinds, which was hosting Uju Studios, an Korean-inspired pâtisserie, for a weekend-pop-up, and ate a tasty mango sticky rice-rice krispie treat. That one I happened upon by a chance Google search, which directed me to an NYC cafe-events site.

Pastries at Renewfinds. Source: @uju.studios

I’m thinking about starting a dedicated blog for food pop-ups, or making it a regular section on this blog. Would you be interested in that? Let me know!

I recently spoke with Cyrus Freshman, developer of web apps such as Minnow, the Indie Internet Index, and Lettercomb (all of which deserve to be checked out!) At the time of the interview, Cyrus was working independently; he now works for Wayfinder Collective, building a sales operations platform. The interview has been lightly edited for clarity.

Micah Blachman: What drove you to create independent projects at first?

Cyrus Freshman: I was working at Amazon Robotics on the path-planning team doing algorithmic cloud software things. I was there for about two and a half years, and I had saved up some money. Amazon Robotics was already functioning, I wasn't adding that much, and it felt like a good time to try to build out as much of my own software as I could. When I left, I had one app, a word strategy game called Wordbase. It had been discontinued, I remade it for the web, and it’s still one of the most popular things on my website. People sign up every couple of days for Wordbase. That was all I had when I left.

Another one of my games, Lettercomb, was also something that had been discontinued, and I remade it. Then, they actually re-released it, so I reached out, and they said it was okay, but that I just can’t use the real name—which I had been doing—so I changed it. Both of those games were ones that I played in high school and really enjoyed. Even those massive apps were going offline or getting discontinued, because it's just so hard to monetize.

MB: What do you think is the overarching theme behind all of your apps?

CF: I’m trying to rebuild the internet. With Instagram, Snapchat, TikTok, when you’re on your phone, you go to an app, which is killing the internet. I've been trying to build web apps that work both on mobile and desktop, that sort of rebuild the internet in a way.

“I’m trying to rebuild the internet.”

I’m just trying to use the internet as it was intended. I have a Twitter-type microblogging social network, and a media upload service, and I don't let you upload things into the microblogging app. I make you put your uploads in the media uploader that I made, and then you link to it. The internet is mostly about linking to things. When you upload your photos to Instagram, you don't actually get a link for them. They're storing your image, but you can't link to it from anywhere else. You can't link to it on your website. They've kind of held your content hostage, so I'm just trying to build things that I feel like are missing. I try to spread out what I’m building into different areas.

MB: Given that you’re trying to make the web more open, how do you feel about making your apps open-source versus keeping them closed-source?

CF: At one point, I was making them all open-source, and I just closed them, because I'm a solo developer. Security-wise, it's just adding more surface area for attacks. If you're a team of people, and you have an app, it would be ideal if it's open-source. If you're one person, and you're building a coding library, open-source that. It just started to feel too sketchy—I'm sure I have some security hole in some app, and I don't want the source code to be out there to be so easily findable. I have been going through and fixing things up now that I can use AI coding tools—I was doing the development all before AI coding tools. I have gone back and asked, “Are there any security flaws with this?”

I discovered that I was building my link preview tool too simply, and that there was a flaw in it. The AI coding tools are still a little oversensitive, but I'm trying to do as best as I can with security. I did get a degree in computer science, so I have a little bit of security training. My biggest worry with open-source is that it’s opening myself up in a way that it doesn’t have to be.

I have some small libraries that are open source, and I maybe have a couple apps that are still open source. No one is opening pull requests on them, and that’s really the benefit of open-source code. The biggest interactions I get on GitHub are for the Wordle word lists I published. People will comment saying it isn’t up to date, and I’ll say that The New York Times changed it.

Ideally, there is a good open-source community, some apps aren't the best fit for that. For me, making open-source changes is something that is hard to get into because there's such a high bar when you're making a proposal, in my mind—or maybe it's easier than I think it is. I’ve always been more interested in making my own apps.

MB: How did you get into creating web apps?

CF: Web apps are really accessible. You don't need an Apple developer account, you don't need to go through App Store review processes, You can very quickly come up with an idea and publish it. I find that really ideal when I'm publishing an app so I've just continued to make web apps. I have made some iOS or Android apps, I'm mostly just making web apps.

MB: How did you build Minnow?

CF: I used to write apps myself. These days, I mostly use AI coding tools, because I can make an app ten times as quickly, and it's two or three times as good. I typically start out with a master rule set, a document that I paste into ChatGPT, which I’m using these days. I have used Cursor in the past, but it just gets really expensive. If you are spending the whole day coding in Cursor, you can spend $100 a day, which is too much. I use ChatGPT, I paste this document in that basically tells it how to bootstrap the project. I then test the basic functionality of the app, see if it’s a good start, and go from there.

MB: What was your inspiration?

CF: Nowadays especially, everyone should have a personal website. You have a computer, and you're not really fully utilizing it if you're not coding with it, and getting what you want out into the world, via the internet—if there is anything, I know, some people don't want to do that at all. 

For Minnow, the idea was, especially with LLMs that can code for you, it's really easy to make a personal website now. The only thing is that while ChatGPT or other LLMs will make you a website, it won’t publish it for you. Minnow puts those elements together: it codes the website, and publishes it on a subdomain of the main site.

MB: Where do you see the trade-offs of AI coding?

CF: If you're in a corporation, and you’re making very small, incremental changes, then hand-coding is great. Even then, you probably still want to be using AI coding tools to get ideas for how to code stuff, and then still go and hand-code it. But if you're a solo developer, and you're building web apps, which aren't mission-critical systems, if you have pushback against AI coding at this point, you're maybe not really seeing it for what it is. Even if you don't want it writing all the code for you, it's still useful for thinking through how you'll build stuff—it’s an incredible tool.

MB: From the user’s point of view, do you think there’s a difference between a hand-coded and an AI-coded app?

CF: It's more about how much effort the developer put into the tool. If someone spent one hour hand-coding versus one hour AI coding, they're probably going to get further with the AI coding app. It'll probably still be better than a one-hour hand-coded app, but it's still not going to be as good as if they put a week into the app. It’s about how much craft and skill the developer has, so, ideally, you shouldn’t be able to tell.

MB: How do you feel it is to try to get your name out there and grow your user base with so much AI-generated software out there?

CF: That's where I'm weakest. I feel like my apps are pretty good, but I still don't really know how to get my apps or name out there? I've gotten around 200K views on my site, but that’s translated to maybe 50 to 100 persistent users. I'm not currently planning to continue doing this full-time, I'm trying to go back to a company. It's still great, I'm still really happy that I spent this time making these apps, just in case they could grow while I'm doing other things. Minnow got three likes on Hacker News, but two weeks prior, I had built the Indie Internet Index, and that got to number 7 on Hacker News. The III is a directory of user-submitted websites with descriptions, and then semantic search over those descriptions to provide a cool indie internet search engine.

It's very hard to get upvotes. I got upvotes because users submit their own websites, so everyone wanted to submit their website, and then they would upvote the posts. It’s hard to engineer a mechanic like that, but it worked.

MB: How does the software you use inspire the software you make?

CF: There are a couple of different directions. There's the websites that I think back to, from when I was a kid, and the websites that made exploring the internet really fun. That's the sort of feeling that I'm trying to recreate, almost. 

Then there's apps like Twitter or Threads. Microblogging platforms are a really great software type, and photo-sharing apps like Instagram are still a really good software type. I'm also sort of inspired by great design in general. As far as my own design goes I’ll hone in on a very specific description or niche that something is going to inhabit, and then I find that really drives the design decisions, and I can take that the whole way through the app. The app will make sense, and it’ll feel correct.

There are also apps I see in the real world that inspire me by being bad in some way, that have scattered design decisions that annoy me. I take that feeling and then I go try to build very polished apps.

Earlier this week, on my family’s trip to Charleston, South Carolina, I ate out at a somewhat upscale modern Filipino restaurant, Kultura.

While there, I saw something increasingly rare: the waiter taking down the order on a physical pad of paper. It feels, nowadays, that restaurants are, more and more, replacing the reliable pad and pen with a handheld mobile device. I suppose it makes it a little bit faster for the order to get to the kitchen, but having a waiter write down the order on a pad of paper just feels so much more… human.

And what if you decide to change something about your order and your waiter has already digitally sent it to the kitchen? Is it actually easier to change than just crossing something out on a piece of paper? Reddit reveals that waiters aren’t so keen on the technology. Point-of-sale programs like Toast (the most popular choice) require the admins (usually the owners or managers) to update the interface with built-in options for common modifications and things like that. Problem is, when the owners are not so tech-savvy, modifications need to be typed in on what is effectively a smartphone keyboard. At some restaurants, the waitstaff have bonded together to use pen and paper, then enter the orders into a POS (point of sale) in a shared location.

This sort of technology use in restaurants I don’t mind. Having a tablet at the host station to manage reservations is perfectly fine—I don’t think anyone’s yearning for physical reservation notebooks again (before my time!). But what I do care about is the honesty and simplicity behind the pad and pen. When the waiter writes down the order on a pad, you can see what they’re doing, and they can make eye contact with you (it’s not too hard to write without looking). Using a handheld POS means that they’re looking down and there’s no way for you to know what they’re doing (for all you know, they’re writing an essay about how bland your palate is in the modifications section).

Having the POS in a central location in the restaurant, like many waiters have informally adopted, feels like a better solution to me. There’s no risk of a paper being lost, and cooks get an understandable, maybe even more accurate, order behind the curtain. But for the customer, even if they’re missing out on getting an order—that may or may not be accurate—one minute faster, they’re getting the authentic humanity that comes with a genuine connection to the waiter, instead of experiencing the waiter as an intermediary.

It’s part of what makes me want to do my trivial computer tasks myself, not outsource to AI or an automation. I don’t want to feel like the middleman. Even if it means that I need to spend a couple more minutes responding to emails myself rather than having the machine do it for me, I’ll accept the tradeoff. In order to have the momentum and energy to create the masterpieces we perceive as unable to be outsourced to AI, we need to get through scheduling those meetings first—and that doesn’t feel like such a big ask.

If we’re outsourcing taking the orders, when will we start outsourcing the cooking? And if we’re outsourcing the menial tasks, when will we start outsourcing the big ones? If the restaurant is just a robot cooking, anyone can get that robot. If you are just prompting the robot to do work, others can prompt the same robot too. I’m not sure if I want to eat in the robot restaurant, especially when the RAM prices keep driving the lamb prices up, too.

Theoretically, file sharing has never been easier. We have at our disposal tools like Dropbox and Google Drive to send anything from pitch decks to cat videos. But, with storage limits and other annoyances, I feel like these programs end up wasting more time than saving it. Especially if you’re only sharing files between your own devices and don’t want to use cloud storage or your own network-attached storage (NAS), what about good old USB drives?

It’s easy to forget about USB drives these days, but they’re private, secure, physical, and give at least an illusion of control over your data. Even if you have no power or internet, assuming your devices are charged/don’t require power/you have a backup power supply (such as a small uninterruptible power supply—UPS), you can still access your files. Can Google Drive let you do that?

As with many unassuming pieces of technology, this storage medium has become sort of trendy recently. In this post, I wanted to explore a few ways that USB drives have taken over the world occupying the gap between tech and art—everything from a file-sharing club to the presentation of a new typeface.

USB Club

Originally created by one of my favorite teams, the Brooklyn-based Teal Process & Company, and now owned by creative collective garden3d, USB Club functions as a sort of Are.na-esque social network. A USB-based social network? Files on your laptop that you’d like to share, or more importantly, that you feel are worth preserving, can go onto your USB drive to then be synced with the USB Club network—and shared. Once you’ve uploaded a file on a particular day, you can view files that others in the community have synced on the same day.

In order to join the club (I am not a member, but it seems cool!), you purchase a USB drive—either the “Classic,” a 128-gigabyte drive ($40), or the “Transport,” a luxury one-terabyte drive ($180!). Alternatively, you can join their waitlist if you’ve already bought a drive.

Image from Designheads Substack.

PUFF

In a launch event at Index Chinatown in Manhattan last weekend, hosted by none other than USB Club’s garden3d, PUFF by Noware was released to the world. PUFF describes itself as a “designer flash drive” intended “to be worn as a pendant, bag charm, or keychain.” When you add files, play music, or move around, the OLED screen lights up and reacts to these movements. While I’m glad I didn’t go to the launch event (they thank their marijuana provider in a blog post), the product feels interesting, in the sense that a flash drive can be as luxury as you want it to be (if you want to pay $179 for it). However, in comparison to USB Club’s Transport, which, for a dollar more, offers almost eight times the storage, it’s clear all you’re paying for is the screen.

Image by Noware.

USB Locket

Ever feel like you need a closer connection to important files? Not me, but probably someone. Especially if you’ve gotten this far down in a post about USB drives 😃 

USB Locket, the brainchild of Manila-based, apparently pork-adobo-loving Alberto Cinco Jr. (he and I appear to share a passion for the NYT mini crossword), is as simple as it sounds: a USB drive attached to a pendant, carrying files important to you. The essay he writes about why he created the project encapsulates a lot of the thought process I had in writing this post. I won’t put it all here, but I’ll quote a particularly interesting section below:

On another webpage, he writes instructions for how to create your own. Creating your own physical USB drive is something completely foreign to the worlds of USB Club and PUFF. Both require you to buy an expensive product in order to participate in the experience; yet personal memories can be kept and savored without a high price, and in whichever way the wearer wants.

Image by Alberto Cinco, Jr.

I’d like to offer one final example, albeit not yet released.

Sleek by Mikey Okei

A type designer I met on typo.social, the Mastodon instance for type designers or those interested in type design like me, Mikey Okei releases type from Ho Chi Minh City (definitely check out his catalog!). For his new typeface, Sleek, he decided to realize his vision of releasing a typeface physically. For this project, he’s creating a physical zine that people can purchase. The font file will be inside, on—you guessed it—a USB drive. I, too, like seeing typefaces physically; seeing them online just doesn’t feel as satisfying. While digital type has brought much good into the design world, and the world at large, there’s a certain element missing from it that can only be found when type is printed. With the USB included, you get the best of both worlds. I’m excited to see the zine come out: in the meantime, here are some pictures!

Image by Mikey Okei.

Image by Mikey Okei.

Image by Mikey Okei.

Imagine: 900 words about a technology you’ve probably never given much thought to, mindlessly plugging into your laptop every time you want that design brief or random TXT file. But now you are Undeniably Somewhat Brighter about those small storage drives that are so easy to forget.

AI executives are often claiming that their product will replace [role] in [amount of time]. More often than not, these predictions are false, or at least overly optimistic. But what if people from a decade or so ago were actually able to predict the future?

That’s the question that guided me as I explored the thoughts of Microsoft researchers, an Oculus forum member, and a Métis Canadian entrepreneur named Matthew Unger. These three articles attempted to describe 2026 in terms of how we’ll interact with technology.

I find it fascinating to explore how individuals from previous decades, centuries, or even millennia felt about the future, and how it would differ from their present time. Yet more interesting is seeing how their predictions actually play out; the places in which they were wrong, the places in which they were right, and the places in which they should have been right—but that just wasn’t the way that things played out.

That being said, so much hype is generated by the tech industry, which, whether genuine or not, can cause people’s opinions to be skewed one direction or another in ways that don’t always line up with reality when we look at the big picture. In a few years (it’s gotten to me, now I’m predicting too!), once AI is less of a buzzword and fades into the mundanity of technology—like Wi-Fi—we’re not going to go around making predictions about it. The first prediction I will look at was made in 2014 (when I was around 1) by an Oculus forum member. This community was obviously thinking about VR, and likely picking up on a lot of the hype around VR from around that time. Virtual reality still finds its way into tech news, but AI has overshadowed a lot of VR developments.

As each prediction has a large amount to unpack, I’ll be splitting this into a three-part series, beginning today with part 1.

Chaoss’s Predictions for the Future of VR and More

As it is now May, it feels appropriate to unpack some of the claims for “early 2026.” Below, I’ll be commenting on each element of the post.

I wouldn’t quite say that computing is invisible yet. Smartphones are, arguably, the dominant form of computing, and are certainly not invisible, nor are laptops. Smart glasses, which are not as easy to see, are nowhere near a mainstream product yet, but I expect more people to have smart glasses in a few years with where the industry’s heading right now. Smartwatches, which have entered the mainstream, are more invisible than a lot of other devices, but I don’t think anyone would call them invisible.

Funny enough, transferring data between devices is still very frustrating and tedious. AirDrop existed at the time Chaoss wrote the post, so I don’t think that’s the intention. On the topic of fridges, some smart fridges have the capability to track inventory—but they’re not ordering food for people.

Internet is certainly embedded into everything now. As the post predicted, people, cars, household items, as well as pacemakers have internet connections. For better or worse, everything sends information back to a server nowadays, and there’s often no way around it (printers are a particularly frustrating example). Starlink is enabling people in remote areas to have a fast internet speed. However, internet speeds are still measured in Mbps, or Gbps, and internet service providers are prevalent.

Certainly not all computing has moved to the cloud. That being said, cloud gaming has been popularized over the last few years, with Nvidia’s GeForce Now leading the charge, as well as Xbox (sorry—XBOX?), Playstation, and Amazon all offering cloud gaming. Another killed-off Google product, much like Google Reader, Stadia offered cloud gaming as well as a controller, and was sunsetted in 2023. As Chaoss predicted accurately, everything does run better locally (assuming you have a decent computer).¹

The forum post’s next paragraph is not quite how things played out. VR and AR do not play that big of a role in today’s ecosystem, and I would argue that AI is the more visible side of computing progress, rather than virtual or augmented reality. VR can be very life-like, but it is not mainstream yet—though mind control is becoming a reality with Neuralink and others paving the way. Resolution is still an issue, and virtual worlds do cover your entire frame of view. I can’t speak to lags or bugs, as I have never tried the Apple Vision Pro or similar headsets (the closest I’ve ever used to those is an old Oculus Quest headset, and I was not blown away by the performance).

Augmented reality is not very popular yet, and certainly doesn’t earn the post’s analogy that AR glasses for the current generation are like smartphones for the previous generation. I don’t think that creating virtual worlds now is quite at the ease and scale that creating a website was in 2015, nor are virtual worlds very widespread. Search engines have become a sort of “hub”—not quite in the way the author imagined, but with AI transforming the way that search engines work, instead of results, giving summaries of information.

Archiving/preserving the web is given a moderate importance, mostly with the Internet Archive. Interestingly enough, working in virtual worlds, in a sense, has been realized. As a result of the pandemic, remote work became popular, and while creating objects in virtual reality does not yet prove a successful occupation, teaching online can be.

Keyboards and mice are still very much in use, and while in recent months we have seen a rise in the use of dictation tools such as Wispr Flow and Monologue, brain readers have not seen use among individuals without neurological disabilities such as ALS.

I recently had the opportunity to correspond with Adam Clegg (also known by his handle Arclegger) about his new project. Adam is the creator of For You, a platform in which you generate AI art, and let it float down a “river” for others to pick up. If they like what you’ve created, they can add it to their collection.

Micah Blachman: What was your process in creating For You?

Adam Clegg: I wanted to both start a project and actually finish it, and create this idea I had in my head since 2021. I'm untested, but I'm fairly sure that I have some level of ADHD/Spectrum/Neurodivergence, some cocktail of all 3 that makes me different from most people. Because of that, it's always been really easy for me to come up with ideas, and it's really hard for me to finish them. I've started more unfinished projects than I can remember, so when I started to see the potential in AI coding grow, I started to get more and more antsy about wanting to actually make something with it. So I sat down and made a promise to myself that I was going to start this project, and I was going to finish it. When I started I didn't really know how it was going to turn out—I just had a vague idea in my head, an image of a floating river and packages floating down it. I knew that I wanted people to make art and give it away, and to receive art from others who are doing the same thing. So I started there and just kept going and going.

A developer view of For You, in its most recent iteration. (Image courtesy of Adam Clegg)

Luckily I've been a game designer and game dev since 2007, so I was working with almost 2 decades of games-making experience and live service. I already knew what it was missing as I was building it, so as it was coming together I kind of knew what needed to be attached to it to make it work. The original art I had made for it was pretty funny and rudimentary—compared to where it ended up it looks better, but also it could look even better if I had a background in art, or had an artist helping me.

The first art Adam created for For You. (Image courtesy of Adam Clegg)

My Thoughts

I thought I’d offer my own thoughts on For You, after playing with it for a couple of days. The text below is my own.

I personally find it difficult to feel ownership and pride over any AI images I might generate. While I may have prompted them, I didn't create them. The same with others' generations: I can find them aesthetically pleasing, but knowing that a human actually drew something gives it so much more meaning.

I think the river concept is a really calming alternative to algorithms and feeds, because you only see the things you click on, and each one is a surprise. But after a while, it's hard to feel excited about AI-generated art—you realize it's all kind of the same. With so much AI-generated content out there, human-created things are getting so much harder to discover. The river is perfect for this, because it gives you the discovery without the stress. Though I would love to be able to share writing, or other forms of media—not AI-generated—and let it float down the river for people to collect.

Perhaps the right way to do this would be to create a sister project of For You for any human-created content—whether that’s art, photos, videos, writing, or anything else. When you stumble across a gem actually made by a person, it's all the more meaningful. Plus, it makes paying for content on For You a lot more sensible—I still don't totally understand the motivation to buy AI-generated art (though I admit I have been experimenting with pricing my own art on the river). Why should the prompter benefit from the money when they didn't create it?

For You in alpha, with art looking closer to its current state. (Image courtesy of Adam Clegg)

A Note to Readers

While I have been doing more (fascinating!) interviews recently, I do not intend for this newsletter to be exclusively interview content. I’m pondering ways to separate the interviews I do from my own musings and ideas. For now, it’s a mix, but that may evolve.

On another note, I also intend to begin posting more regularly again. It’s been a few weeks since my last post, and I want regularity just as much as you do (hopefully, you want regularity, then). Armed with lots of interviews, I’ll be attempting to resume a weekly posting schedule, at a minimum.

If there’s something you want to see more or less of, don’t hesitate to let me know. As Marcin Wichary of Unsung (a fantastic software blog) put in his call for feedback a few weeks ago, “If the very idea stresses you out, I want to give you permission to send me just your bit of feedback without any greetings, or small talk, or ‘compliment sandwiching.’”

I recently spoke with Kristján Pétursson, the creator of a new RSS tool called Sponder and developer at Haha Moment. Short for transponder, Sponder allows you to create custom feeds tailored to your specific preferences that you can then import into your RSS reader (you should check it out!). The interview has been lightly edited for clarity.

Micah Blachman: What was your inspiration behind Sponder?

Kristján Pétursson: It started as a different idea: a long time ago, a friend of mine and I had this idea of an RSS social network but where everything is purely feeds. He actually texted me this morning, he forgot I was building Sponder and he’s building a similar thing now, too.

But the idea was that you subscribe to whatever you want. If you like it, you can publish it to your outgoing feed of things you find interesting. Then you could also publish things to your friends, like inbox feeds. So there would be a “things Kristján thought were interesting” feed and a “things Kristján thought Micah might like” feed. So I could publish in general or just to you.

I started intending to build that, but I realized that to get that working well, you needed somebody to be using a client for that product, and I didn’t want to get into a fight with everybody’s favorite RSS readers. People like what they like, I like what I like, and it felt like a very small product to say, “Abandon everything that you're doing now and come use this instead.” But next to that I had something I wanted to do. Most things I build are just because I want something and then I think that maybe another person wants this too.

There’s a blog I read called Bits About Money that’s about the intersection between tech and financial infrastructure. It’s fascinating in a way that I never want to actually have to do myself, but it’s fun to read about. I wanted to read its archive, but RSS feeds don’t do archives. They do what’s happening right now. I needed this thing that would put together the archive with the live feed, and then I could replay history at a compressed rate because this blog started 15 years ago and publishes once every month or two. I figured I’d read maybe an article every week or two until I caught up.

I thought that kind of tool was extremely specific and hard to build because I have to do web scraping. And so I thought, “Okay, let’s just get the system set up first to do some basic filtering.” Maybe other people can use that. That was actually the first version. It was just an extremely simple filter and the ability to make an RSS feed.

Then I started building in the things that I actually wanted that were harder to do—all of this difficult stuff under the hood that hopefully looks very simple on the front-end. Along the way, my friend mentioned, “Oh yeah, I wish I could skip reruns of podcasts.”

I thought, “Podcasts are RSS, I can do that.” I started putting on all these large features for what was, in theory, a small thing. But now Sponder does that! Once it had that feature, I felt like it was ready to push out onto Hacker News and other places because it was more than you could do with a simple extension to an existing system like FreshRSS. The rerun filtering involves a ton of historical analysis and statistics about what phrases tend to appear in titles when it‘s a rerun versus when it’s not. It felt bigger than somebody who just built something in an afternoon for their existing reader at that point, so I said, “Okay, I’ll give it to everybody else and see if they like it.” I’m still trying to get the word out there, though.

MB: I’m curious to know what made you decide to not have Sponder be an RSS reader with added capabilities?

KP: It’s mostly that I didn’t want to start that fight. It’s hard enough to get people to come and use something new. It’s even harder to get them to replace something they already have, whether or not they particularly like it. I’ve been meaning to replace Feedly for ages after they started cramming all their stuff into it.

But then, I know it’s going to be a 15-minute process where I download the OPML [a file format RSS readers use to exchange data] and then upload into my new reader. I know it’s not going to be hard, but I just haven’t wanted to bother spending those 15 minutes. And that’s a thing that I actively don't like using anymore. I just didn’t want to have to convince people to replace something even that they didn't like, but certainly something that they do like—if they’re already using FreshRSS or Feeder or whatever they’re happy with.

There are lots of good clients out there already. There are really good podcast clients all out there already too, and I really like when things are composable and just do their one job. It’s like the old Unix philosophy. I thought, “Okay, I see all these clients, whether they’re for RSS or podcasts or whatever, and they all have different features, right? They all organize your list in different ways. They all do different filtering. They all have their different stuff they chose to approach. But none of them do the entire gamut of every feature you could want.” But because it’s all just an XML file under the hood, Sponder can do whatever you want. Plus, you can pull from your sources that already exist. You can use it in your client that you already love. And it feels like an easier lift for me, and substantially less competition with things that are already out there and good.

MB: Where do you see Sponder as going in the future? Who do you feel is the target audience?

KP: It’s changing as I find new use cases, and I think it’ll change more as I try to introduce it to more people. You can probably tell from the demo UI—the fact that you can switch to a YAML mode [YAML is a human-readable format developers use to structure and organize data] if you just want to type code, I made this for technical people. I made this for myself and thought, Okay, maybe I can find a hundred other people like me.” I started to see that perhaps it could be broader pretty soon, though.

I have a bunch of friends who are doctors—my wife’s a doctor, and so I stole a handful of her friends and I play games with them pretty frequently. One of them mentioned that they use RSS to get their medical journal publications. But they’ll get a quarterly dump of 250 articles and probably care about 15 of them because they're doing cardiac electrophysiology. They don’t need all of the cardiology journal or even all of the general medicine journal, they just want electrophysiology. So I thought I could save doctors some time. They have enough on their plate, maybe I can filter for them.

When my friend mentioned podcasts with reruns and cross-promotions, things you didn’t quite sign up for, I thought I could do the analysis there, that’s an extremely broad thing. As much as I built Sponder to be a general and relatively technical tool, I feel if it gets any mass adoption—and by mass maybe more than a thousand people—the skipping podcast reruns sounds like maybe the most likely path to that. In the next phase or so I am going to try to make that good, easy, robust, and work obviously instead of being as finicky like it is now, because I could talk to just some normal human being, right? Normal human beings don’t want to come type YAML.

I could say, “Hey, are you tired of your podcast reruns?”

They might say, “Yeah, they’re kind of annoying.”

I’ll say, “For $2 a month, I can solve that for you. Put in the URL here, sign up. Here’s a new URL. You're finished.”

That seems like the best wide adoption path. I’ll be pleased if I can be useful for a hundred or two hundred people. I don’t expect it to be larger.

MB: How do you see Sponder as paving the way for future projects you might do with Haha Moment?

KP: The Sponder code itself took roughly three weeks to put together, and then added-on features like the hosting aspect turned into probably two months. But overall, it’s been about three months, because then you have to go set up the entire infrastructure. I put it on Google Cloud because they have a lot of good free tiers. Before Sponder is making me any money, I would like to avoid paying money just to keep it sustainable. But to do all that gave me all the tools I need to know how to push the link up to Google Cloud. Just having all that infrastructure work available to me now will be handy for the next things I do, and I expect them to go faster.

MB: What do you think people are getting wrong with RSS, and not really understanding that you think you could help people with through Sponder?

KP: That’s a cool kind of advocacy and education question. I don’t think people get anything wrong. I think they don’t know RSS is there. RSS was great between 2000 and 2010. Then Facebook, Twitter, Instagram, TikTok, and all of the current social media came and took over.

I think one of the big things that caused that shift is that RSS doesn’t natively have a shared nature, the actual networking part of it where you go to one spot and you see all the things your friends are doing and it grows from there. With RSS, you have to go to a site, wonder if it has an RSS feed, find the RSS feed, and put it your reader.

It’s just a bunch of steps that the average person or even probably the modal person doesn’t know is even an option. Just having people use RSS more is a massive outreach and education problem to solve.

MB: Do you think RSS is still thriving? From projects I’ve seen online, it seems like there are still people actively developing new RSS tools, like Sponder. But from your perspective, how do you see other people using it?

KP: I don’t feel like I have great transparency into what’s going on. From what I see, there seem to be two camps.

First, there are people probably like you and me, who subscribe to anywhere from 10 to 200 feeds and read stuff as it comes through whether or not they’re being completionist about what they read or just using it as a stream that goes by.

Second, there seems to be the more enterprise side of things where feeds roll up publicly available data and present it to a team that very much cares about it. For the security team in a large enterprise organization whose job it is to be on top of this stuff and react to it fast, for example, RSS is a good medium to push information.

And then, between the niche reader and the enterprise team, there doesn’t seem to be a middle ground. That middle ground has been eaten by more modern social media. I don’t know if it can ever really come back from, or if it will always be this almost-nostalgic technology, the way people look at having a fancy car with a manual transmission, or riding expensive and fast race horses. We know that horses are not a part of the modern infrastructure. We invented engines and now horses are gone. So why do we even still have horses? Just because people like them and people race them.

It feels to me like that's where RSS is. People like it, but most people don’t think about it. I expect that to continue to be true.

MB: I think it’s interesting to compare it to email, a legacy format which is still thriving. People are still using email—maybe some people have switched mostly to using other communication formats, but email is certainly not going away. I think it’s interesting to see that RSS is losing while there are still so many people using email.

KP: The first mover advantage is huge in a new space. Facebook wasn’t quite the first mover in social media—there was MySpace and some things before that. But various things happened: MySpace failed, Facebook took over. Email was after the older technologies, like the Bulletin Board System, which was a little more like RSS in the feel that you were dialing in somewhere to get the things that are there, instead of something being sent to you at your home.

I think email was the first one that seemed to just become the common denominator. It was a way that you could have a box where people could dump things in. And it felt very much the same as physical mail, in that you have a mailbox and people dump things into your mailbox. And we all hate the spam that gets dumped into both of those places, so there are obvious downsides. But email was there first, and it was an understandable metaphor for people. Now it is so widespread, and so relied on by the companies you sign up for and the companies that advertise to you.

It’s such a large ecosystem that the momentum there is going to make it stay alive. Then we went and started putting all of our login and authentication into email because it was available. Now everybody is fundamentally bound to their email address as their identity online. If they lose access to it, they also lose access to their bank and to all these other crazy things. I don’t think it's going away, but I do wish that email was more pull-based and less push-based.

I want to sit down and read my RSS articles when I have a down minute and I’m just relaxing, as opposed to somebody deciding that I need to buy their new product and emailing me, and that causing a ping on my watch. It makes me reasonably upset every time that happens.

I mostly turn off notifications for everything. I will not listen unless I’ve gone to look for the thing. So I really want a pull society, to the extent that anybody’s going towards the pull model that RSS matches. But nobody’s replacing email at this point.

MB: It’s interesting—you were mentioning the social network you thought about creating, and I think that has interesting parallels to email, in that you have an inbox and you are filtering out everything you don’t want. You’re left with only the things you are actually interested in, which is one of the pros of RSS. I think a hybrid between the two might be something that people would want, to remedy the shortcomings of both email and RSS.

KP: Maybe, because RSS definitely lacks the social aspect in large part. There are a couple of services that let you mention people when you publish things [webmentions]. But that’s why my friend and I had this idea and we liked how you had an inbox feed, but it’s not pushing anything at you. You’re going and checking the box when you’re curious.

Contrast that to email, where advertisers are telling me they have new pants and I need to buy pants. I don’t need to buy pants. When I have a hole in mine, that’s when I’ll go buy pants. It makes me a little sad that I don‘t think anything will ever unseat email—but we can still have a little bit of optimism.

MB: What would you advise to someone who wants to do more with RSS?

KP: I think I would encourage people to think about how to replace their stuff. Look at what you are consuming, think about what parts of it you actually want to consume versus having somebody throw a thing in your face, and then find where you can get that in an RSS feed.

I’m building out a few tiny tools in Sponder to help you do that, to find RSS feeds that might be harder to find, and it’ll try the 15 places it might be, and then maybe that’ll get someone to replace TikTok.

But you should just figure out what parts of your content make you happy and then find an RSS feed for it. And then you don’t have to think about it anymore ever again. You’ll get the things that make you happy, and then you’re done and can ignore everything else. Nothing will ever be added without you taking more action and saying, “Oh, I thought of a new thing that makes me happy, and I’m going to find a feed.”

MB: You had mentioned in your email that you’re working on a puzzle game—what is it and what was the inspiration?

KP: The inspiration was Pips, in The New York Times—it popped up recently, and it’s a fun game. I’ve been doing Sudokus and a bunch of variants of Sudoku for years, when I’m sitting at the bus stop and want to think—rather than when I’m sitting at the bus stop and don’t want to think, which is when I read RSS articles.

Pips looked really interesting to me, but I couldn’t quite tell if the puzzles were arranged in such a way that there was always a purely logical path to place the pieces. My wife would usually beat me in terms of time, because she was more willing to place a domino and see if it worked out—whereas I wanted to start from axioms and never place a piece wrong, which is what you can do in Sudoku.

I'm calling my game Scintle, short for scintillate. It’s basically Pips, but you use colorful triangles. The shapes are more interesting, but also they can overlap and mix colors. I think it’ll be interesting. Other than that, it's the same regional constraint idea of “this has to be red” or “these have to be all different colors” or something like that. It’s very clearly Pips-inspired, you’re just building a stained glass window out of overlapping pieces instead of dominoes.

I found my interview with Kristján to be really fascinating, because I too wish that more things were pull- rather than push-based. It’s a little ironic, then, that this is an email newsletter, but hopefully you’re glad it comes into your inbox! (Of course, there’s always an RSS feed available, too.)

Recently, the Beehiiv Analytics tool I’d been beta testing came out of beta and became a feature of the Scale plan. Beehiiv had warned me of this when I signed up, but it still came as a blow to me, since I had been so heavily reliant on the lightweight, privacy-focused tracker. After coming out of beta, it will only show me very basic information: pretty much just the number of page views in the previous 30 days. I dealt with it for a while, but, the other day, I decided I wanted to have insight into who was looking at my blog again.

I had remembered looking at Jeff Porter’s list of European alternatives to popular software products, so I figured I’d give one of his alternatives to Google Analytics a shot. I quickly landed on Alceris, a German-hosted, lightweight analytics tool which was free for up to 10k monthly page views. I get a little under that, more like 7k, so I decided to go with it.

After creating an account, I was brought to setup, where I received a snippet to paste into the HTML code of my website. In a normal situation this would be easy, but Beehiiv demands a lot of control. I couldn’t put any custom code in the website, at all. To do that, I would need to upgrade to the Scale plan, which, again, I didn’t want to do.

I tried to find a workaround, and realized that Beehiiv will let you put a custom Google Tag Manager tag on your website. (Now, if I were really trying to make a statement about how I feel about American tech—like I see so many in the fediverse doing—this would sort of defeat the purpose. However, I’m American, so I don’t particularly care. I use Gmail and Google Docs every day anyway.) I added an HTML Snippet tag to GTM, and published it.

I then went back to Alceris, where I received a number of page views from “https://gtm-msr.appspot.com/render/.” This was a minor annoyance, as I realized I was receiving false page-views. I quickly sorted those out and did some proper testing, where I got some real data into my pageviews.

Alceris is privacy-focused, lightweight, and doesn’t collect any extra data, so I’m very happy with where I’ve landed.¹ However, I’m upset at Beehiiv on two accounts: both that they took away my analytics and that they made it so hard for me to add custom HTML snippets.

I know that I’m on the free plan, but it seems to me I should still be treated with some respect as a customer. I signed up for Beehiiv hoping it would be more freeing than Substack (and I had no need for as much freedom as Ghost offers). I’ve been pleased with it for the past five months. But I’m starting to rethink my decision. I want to be able to add whatever I want to my blog and not be artificially limited by basic software. I'll probably stay, but not without my reservations.

This got me thinking about freemium software as a whole (hence the title of this post). If Beehiiv’s goal is really to convert me to a paying customer, which it is, then shouldn’t the strategy be to get the user to be so happy that they want to pay, rather than being forced to upgrade because of gatekept features?

I think too many freemium companies go with the latter, because it’s easier. They can simply decide that their advanced features/full-fledged software is behind a “Pro” plan, and not have to go through the pains of making their users love the free version of their software.

But as I’ve said before, people should love your software product, and you should design it with users loving it in mind. Otherwise you’ve failed to create a worthwhile piece of software—when some of the best programs (like Arc), were things that became instant fan favorites.

If Beehiiv had been studded with great UX features and felt like somebody cared about taking the program to the next level, I would have been more than happy to consider upgrading (I probably wouldn’t have, though, as I’m only twelve). But gatekeeping features just so I’m forced to upgrade isn’t going to make me happy.

In my opinion, Beehiiv should want millions of happy, free, users who can write about how much they love Beehiiv and refer others to the platform, who then might become paying users. Trying to turn all of your free users into paid users isn’t necessarily going to keep them happy, and they won’t write about it. In turn, you don’t get the positive community, and I would guess in most cases don’t actually get more paid subscriptions in the end.

If any SaaS executives are reading this: It might be time to rethink making customers pay. Invest all of the effort you’re spending on making dozens of subscribe pop-ups into making your customers happy, and the payments might come flowing in. But I’d love to hear from any actual founders who can tell me why I’m wrong!

With the internet, it’s easy to find people who share an interest or hobby with you, even if it’s quite niche. If you’re obsessed with the Arc browser, there’s a subreddit for you. If you have a passion for mechanical keyboards, there’s also a subreddit, as well as a bunch of dedicated forums, namely Geekhack and Deskthority. If you love the indie folk music of Gregory Alan Isakov (check it out!), there’s a community on Reddit for that, too. And if you’re a type designer, calligrapher, or in some way involved with or interested in letterforms, there’s a whole Mastodon server for you.

Yes, those are my interests, among (many) others.

What’s harder to find, but all the more so rewarding, is when two (or more) of one’s interests are combined. Now, this experience can take a variety of different forms. It could entail meeting someone with whom you share two or more interests (always fun!). Another form, which I will focus on, is when some sort of project, post, video, or really anything, interests you because of two different passions you already had. Since I’ll be referring to this experience a lot throughout the post, I’m going to call the noun a “crosspassion” and the adjective “crosspassionate.” To clear any confusion up, I’ll share a couple of examples of crosspassion that I found recently.

As I delved deeper into the world of mechanical keyboards, I learned about some fascinating projects. One company that I re-discovered last week through the fantastic mechanical keyboard newsletter clicks ‘n clacks is Typface, which—you guessed it—makes keyboards inspired by famous typefaces. When I saw their keyboards, I was interested on two levels. Yes, I love looking at interesting keyboards, but what really drew me in was the fact that they appreciated both keyboards and fonts. It’s not that I wouldn’t have explored Typface’s offerings if they didn’t have that added tie-in. It’s just that it amplified the curiosity I already felt, and made me feel a connection even before I knew much about them.

As I thought about the crosspassion between me and Typface, I realized that crosspassionate experiences can occur in even the simplest of ways. When someone is interested in something you’re interested in too, that creates a bond. The bond only deepens when you have multiple relatable things, which is what makes crosspassions so powerful, even if it’s just looking at a website, and thinking, Oh, that’s cool.

As well as what I mentioned earlier, I enjoy trying new foods and going to new restaurants, as well as the music of singer-songwriter Ben Rector.

Those two interests remained completely unrelated, until I stumbled upon a crosspassion experience this week, in the form of a new initiative launched by the Tennessee Department of Tourist Development, pairing Tennessee musicians with chefs from across the state. It explores “sonic seasoning,” which is a phenomenon in which music impacts the way we perceive taste.

In the 8-minute film I watched, Nashville-based Chef Philip Krajeck of Rolf and Daughters was paired with Ben Rector to create an original song in 24 hours, based on Krajeck’s signature dish, tomato pasta. The song that resulted, titled “From the Ground Up” both sounded great, and, in my opinion, evoked the feeling of a garden.

Why was I so interested in what was effectively a promotional video to visit Nashville? Well, in this example, had it been another musician, whose music I hadn’t already discovered, I don’t think I would have watched it. But the video appealed to me because it took two things I loved, and mixed them together (pun intended) in a way that I would never have thought of. Who knew that music affects taste (or did I just scroll past that headline)?

Crosspassion can occur with any two interests, not just obvious ones. My dad, for example, who likes baseball and sushi, would find it rewarding if there were a group of baseball fans who wanted to go out to sushi restaurants—and would be more rewarding than just a group of baseball fans who talk about baseball, or a group of sushi-lovers who eat sushi.

I think that someone should create a new platform for niche interests, but for people who have the same two (or three). I think it would look something like this: I list my interests, and in some way, they get added to some big grid, and the intersections of the lines (hobbies) are where the magic happens. If anyone wants to build this, I’d be happy to advise!

As people, we focus on who shares one hobby with us. But what if we can create a world of newfound friends who share all of our hobbies? Now, who wants to solve my Ben Rector-themed 15×15 crossword puzzle that uses a sans-serif custom font, must be done using with a Topre keyboard, and if completed correctly, you get a free trip to Tennessee?

Lately, I’ve been very busy interviewing and reaching out to small, new type foundries for my niche type design zine (a lot of people have been asking me if this will be published online—I’m not sure yet; it may just be in print). I have a bunch of different methods for finding foundries. At the beginning, I relied upon dozens of Are.na lists to guide me to foundries. When I realized that the newest foundries (launched within the past year) have great perspectives for my zine, I started diving down the rabbit-hole of type foundry indexes (typefoundries.info, Type Foundry Directory, and Type Atlas were most helpful). When those didn’t surface enough for my liking, I turned to Google Search operators to find foundry listings on Fonts In Use that said something along the lines of “launched in [month] 2025.” That gave me some great ones.

While my methods were not the most elegant, my conversations certainly have been. I totally overshot, and reached out to way more type designers than I need to for the zine, but talking to them and hearing their unique perspectives is just… so much fun. It’s forced me to be organized and keep a crazy spreadsheet with dozens of names and statuses and seventeen different colors.

Something that I learned from this part of the process: There are so many type foundries. I’ve probably reached out to everyone who’s even remotely relevant to my zine by now, but it’s incredible to me how many of these small businesses are able to support themselves.

One foundry that I stumbled upon in my research, launched in November, is GD Foundry, the type design practice of Stockholm-based studio Gärde Design. While I’m fascinated by many of their typefaces, GD Optio caught my eye when I was browsing their catalog. Optio is a reimagination of the classic humanist sans-serif Optima, designed by Hermann Zapf in 1958. It’s long been a favorite of mine, always eluding me, “Is that a serif that I see?” I would ask myself. This reimagination preserved that spirit I loved, while modernizing the font. I wanted to learn more: about this version, about the original, and humanist sans-serifs in general.

So I turned again to Fonts In Use, this time carefully examining all typefaces tagged with Optima. I opened them all up in the browser, determined to find all Optima-inspired typefaces from the past few years. I expected to see not more than a couple, spread out by a few years. But no. This process yielded me with four Optima reimaginations… all from 2025. That was when I got confused. Why, in 2025, would there be four typefaces inspired by Optima, but none in the previous few years?

I wanted to find an answer. After all, I had created this whole question in the first place. So I dug up all the information I could find on each of them: Amarelle Sans, Museum, TWK Issey, and, my personal favorite, GD Optio.

Amarelle Sans

This one has the simplest backstory: on typeface from a three-piece set reinterpreting three of Hermann Zapf’s most famous designs: Palatino, Chancery, and Optima—as Serif, Chancery, and Sans, respectively, and meant to go together. Amarelle Sans, the Optima-inspired one, was designed by Alex Blattmann, Ben Jones, and Diogo Rapazote, and released to Dalton Maag’s catalog.

Museum

Released by one of my favorite bigger foundries, Pangram Pangram, in March of 2025, and designed by Andrea Biggio, Mat Desjardins, and Francesca Bolognini, Museum finds its inspiration through Optima sketches found on old banknotes. While I think it’s a fantastic typeface, it strays a bit farther from the roots of Optima. That’s not necessarily a bad thing, just an interesting thing—one of the many considerations of picking a typeface.

Museum was recently used by Sporthotel Lorünser, an Alps ski hotel, in a rebrand by Teel, paired with Pangram Pangram’s Editorial Old, providing (in my opinion) a really beautiful effect.

TWK Issey

Designed by Paris-based Rimasùu Studio, Issey was released by Weltkern in October. According to the description, it is, “Inspired by French vintage fashion show invitations and Japanese aesthetics… [and] rooted in the tradition of incise typefaces such as Optima.” Like Museum, Issey uses Optima as a jumping-off point—not as a limiting anchor. While I do see more of a resemblance to the original than Museum, I still think it takes more of a modern approach to the reinterpretation.

GD Optio

As I mentioned, this one is my favorite. Not only does it come from a small, new, yet highly talented foundry, it has all the qualities one would want from a reimagining of Optima—and the name is spot-on. Optio was created by Marcus Gärde, and Anders Wikström (read Anders’s thoughts on it here), and released by the brand-new GD Foundry in November. It achieves what any of these strive to do: to preserve the legacy of Optima, but to breathe new, digital life into it. Oh, and it comes with one of the best lines in a typeface description I’ve seen, “… because good design should be both functional and flavorful.”

So, Why?

Now that we’ve established our facts, understood the four Optima-inspired typefaces released in 2025, and learned about digging through history to create something all-new, I’ll attempt to answer the question I posed at the beginning: Why is it, that in 2025, there were four typefaces released inspired by Optima?

Optima has a hand-drawn element to it. While even Anders Wikström, the designer of Optio, notes in his post about his font that, “Personally, I don’t think it reaches the lofty standards of Zapf’s calligraphy. But it still has to be said that it originated a new subgenre of sanserifs.” Hand-drawn elements are something we’re missing from our everyday lives, in this era of AI. The type world faces the risk of being eaten by AI, by fonts created from a dataset of other fonts. So what do they turn to? The most calligraphic of fonts, found by digging through history, at fonts created before there was such an notion of fonts created by AI. And these type designers take these historic fonts, and use them, as well as a variety of other sources, to create new fonts.

I see an analogy there. Instead of leaving it up to AI to create new fonts which are really combinations of old ones, the designers take it upon themselves to pay homage to these designers that have inspired them, and create something new—not destroying the soul of the old font, but amplifying it. Some would even say making it better.

With Optima, it was no coincidence that four fonts released in 2025 were inspired by it. Optima is a masterpiece of creativity and the human hand, with flared terminals and a sophisticated texture, literally the opposite of something an AI font creator could make. These designers, in a sea of AI, chose to augment this human creativity, rather than let AI swallow the world.

I’m really excited for 2026 in type design. Not just new font releases, but all the new foundries that will come about (Type Atlas lists 27 foundries launched in 2025), and I’ve already discovered a couple established this year. Most of all, my interest is piqued to see what 2026’s Optima will be—a font cherished by type designers, that leads them to create new versions, inspired by the original.

Will 2026 bring a bunch more Optima redrawings? Maybe. But chances are, there’s something new, waiting just around the corner. After all, the first one that I mentioned from this year was released last March, so let’s just wait and see…!

If you’re receiving this via email, you’ll notice my redesign. I didn’t change too much—I kept my base typography and formatting, but I updated my color palette and heading styles pretty significantly. I have to admit—I was getting a bit tired of the black on white, and wanted to design something better. I’d love to know what you think. If you have 30 seconds to spare, I’d love for you to fill out this quick poll! If you’re not receiving this via email, now’s a great time to subscribe (it’s free!).

Until very recently, my email strategy has been very simple. When I receive an email, I read it. I usually respond very quickly, especially since I use the Gmail feature that shows the number of unread emails in the tab icon. After I deal with it, I leave it in the inbox, so I have an almost complete record of all my emails (I did try deleting unimportant emails a few years ago). As you might imagine, with all my email lists and password reset notices and actual important email, I’ve accumulated, over the course of twelve years, fifteen thousand and six emails (I’d like to reach 20,000, so if you email me with your email thoughts, you’re a drop in the bucket). That sounds like a lot, but it’s not, when a good chunk of it is email lists, or emails from years prior that are no longer relevant to me.

In the past week, as I’ve been emailing a bunch of type designers to be interviewed for my type design zine, and there are some emails that are going to take a bit longer to respond to, I’ve been marking them as unread. It’s certainly not the most efficient way (I really like what Talanoa is doing with kanban to-do lists in email, but I’m so used to Gmail it’s hard to fully switch). Combined with the notice in the tab icon, and my glancing every five seconds at my email (did it again!), I’m reminded of the email that I have to deal with, and it makes me feel like there is one small thing hanging over my head. Not enough to stress me out, but enough to give me a quick nudge to stop perfecting my over-the-top spreadsheets and do something more productive. That one lingering email is actually a motivator.

As I thought about it, I realized that Inbox Zero is not the right strategy. While I like that, in what I do, all of my emails are neatly in a list, striving to have zero unread emails is not actually a useful productivity strategy. I don’t want to have thousands of unread emails, like my teachers (the most I’ve ever seen in unread emails by one of my teachers is somewhere around 157,000, which is crazy), but keeping an email or two open actually helps. It doesn’t feel so far off from Inbox Zero, which is part of why it’s so useful. It’s only one or two tasks that you need to attend to, yet it pushes you to knock the rest of the items off your to-do list. And if you’ve done everything you need to do, just mark any random email as unread. Just seeing that you have an unread email will push you to achieve more.

You probably don’t believe me, especially since you’ve most likely been using the same email strategy for decades, and I’m only 12. But if you have thousands of unread emails, just mark them all as read and start from a clean slate—if you have that many, it can’t possibly mean that you have to deal with that many tasks. And if you’re like me, an Inbox Zeroer, just mark one email as unread and continue on whatever task you were working on.

I encourage you to try it for a day: if you receive this in your inbox, mark it as unread after you finish reading it. And if you’re reading this on the web, you’ll need to subscribe! 😀 

Let me know how it goes, and let’s see how long it takes for me to reach 20,000 emails (19,999 if you don’t count the unread email).

This is somewhat of a continuation of a previous post: if you haven’t read that yet, you can read it here.

As a seventh-grader, I use a variety of tools at school, and I thought it might be interesting to comment on many of them, both positively and negatively (mostly the latter), and perhaps humorously.

Nearpod. Nearpod is an interactive presentation tool that we often use in class. It displays the teacher’s slideshow onto our computers, so if you don’t have glasses but you need them, maybe it’s easier. It also lets students answer questions and complete a quiz game called “Time to Climb,” in which you need to answer correctly, and fastest. I’m pretty sure that in real life, climbing as fast as possible up a steep mountain with carnival decorations (?) isn’t the best idea. One of my teachers reportedly said at the open house that she is “in love with Nearpod.” Perhaps this is better than being in a relationship with a chatbot, but I’m really not sure.

Toddle. Toddle shows me what assignments I have due and my grades. I think it could show more, but that’s all my school has adopted. They migrated learning management systems to Toddle from Blackbaud last year, and the IT Department thinks it’s the best thing ever. No one else feels so strongly. There are some things I miss from the old program, like being able to add an assignment that isn’t posted by a teacher from within the program (like a to-do list), but overall, it’s fine and mildly infuriating when they do a redesign (every few months, it seems).

Edpuzzle. Edpuzzle, which I touched upon in my previous ed-tech post, allows my teachers to assign my classes videos to watch, and it embeds questions within the videos that we have to answer. For some reason, this year, my school decided to use the embedded version within Toddle, which is annoying. The way Edpuzzle works is it embeds a YouTube video using the youtubeeducation.com domain, which is not blocked, while the main YouTube is blocked by the school’s blocker, Securly. Using our school accounts, embedding a video onto a Google Slide also uses the YouTube Education site, so now everyone just watches videos on Google Slides. Unfortunately, the school can’t do much about that, because then Edpuzzle wouldn’t work, just like the MIT Scratch paradox (if they block Scratch, their programming class won’t be able to use it to make games).

Canva. This one is certainly not just used within my school, and is actually very popular. However, I decided to include it because it’s so strange how my teachers feel about it. Some believe that it is a waste of time, and there are better ways to do projects. But others of my teachers love it, and have us use it for everything, from posters to presentations to documents to videos. I don’t hate it, and I think it gives a lot of creative freedom. However, what I see so many kids doing is just reusing a boring presentation template, but adding images in random places. It always looks almost “too perfect” to me, which is why I never use a template and create my slideshows and videos from scratch. As it works out, educational facilities get free Canva Pro, with which I can do other things.

AI. This one is a bit of the odd one out. As students, we are forbidden from using AI for school-related purposes. However, I see my teachers using it to create assignments, mistake-filled example essays, lesson plans, and class discussion questions. some more than others. It makes me wonder: why am I spending so much time doing this assignment that was obviously created by ChatGPT or Claude (there’s literally a tab with a ChatGPT icon in the teacher’s browser!)? I wouldn’t say that AI is necessarily bad at creating assignments. But there’s sometimes factually incorrect things, or questions that don’t make sense, or analysis that feels far-fetched in a class discussion. If the students can’t use AI, why is there a double standard for the teachers?

SMART Notebook. I don’t get why my teachers love this interactive whiteboard tool so much. They never even put it in presenting mode! It lets them annotate slideshows and PDFs with a pen tool, and that’s all they use it for. It can actually do a lot more cool things, like hooking up to a SMART Document Camera to display on the screen, but of course my teachers only use the strangest features. My school phased out SMART’s projectors a few years back in favor of TVs, but instead of buying SMART’s TV-style boards, they bought Newline TVs, which run a modified version of Android, and connect via HDMI. Since it has four HDMI ports, connecting is very difficult for my teachers, and it couldn’t be that whatever one is hooked up to a computer is highlighted in white, while the others are gray. No way! Sarcasm and interlude aside, SMART Notebook is definitely not always the best tool for the purpose, but it’s whatever my teachers are used to, so we use it. And the other day, we carved our math answers into stone tablets.

So, for all the K-12 Systems Administrators reading this, maybe think twice about the tools you’re paying thousands of dollars a year for. Are they the best options — not even from the students’ perspectives, but from the teachers’ perspectives? What is the value add over Google Slides or another free tool? Do you need to pay for SMART Notebook, or is OpenBoard a better alternative (it is, and it’s free and open-source!)? If my school switched to OpenBoard, they could save about a thousand dollars a year, and while many of my teachers have decades-old libraries of SMART Notebook files on their hard drives, they can be converted to work with OpenBoard, albeit with some buggy elements, and depending on the way it’s done, non-editable components. Or try annotate.net (which my fifth-grade teacher used), which includes compatibility with SMART Notebook, and is web-based.

I don’t know what school was like before all of these tools, but I don’t necessarily feel like any of them really help me learn. As opposed to my classmates, when studying for my math tests, I print out all the SMART Notebook files my teacher shares to go over, and when making creative assignments like posters, I do it by hand instead of turning to Canva. A lot of times, these analog methods are completely ignored in favor of modern, technological approaches. Instead of teachers writing on the whiteboard or chalkboard, they’re sending their presentations to our computers, and we take notes on Google Docs. It almost feels “over-engineered,” like Apple putting the guts of an iPhone into the Studio Display, their 27-inch monitor, to enable advanced camera, audio, and voice assistant features, that aren’t really necessary. So maybe ed-tech isn’t all teachers and admins think it is, and some analog methods are truly the best methods. Plus, chalk is a lot cheaper than dozens of software licenses.

This week, I was starting work on an early mock-up of my niche typography zine using the newly free Affinity. Downloading the software, my thought process was something like: This used to be a paid creative suite, now it’s free, so it’s probably good. While I can report back that it did indeed allow me to create my mock-up, albeit with some frustrations on my end, its weak spots disappointed me. (If anyone would like to give me advice on publishing a zine and take a look at some very rough drafts, feel free to reach out. I’ll be sharing more about it in a few weeks, so stay tuned!)

My naive assumption, upon beginning the work, was pretty much that I was getting InDesign for free. But the Affinity creative suite, after being acquired by Canva, has been unified into one app, and a lot of its former functions have been dumbed down, gotten rid of, or hidden in layers of menus. What I was really getting was software that you’d expect from a free download: serviceable, if that, but certainly not more. Luckily, I didn't have to learn to use Scribus, an unintuitive desktop publishing application.

Most software falls into one of a few categories: free open-source, free closed-source, one-time payment, and subscription. Each one has its individual trade-offs, and some more than others.

Free open-source software can be great (think Handbrake, for video converting, and Blender, for 3D modeling). It gives you the power to do whatever you want with it, and if there’s something you don’t like about the program, you can hack it in. (Of course, there’s licensing and all that, but that’s not the focus here.) But open source software can sometimes also be difficult to use and not that polished (like Scribus).

Free closed-source software addresses some of those issues. They have a real company behind them, and as a whole, are probably more polished than free open-source software. Apps like Arc (free) and Raycast (freemium) bring a whole new level of power to their users. The downside is that they’re proprietary and can’t be modified like open-source software. Affinity certainly aided me in creating my design, but it felt like if it was paid, it would’ve been even better (sounds like I should just buy InDesign).

Finally, one-time payment and subscription software often have more polish and functionality than all of the other options. I’m not at all against paying for software (or at least my parents sometimes aren’t). Right now I pay for Mood Camera for taking beautiful pictures, Crossword Compiler for making my crossword puzzles (I haven’t gotten one in the NYT yet, but hope to be the youngest constructor), and Suno, for generating music with AI. I think that, sometimes, there’s really nothing that can match the power or ease of a certain paid software program, just like Affinity and InDesign.

Paid software, especially subscription-based software, gets a lot of negativity online. In a lot of cases, there’s a free alternative that’s just as good, and I think I use more free alternatives that anyone. I’m not saying that you shouldn’t use open source software or free software. Every type has its advantages and disadvantages, and the important thing is to weigh your options. If you have a big budget, and the paid choice is the best one, then you should probably get that. But in other situations, the decision might be a little bit more nuanced.

Free software is often limited by lack of funding, something that paid software doesn’t need to worry about. By buying indie paid software, you’re supporting the developers. When you buy regular paid software or subscription software, you’re supporting a company that you might not want to necessarily fund — like Microsoft or Adobe.

But when you support an indie developer creating a program that is genuinely useful and beautiful, you feel good — like you made a positive contribution to the world and to the developer. I personally aspire to support more indie developers, but if there’s a free alternative, I won’t necessarily jump to paying for a program.

I’m curious to know what my readers pay for, and how they feel about open-source versus closed-source software. You can be my first commenter!

As 2025 is coming to a close, I decided I’d put together some of my favorite things from this year.

Fonts: I am obsessed with fonts, and love seeing new ones get brought into the world. All of the following fonts were released during 2025. A little sneak-peek for what’s to come: I interviewed a bunch of designers at niche type foundries earlier this year, and am thinking about maybe publishing a zine about it, so stay tuned, and if you know any type designers who’d like to be interviewed, feel free to shoot me an email.

• Swim Club, by Old City Mailroom. I love the versatility of this font. It comes in sans and slab, tall and wide, and textured and solid. While the sans looks like the kind of thing you’d see in an old beach town (hence the name), I could see the slab going pretty much anywhere.

• Print, by Faire Type. This family, based on copperplate script, evokes the feeling of an old book. Print Script, the cursive-inspired typeface, seems great in its display form for a standout portfolio, while I can see the regular Print becoming the next Instrument Serif (which I use for the heading of this newsletter).

Keyboards: Just a heads-up — I do not own any of these keyboards (although, once my newsletter takes off, maybe… ). However, I do follow the keyboard news very carefully, and a couple from this year really stood out to me.

• Cleaver, by Serene Industries. The successor to the Icebreaker, this keyboard continues the all-aluminum tradition. While the brutalist design language of the Icebreaker was… interesting, I wouldn’t say that it really spoke to me, with all of its sharp corners. However, the Cleaver embraces the minimalist nature that I would’ve wished for in the Icebreaker and it looks really cool. Probably not so great to type on though.

• Encore Series 3, by Mod Musings and Mode Designs. Interestingly, my keyboard preferences have changed from liking 100% to gravitating between 65% and 75% boards. The design language of the Encore has caught my eye for a long time, and while I slightly prefer the wood of Series 1 and 2, the brushed copper bottom of this batch looks really nice.

What I Read: I read a ton of different things, and thought I’d share some of my favorites.

• The Verge. This is my go-to for tech news and reviews. I feel like nearly every article I read is super high-quality and informative, and I can always trust their tech reviews to be useful and true — their ethics policy seems very, very important to them. It feels like most of the other tech news sites are owned by Future and are filled with clickbait-y articles and very little original content.

• Installer. This newsletter, written by David Pierce of The Verge, is one of my favorites (and the inspiration for this post). Every Saturday, I get the latest tech products (with commentary), a Q&A + home screen share of someone in the tech industry, and a crowd-sourced section, where people can write in/comment on social media with what they’re into that week. Highly recommended.

• Workspaces. A fellow Beehiiv publication, this newsletter delivers a desk setup and a Q&A with a tech/creative professional. While not the most writing-heavy, it’s still interesting to get a window into the life of a new person.

• Dense Discovery: Similar to Installer, this newsletter includes links to interesting things on the web, plus a section with art picks and a new font.

• Braise: Somewhat of an odd one for this list, Braise is not related to tech. Written by Dominic Preston of The Verge, it comes twice-weekly with London restaurant recommendations and recipes. I live in New York, but I still find it interesting, because I love reading about food, whether or not going there is in the equation.

• The Manual: This newsletter comes monthly, and is written by Oliur, a design influencer. It shares anecdotes about his career, plus a section (my favorite) at the end, with links to cool internet things. I mentioned Link Lowdown in a previous post, which is the website version of the (more curated) links section of this newsletter.

• clicks ‘n clacks: I linked to this newsletter in last week’s post — one of my favorites, it includes three interesting links (not the same ones as my previous picks, don’t worry!), and three new keyboards. Plus, it’s on Beehiiv just like me!

• Product Identity: I wrote an article for this publication in the spring about Deta Surf, a new web browser which recently went out of business. It talks about niche design and small startups and offers a fascinating perspective on the worlds of tech and design.

An Anti-Recommendation: At school, we play a lot of interactive quiz games in class to supposedly study for tests and the like. However, what I’ve found with the most popular one, Gimkit, is that it becomes so far removed from the content (in most of the game modes, you answer questions to earn energy in order to move around in the game) and the complicated, involved games heavily distract from the learning material. While the other kids in my class find it fun, and I don’t necessarily dislike playing it, it becomes less of a study tool and more of a game that is disguised as a study tool. I don’t think the games are bad — I just think the intention behind the games is confusing.

While I’m pretty sure I wrote the word “newsletter” about a hundred times, I do hope that my recommendations were as helpful as other yearly recaps!

The Magic Mouse is the most controversial computer peripheral. The haters eschew it in favor of a big, bulky, button-packed Logitech MX Master 3S. But why does half the web so strongly dislike it, and why does that drive them to create the anti-Magic Mouse?

For one, the mouse’s charger port is on the bottom, which means you can’t use it while you’re charging it. Since you only have to about charge it once a month, it’s not nearly as big of a deal as Reddit would have you think. (My dad, an iMac user, finds that not being able to use it upside-down is quite annoying when he forgets to charge it, but he’s been using it for years and happy with it overall.)

Also, some people find that the mouse’s shape is unergonomic and uncomfortable. Again, a valid concern, but not something everyone cares about/notices. People talk all about how the Magic Mouse is horrible, how could Apple sell such a product. But the issues aren’t that big.

People have tried to fix both concerns. Of course, buying a standard computer mouse can solve both issues, or, if you want to be fancy about it, get an MX Master 3S. But over the past few years, we’ve seen products pop up like the $30 Magic Mouse sock, an off-brand Amazon product that attaches to the bottom of the Magic Mouse and adds a more ergonomic grip and MagSafe charging (still not as cool as the Razer mousepads that beam charge to your mouse).

I saw another, even more interesting product to address people’s hatred, the Melt Mouse (courtesy of clicks ‘n clacks, one of my favorite newsletters). The Melt Mouse is a distinct computer peripheral that keeps all the same swipe-y gestures of the Magic Mouse, but gives it a much more defined, ergonomic shape and a shortcut pad. It also uses LRAs, which provide haptic feedback, instead of buttons to click. LRAs are the same thing that Apple uses in its Magic Trackpad and the trackpads that come with its MacBooks (another thing you could get if you hate the Magic Mouse). I think that when Sensel, the company that makes the haptic trackpads in pretty much every PC (though I don’t believe Apple uses them), goes public, they’d be a great investment, because they’re going to be in every PC laptop within the next few years.

The Melt Mouse seems really interesting, and might be a good purchase if you have a very strong dislike for the Magic Mouse. One thing to note is that the charging port appears to be on the bottom from the renders, but looks like it’s in a spot that you can still use it while it’s charging (fingers crossed). The most useful feature, for me at least, is that you can turn the mouse into trackpad mode, which lets you use things like pinch-to-zoom. I do think that Melt Interface should have done slightly more market research, because just the potential for the charging port to be on the bottom might lead to a boycott. I also think they need a better name. The current name evokes images of computer mice disintegrating, which probably isn’t their goal?

I’d say that if you hate your Magic Mouse, get a Magic Trackpad. If you want a mouse, get a Logitech MX Master 3S. If you miss the gestures but still hate your Magic Mouse, get a Melt Mouse (assuming you have money to spare, because every single one of those is quite pricey, including the Melt Mouse, which you can buy from Kickstarter for $270).

Or just pause and take a moment to think — these supposed issues are barely issues. Why am I wasting my time throwing mice at the wall?

Over the weekend, I was playing around with Suno, the AI music generator, having read the article about it in The Verge.

I was inspired by the article to write a country song and see what it could do with it, and the results (especially after refining by making a second version) were pretty remarkable. Embed, with Suno AI-generated image included:

Nashville Music Career

That got me thinking — if Suno could take my lyrics that were thought up in about ten minutes and turn it into a song that could be on country music radio with a little bit of touching up, what could it do for real country musicians?

The article has a clip of a song by some songwriters that’s just vocals and guitar on a voice memo — no mastering or anything, side-by-side with the same clip, remixed by Suno, which keeps the melody and rhythm, but makes the song radio-ready.

With this new technology where pretty much anyone can make any songs they want with lyrics and a prompt, professional musicians are going to be forced to be better than Suno. Beating users on the free plan is easy — v4.5-all, their free model, isn’t going to match the creativity of a real artist. However, their paid option, v5, is more accurate and less predictable than v4.5-all.

Suno’s not going to get rid of musicians yet, or at all. For celebrities, part of why people are listening is because they want to listen to the celebrity — or at least at this moment — because music, since its very start, hasn’t just been about the sound. It’s about who’s singing, why they’re singing it, and how it makes people feel. Suno is highly impressive, and I’ve been blown away with the quality of the songs I’ve created. But musicians, for now, have creativity, something that AI doesn’t have. Artists can decide if their song needs a trumpet or xylophone solo and make their choice for a reason. AI can make the choice, but even if it pretends it has a reason, it might’ve still made the wrong decision.

Where Suno really shines is empowering budding songwriters with the tools to create great tracks from their own music. As the article explained, musicians previously needed to pay somewhere between $500-1000 to get a “studio guy” to record a demo track to pitch to a publisher. Now, they can just use Suno, and get a pretty much equal result in thirty seconds. These demo tracks are never meant to be heard. They are just to send musicians to record. These demo recording artists aren’t making those same choices that the artists can make once they received the song — they are just there to serve a role in the process. That these people are being replaced by AI has no impact on the final creativity of the song. It makes it easier for songwriters to do more at once, and, if anything, might result in a positive change in the music industry — more songs from skilled songwriters.

It also allows non-musicians to dip their toes into creating music — something that used to have a very high bar to entry if you wanted a professional-sounding song. Now someone who’s never recorded a song before can just type lyrics into Suno and a prompt for what they want the song to sound like, and they’ll have a song. Suno lets people not have to hire instrumentalists to make a song for fun.

I think that Suno threatens the “track guys” only. For songwriters, it allows them to do more. For the session players, a real musician would hire them if they want their song to sound even better. And it lets normal people create songs for free, which opens up the music industry to more people.

I feel like they could charge a lot more. If the article in The Verge is true, that every Nashville songwriter is using it, they could charge more than $72 a year for 500 songs with their v5 model, under their Pro plan.

Overall, I’m super intrigued by Suno, and am thinking about upgrading to the Pro plan mainly for access to v5 and uploads longer than one minute. It makes the whole process of creating a song a whole lot less intimidating for beginners like me, but it’s also really interesting to see just how much it can do. To me, Suno feels like one of the best uses of AI I’ve ever seen. Using AI to write or to code doesn’t leave you walking away with something that you feel is yours. When you use Suno and put in your own recording for it to make it better, you get something that you made — just amplified to a level higher than you’d ever be able to achieve.

I hope that Spotify doesn’t get overrun with Suno artists, but at the same time, I know it will. With the power that it gives the public to create music, we’ll have to adjust, just like we’ve done to detect when a piece of text or an image is AI-generated: we’ll have to filter through the noise and find artists who are making music because they’re passionate about it. Right now, no one is really suspicious of music being AI-generated in the same way we are of videos, images, and text. But I think we’ll soon have to be. I’ll close with an amusing comment on one of my previous posts from Hacker News that feels relevant here:

I didn’t use AI — all of my words are mine — and so are all of my em dashes — just not my song — I mean, sort of — I wrote the lyrics.

Take browsers. Chrome is used by three billion people. Yet you wouldn’t know that a human actually put time and effort into it from the generic design and merely functional layout. Now take Arc: it still has tabs and windows just like Chrome, but it’s different. Arc brings a magical feel to the user, like it was made by people who care — the music notes when you play Spotify, the bouncy download icon when you save a file. These features don’t need to exist. Arc could’ve been just as corporate as Chrome and Safari, without the little touches — but The Browser Company chose not to be just like Google and Apple.

Take mechanical keyboards. Most mechanical keyboard¹ users use Keychron. Yet Keychron keyboards are like Chrome — good for a beginner, leaving something to be desired for an experienced user (build quality, uniqueness). Now take a Norbauer Seneca, or a Mode Encore: Ryan Norbauer spent years fixing the space bar — one key out of a hundred. Every tiny piece of the Seneca is agonized over, from the material choice to the innovative retro-profile MTNU keycaps. In the Mode Encore, every detail of the cherry lip, the chamfered edge on the front of the keyboard, is thoroughly thought about and looked at.

Take fonts. 50 million websites use the Google Fonts API. Yet Google Fonts, for the most part, aren’t going to be as pretty, as thoughtful, as paid fonts (money is a good incentive!). Now take Old City Mailroom: same letters as Google Fonts, but handcrafted and perfected to an insane level, producing some of the most original and unique fonts I’ve ever seen.

We need more of those niche products. Right now, all AI chatbots look the same — they all have a chance to stand out, but they deny it. They could include novel user input methods, a slightly different layout from the all-too-common sidebar and chat box. They could even use Old City Mailroom fonts! Products that try to stand out might not have billions of users like Chrome, but a few thousand dedicated users easily trump billions of forced ones.

Companies can stand out in many ways: original design, quirky copy, efforts to build a community. When you try to stand out — really stand out — that’s when you get the dedicated user base. The Norbauer Seneca’s not going to be for everyone. But enough people will deeply love it.

I love trying to find niche products — searching for that sense of satisfaction when you stumble upon something that will really make a difference in your life. Currently I’m enjoying Mood Camera, a replacement for the stock camera app which comes with tons of film emulations and presets. if you’re looking for the same kinds of products, whether it’s hardware or software, I’d suggest looking at Oliur’s Link Lowdown, a collection of links to things that are niche and useful.

Companies that create niche products change the industry around them. When products all start to look the same, it signals functionality over experience. I’d make the case that people should switch to personality-filled tools: from Chrome to Arc, from the default to Mood Camera, from Google Fonts to Old City Mailroom (not saying you should necessarily buy a Seneca). Niche tools are created by people, and when you buy them, you fund people. Mainstream tools are created by large companies, and when you buy them, you fund companies, not people.

If you’re interested in reading more about niche products, I’d recommend the fascinating Niche Design Zine, by Itay Dreyfus, with whom I collaborated earlier this year. It’s filled with interviews and Q&As with designers and developers of niche products and personal essays about the world of niche design.

I’d also love recommendations for niche products I’ve never tried — send me an email if you know of any!

This week, Deta announced the shutdown of their managed services.

This came as somewhat of a shock to me, having felt more connected to Deta than the other browser companies. Last year, I did two calls with a design engineer at Deta, Felix Tesche, where I offered feedback (if you want to read more about that, I’d suggest my article from March about Surf).

It was obviously coming. The Browser Company, much larger than Deta, had been acquired. ChatGPT Atlas and Perplexity Comet had come out — companies that already had a foothold in the market.

Deta Surf was such a gem. While OpenAI and Perplexity’s offerings might have had more capability, they felt like the next Chrome. Surf felt like the next Arc, bursting with personality and quirky details. Where Surf has notebooks, documents you can fill with your own findings and those of AI, Comet does everything for you. Where Surf has Surflets, apps you can create with Claude and and pin to your home page, Atlas has browser memory.

I believe that within the next few months, the only AI browsers people will be talking about are Atlas and Comet, and those will look just as corporate and impersonal as they do now. Without smaller players like The Browser Company and Deta, there’s nothing stopping the others from becoming just like Chrome.

I don’t hate Atlas and Comet — I think they’re really cool tools. But every cool tool has its weaknesses, and a big one for me is the lack of personality. I want to feel the designer behind it, not a robot. I want to see intentional choices — even if I don’t agree with them. I want to see interesting typography, new layouts, and fearless changes to traditional elements like the tab or URL bar. I want the same advanced AI features as Comet and Atlas, but infused with a sense of personality that Comet and Atlas just lack.

The Browser Company left a huge impact on the browser industry and the tech industry more broadly, whether it’s vertical tabs, favorites, spaces, or the brand identity itself (looking at you, Interaction Company of California). But it’s also about surviving, yourself, as a company.

In an ideal world, every browser company would be Deta. Unfortunately, that’s not going to happen. Or will it?

I love keyboards, and at the top of the market, there’s nothing like a Norbauer Seneca. In case you’re unfamiliar, the Seneca is a mechanical keyboard made by Norbauer and Co. with an incredibly high build quality and attention to detail, and an even higher price.

Don’t get me wrong. From all I’ve read, the Norbauer Seneca seems great. Yet, it was never meant to be a mass-market product. Its purpose is as a luxury good, bought by only the most dedicated mechanical keyboard fanatics — and its price, starting at $3600, reflects that.

I also love app awards. Whether it’s mainstream ones, like those of Apple or MacStories, or ones like the Expo App Awards, which are meant to display apps made using a specific platform (in this case Expo), app awards showcase the best apps of the year, both giving the creator satisfaction, and granting the app more fame. And while the inaugural winners of the Expo App Awards should certainly get a prize, the Norbauer Seneca doesn’t feel right to me.

The people buying a Norbauer Seneca are a self-selected group, who are willing to pay thousands of dollars for the perfect typing experience. Most other people, even if they type all day, wouldn’t want such a product. My dad is a writer and types thousands of words a day — yet he’s perfectly happy with the Magic Keyboard that came with his iMac, and can’t stand the noise of a mechanical keyboard like the Norbauer Seneca. Developers are heavy users of the keyboard, too, but that doesn’t inherently mean that they’re mechanical keyboard fans.

So to give a luxury keyboard to people who won’t really appreciate it feels like a waste. For a select few, a Seneca would be the perfect gift. But for the creators of the five best apps of 2025? Maybe.

The winners will be revealed on November 20th, pushed back from November 13th. Stay tuned for more on this — once the winners are announced, I’m going to reach out to hear their thoughts on receiving a Seneca (maybe they’d like to regift it to a certain twelve-year-old writer?).

Educational technology, specifically to monitor student activity, only works as long as it can keep up with the students. As someone who experiences these solutions in use every day, I can affirm that the loopholes might outnumber the blocked sites.

For example, some students might want to play video games during class. A school can make efforts to mitigate this, such as using a platform like Securly to automatically block sites flagged as games, but that only goes so far. A coding class at my school uses MIT’s Scratch to create simple games — which means that they have to unblock it, allowing kids access to any game on that site. Additionally, sites used by teachers for quiz games, such as Blooket and Gimkit, can be repurposed by students with “all-correct” question sets to be used as games. Thirdly, I’ve seen that when a game happens to be unblocked, the link spreads like wildfire from email account to email account. As one example, I know of an unblocked website with an embedded game inside that involved moving a truck through an obstacle course (it looked pretty boring to me, but I guess the bar is pretty low, no pun intended).

That’s not to say a school’s system is necessarily completely ineffective. Last year, my school had left unblocked the spammy-sounding Unblocked Games 66. It’s now blocked, but in a lot of ways, relying on IT to stay a step ahead of tech-savvy students seems like a recipe for ineffective policing.

Here’s something interesting: Even if a school blocks YouTube, there are educational tools teachers can use to create questions and those sites may well be unblocked. If you create a teacher account, you can watch any YouTube video even if the YouTube site itself is blocked.

Blocking sites just encourages kids to find loopholes and can only be a temporary solution. Schools should instead teach kids how to use technology responsibly and to trust the students at least somewhat — of course they should block actually inappropriate content. Is blocking easier for a school? — certainly. But is it actually better in the long run? — maybe not.