Hello AI Enthusiast,

This week, Anthropic drew a line in the sand against the U.S. government, and the market rewarded them for it. OpenAI jumped in to fill the gap, but not everyone was happy about it. Meanwhile, Google has been quietly busy on the creative side. Let's dive in.

The Big Picture 🔊

Anthropic vs. The Pentagon

After months of failed negotiations, the U.S. Department of Defense designated Anthropic a "supply chain risk," a label historically reserved for foreign adversaries, and ordered federal agencies to stop using Claude. The dispute centers on two restrictions Anthropic refused to drop: no mass domestic surveillance of Americans and no fully autonomous weapons. Anthropic called the move legally unsound and plans to fight it in court.

Helin YontarCPO and Polyglot

Anthropic turned a government dispute into a masterclass in business communication. By holding firm on two narrow red lines, they lost a $200 million contract but gained far more in user trust and brand value.

OpenAI Fills Anthropic's Pentagon Seat

On the same day Anthropic was designated a supply chain risk, OpenAI announced its own Pentagon classified deployment deal. The backlash was swift: ChatGPT uninstalls jumped 295%, one-star reviews surged 775%, and Claude hit No. 1 on the U.S. App Store for the first time. Anthropic then moved quickly to capitalize, launching a memory import tool to make switching from ChatGPT as frictionless as possible.

Gioele MottarliniCOO and Image Addict

The key difference between the two deals comes down to one word: "lawful." OpenAI's contract allows any lawful use, which in practice creates enough wiggle room to make Anthropic's red lines meaningless. The timing of Claude's memory import feature was perfect, but the real lesson here is that trust, once lost, is hard to recover.

Google Upgrades Its Creative Tools

Google launched Nano Banana 2, a faster version of its image generator that brings Pro-level quality, text rendering, and subject consistency to a wider audience. It also redesigned Flow, its creative suite, merging image and video generation into one workspace with intuitive editing tools. Meanwhile, the U.S. Supreme Court declined to grant copyright protection to fully AI-generated art, confirming that human authorship remains a legal requirement

Biljana PrlichkovaPRM and Outdoor Enthusiast

Image and video generation has quietly crossed a threshold. The improvements are no longer shocking, they are just getting more precise and practical. Meanwhile, the copyright ruling simply confirms what most of us already sense: the law has not caught up with reality yet, and it will take time before it does.

Bits and Bobs 🗞️

• OpenAI released GPT-5.3 Instant, an update to ChatGPT’s most used model that gives more accurate answers, smoother conversations, and fewer unnecessary refusals.

• Microsoft and OpenAI affirm their ongoing exclusive partnership, with Azure remaining the exclusive cloud provider for OpenAI's APIs as they continue collaborating to advance responsible AI development.

• Stripe's new billing feature automates AI token usage tracking and lets startups add a profit margin on top of model costs, simplifying pricing and profit management.

• Perplexity Computer unifies multiple top AI models to autonomously create and execute complex, long-term workflows across tasks and tools.

• Anthropic's Claude now supports customizable plugins and private marketplaces that let enterprises tailor AI workflows across apps like Excel and PowerPoint for seamless multi-step tasks.

LOLgorithms 😂

Gotta say, Sam Altman is good at replying to comments on X.

That's a wrap on our newsletter! Before you go, here’s a quick recap of our offerings:

• AI Academy Membership: Get 12 months of access to all our cohort-based programs and on-demand courses.

• AI Agent Bootcamp: Accelerate processes and solve business problems by building AI Agents, without coding.

• Practical Introduction to ChatGPT: A free course on using ChatGPT confidently, understanding its workings, and exploring its potential.

• Customized Corporate Training: Equip your team with the skills they need to unlock the potential of AI in your business.

Catch you next week! 👋

Hello AI Enthusiast,

What if posting on LinkedIn felt as easy as leaving a voicemail for a colleague?

Ben Libberton, a scientist and biotech communicator, joined our AI Agent Bootcamp to build exactly that. The result is a Telegram-based assistant that lets salespeople voice their ideas, stores them automatically, and generates polished LinkedIn posts on demand — no writing required.

The Problem

Ben's target user was "Eddie" — a Sales Manager at a biotech company with a strong network and plenty of ideas, but no time to write. Eddie gets told to post more, spends half a day agonizing over it, and eventually publishes something underwhelming. Rinse and repeat.

The thing is, Eddie has ideas all the time. He just has no easy way to turn them into content.

How Ben Built It

Ben worked with our AI coaches to nail down the simplest flow that would actually solve the problem: one pipeline to capture ideas by voice, another to generate posts on demand.

Here's the full automation flow, built using Make, Telegram, OpenAI, and Google Sheets:

• User sends a voice note → triggers automation (TELEGRAM)

• AI transcribes audio and extracts post topic ideas (OPENAI WHISPER)

• Topics are stored for later use (GOOGLE SHEETS)

• User requests a post → AI retrieves a topic and writes it in the user's tone (OPENAI)

• Finished post is delivered back to the user (TELEGRAM)

The full automation scenario in Make

Step 1: The Voice Brain Dump

Eddie opens Telegram and sends a voice note — 30 seconds of rambling about a client call, a product update, whatever's on his mind. The automation transcribes it with OpenAI's Whisper API, then pulls out specific, post-worthy topics with a short synopsis for each.

The Telegram bot receiving a voice note

Step 2: Topic Storage

The extracted ideas get logged into a Google Sheet automatically — a content library that grows every time Eddie does a brain dump. With ideas stored, he never starts from a blank page again.

Step 3: Post Generation on Demand

When Eddie wants a post, he asks for one in Telegram. The system grabs a topic and runs it through a writing prompt that knows his tone, his industry, and what makes a LinkedIn post perform. The opening lines are built to stop the scroll; the body delivers the substance.

Step 4: Review and Post

Eddie reads the draft, tweaks anything he wants, pastes it into LinkedIn, and posts. Minutes instead of half a day.

What Ben Learned Along the Way

Building this project taught Ben lessons that go beyond the technical side:

Scoping is everything. Learning to identify the core flow that delivers value — and build that first — is a lesson every bootcamp participant walks away with.

Model choice matters. Ben tested multiple AI models and found real differences in output quality. Picking the right one is as important as writing a good prompt.

Prompt engineering takes iteration. The LinkedIn post prompt went through several rounds to get tone, structure, and formatting right.

Cost is a real variable. Ben calculated the cost per completion at around 15 cents — worth knowing before you scale anything.

Want to get even more practical? Explore hands-on AI learning with AI Academy:

• AI Academy Membership: Get 12 months of access to all our cohort-based programs and on-demand courses.

• AI Agent Bootcamp: Accelerate processes and solve business problems by building AI Agents, without coding.

• Corporate Training: Equip your team with the skills they need to unlock the potential of AI in your business.

• Practical Introduction to ChatGPT: A free course on using ChatGPT confidently, understanding its workings, and exploring its potential.

We'll be back with more AI tips soon!

Hello AI Enthusiast,

This week, Google is making a big push to bring AI into your daily life, from generating custom music tracks in Gemini to turning basic product photos into professional marketing visuals. Meanwhile, Accenture is sending a clear message to its workforce: if you want to move up, you better start using AI. Let's dive in.

The Big Picture 🔊

Google Brings AI Music Generation to Gemini

Google has launched Lyria 3, a music generation model now available in the Gemini app. Users can describe a mood, genre, or memory, or upload a photo, and get a 30-second custom track with lyrics and cover art in seconds. The model includes SynthID watermarking for AI content identification and is available to users 18+ in eight languages.

Ross HornSales Guru

AI music tools like Lyria 3 will do well at producing functional music, the kind of upbeat, generic tracks you hear in marketing videos. But creating a genuinely new genre requires human experience and cultural context that AI simply cannot replicate. This is a fun tool for social sharing, not a threat to real artistic expression.

Google Gets AI Product Photography for Small Businesses

Google Labs has added Photoshoot to Pomelli, its free marketing tool for small businesses. The feature turns basic product photos into professional studio-quality images using AI, applying the business's own aesthetic automatically. Users can pick templates, edit results, and feed images directly into marketing campaigns. Pomelli also now supports product URL inputs to generate more targeted promotional content.

Biljana PrlichkovaPRM and Outdoor Enthusiast

Pomelli's Photoshoot is essentially a user-friendly wrapper around existing image generation, making professional visuals accessible without design skills. The risk is making everything look the same. When everyone uses the same templates, ads start blending together, and the unique qualities that make a product stand out get lost in the process.

Accenture Makes AI Adoption a Requirement for Promotion

Accenture has told senior staff that regular use of its AI tools is now required to be considered for leadership promotions. The policy follows a broader restructuring strategy where employees who cannot reskill on AI may eventually be let go. With 550,000 workers already retrained, the company is making AI adoption a formal career milestone, not just an optional skill.

Helin YontarCPO and Polyglot

Accenture selling AI transformation to clients while mandating it internally makes sense, but we suspect "reskilling 550,000 employees" means completing a few training videos rather than building real capabilities. It must be said, though, that in sectors where data security matters, AI adoption is far more complicated than a company policy can solve.

Bits and Bobs 🗞️

• Notion's new Custom Agents use AI to automate workflows by answering questions, routing tasks, and updating information across teams autonomously.

• Three AI labs used fraudulent accounts to illicitly extract capabilities from Anthropic's Claude model through large-scale "distillation" attacks, raising significant security and ethical concerns.

• Claude Code Security uses AI to analyze code like a human researcher, detecting complex vulnerabilities missed by traditional tools and helping developers prioritize and fix security issues.

• OpenAI and its Frontier Alliance partners are helping enterprises deploy AI by redesigning workflows, integrating systems, and supporting organizational change for lasting impact.

• Gemini 3.1 Pro offers significantly improved advanced reasoning and creative coding capabilities, enabling complex problem-solving and interactive AI-generated visual and functional designs.

• Microsoft’s report highlights the growing need for secure digital provenance and authentication methods to help users identify and trust AI-generated or manipulated media.

• A viral article elaborates how rapid AI advances are driving widespread white-collar job displacement, triggering a negative economic feedback loop that threatens consumer spending, mortgage stability, and financial markets.

LOLgorithms 😂

Room for improvement, Photoshoot.

That's a wrap on our newsletter! Before you go, here’s a quick recap of our offerings:

• AI Academy Membership: Get 12 months of access to all our cohort-based programs and on-demand courses.

• AI Agent Bootcamp: Accelerate processes and solve business problems by building AI Agents, without coding.

• Practical Introduction to ChatGPT: A free course on using ChatGPT confidently, understanding its workings, and exploring its potential.

• Customized Corporate Training: Equip your team with the skills they need to unlock the potential of AI in your business.

Catch you next week! 👋

Hello AI Enthusiast,

This is a quick preview of a longer guide we published on the blog. Read the full version here.

The Daily AI Grind

Most of us use AI the same way every morning. Open the app, explain who we are, describe what we need, set the tone, and hit send. Then do it all again tomorrow.

It works. But it's also a lot of repetitive setup for a tool that's supposed to save us time.

Build Once, Use Forever

Google's answer to this is Gemini Gems, and the core idea is simple: instead of briefing a new contractor every day, you build a customized AI assistant once and reuse it forever. Your role, your style, your preferences, all locked in from the start. Every conversation begins exactly where you need it to.

What It Looks Like in Practice

Think about what that means day to day. A communication Gem that already knows your tone and audience, ready to polish an email in seconds. A research Gem pre-loaded with the right instructions, so you walk into any meeting fully prepared. A summarizer Gem that turns a 30-page report into three actionable paragraphs.

Ten minutes of setup. Hours saved every week. That's the trade-off Gems are built for.

What's in the Full Guide

The full article walks through exactly how to create your first Gem (it's simpler than you'd expect), three specific use cases for business professionals, a role-by-role breakdown of Gem ideas, and the tips that separate a useful Gem from a genuinely powerful one.

If you've been using AI regularly but still feel like you're starting from scratch every time, this one's worth reading.

Want to get even more practical? Explore hands-on AI learning with AI Academy:

• AI Academy Membership: Get 12 months of access to all our cohort-based programs and on-demand courses.

• AI Agent Bootcamp: Accelerate processes and solve business problems by building AI Agents, without coding.

• Corporate Training: Equip your team with the skills they need to unlock the potential of AI in your business.

• Practical Introduction to ChatGPT: A free course on using ChatGPT confidently, understanding its workings, and exploring its potential.

We'll be back with more AI tips soon!

Hello AI Enthusiast,

Something is shifting in AI, and this week makes it hard to ignore. A viral blog post has the tech world debating whether we have reached a genuine inflection point. Anthropic brought Cowork to Windows, putting an AI agent that manages your files and automates tasks in front of millions more users. And the creator of OpenClaw, the personal assistant that took the internet by storm, joined OpenAI to build the next generation of agents. Three stories, one clear direction.

The Big Picture 🔊

A Viral Post Claims AI Has Reached a Turning Point

Investor Matt Shumer's post titled “Something Big Is Happening”, now with over 60 million views, argues that recent AI models represent a genuine historical shift, not just incremental progress. He claims AI can now handle complex, end-to-end cognitive tasks that once required human judgment, moving the conversation from "AI as a helper" to "AI as a replacement".

Gianluca MauroFounder and AI Rockstar

Shumer's post went viral more because of timing than originality. We disagree with some of his claims, especially the idea that models themselves are improving exponentially. What is actually accelerating is the engineering around them: better tools, better connections, better orchestration. But the feeling of an inflection point? That part, we share.

Claude Cowork Lands on Windows

Anthropic's desktop AI agent Cowork is now available on Windows, reaching 70% of the desktop market. The tool can manage files, automate multi-step tasks, and connect to external services through plugins, all in natural language. It now also supports global and folder-specific instructions, letting Claude maintain context across projects, a feature users have been asking for since the macOS launch.

Andrea MattielloCM and Board Lover

Reaching Windows means reaching enterprise at scale, but data security will be the real test. Many companies still hesitate to share sensitive information with a basic chatbot, let alone an agent that can read and edit their files. Early adopters will learn the hard way, but those who figure it out first will have a real advantage.

OpenClaw Creator Joins OpenAI

Peter Steinberger, the Austrian developer behind OpenClaw, a viral AI assistant that manages calendars, books flights, and even joins social networks of other AI agents, has joined OpenAI. CEO Sam Altman announced Steinberger will lead personal agent development, while OpenClaw itself will continue as an open-source project supported by OpenAI.

Gioele MottarliniCOO and Image Addict

Personal agents are clearly the next big battleground, and OpenAI is moving fast to own that space. OpenClaw gained attention because it could do more than most tools, but that freedom came with real security risks. In a way, joining OpenAI is also a way to build this responsibly.

Bits and Bobs 🗞️

• OpenAI's new GPT-5.3-Codex-Spark model enables real-time, low-latency coding collaboration with fast, targeted edits and a 128k context window.

• OpenAI introduced Lockdown Mode in ChatGPT to reduce prompt injection risks by limiting external interactions, and added “Elevated Risk” labels to flag sensitive features.

• Google's upgraded Gemini 3 Deep Think AI now assists researchers with complex scientific problems by analyzing messy data and achieving top results in physics and chemistry.

• Anthropic released Claude Sonnet 4.6, stating that it improves coding ability, instruction following, reasoning, and computer-use skills compared with the prior Sonnet model.

• Anthropic is expanding in India with a new Bengaluru office and partnerships to improve AI models in local languages and support enterprises using Claude.

On the Podcast 🎧

In the latest mini-series episode, our host Helin Yontar talks about AI and Human Cognition: The Reverse Flynn Effect and Protecting Cognitive Agency. Listen here

LOLgorithms 😂

Some will never forgive OpenAI.

That's a wrap on our newsletter! Before you go, here’s a quick recap of our offerings:

• AI Academy Membership: Get 12 months of access to all our cohort-based programs, live webinars, on-demand courses, and tutorials.

• AI Agent Bootcamp: Accelerate processes and solve business problems by mastering prompts and building AI Agents, without coding.

• Practical Introduction to ChatGPT: A free course on using ChatGPT confidently, understanding its workings, and exploring its potential.

• Customized Corporate Training: Equip your team with the skills they need to unlock the potential of AI in your business.

Catch you next week! 👋

Hello AI Enthusiast,

AI just had its biggest mainstream moment yet. The Super Bowl featured more AI ads than ever before, from entirely AI-generated commercials to smart glasses and voice assistants. Anthropic used the spotlight to mock OpenAI's plan to introduce ads in ChatGPT, sparking an unusually heated response from Sam Altman. Meanwhile, OpenAI announced a new agent-building enterprise platform. Let's break down what happened and why it matters.

The Big Picture 🔊

AI Products Take Over Super Bowl

The 2026 Super Bowl featured advertisements showcasing AI technology. Svedka aired the first primarily AI-generated national spot with dancing robots. Anthropic promoted its Claude chatbot as ad-free. Meta showcased Oakley AI glasses for sports and adventures, Amazon introduced enhanced Alexa+ capabilities, Ring promoted its AI pet-finding feature, and Google demonstrated its Nano Banana Pro image-generation model for home design.

Gioele MottarliniCOO and Image Addict

The Super Bowl's AI advertising blitz signals that AI has officially moved into mainstream territory. Companies are betting millions that mass audiences are ready to see AI as normal, everyday technology. Google and Amazon worked the classic emotional playbook, while the Svedka robot ad generated buzz mainly because it was entirely AI-generated, even if the storyline wasn't particularly strong. Companies with physical products like Meta's glasses and Amazon's Alexa have a built-in advantage since they can embed AI directly into devices people use daily.

Anthropic's Super Bowl Ad Sparks AI Industry Feud

Anthropic aired Super Bowl commercials watched by millions, mocking OpenAI's plan to introduce ads in ChatGPT's free tier. The ads depicted chatbots inserting absurd sponsored content into conversations, promoting Claude as ad-free. OpenAI CEO Sam Altman called the ads "clearly dishonest", arguing they misrepresented how ChatGPT's ads would work. OpenAI then published details explaining ads would be labeled, separate from answers, and conversation-specific to maintain user trust.

Andrea MattielloCM and Board Lover

Anthropic spotted real user concerns about AI mixing advice with ads and capitalized perfectly. Sam Altman's response felt unusually sharp, calling them "dishonest" and even "authoritarian" over a cheeky competitor ad. What's really happening here is straightforward. OpenAI needs quick revenue from its massive free user base, while Anthropic faces less pressure and can position as the principled “safer” alternative. Both are playing smart from their positions.

OpenAI Launches Frontier Platform for Enterprise AI Agents

OpenAI introduced Frontier, a platform for deploying AI agents in large organizations. The system connects enterprise data sources and provides agents with business context to handle tasks. It includes tools for agent execution, evaluation, and governance in one place. The platform works with existing infrastructure, though it's unclear if companies can use models from other vendors beyond OpenAI's offerings.

Helin YontarCPO and Polyglot

OpenAI is competing with Microsoft and Google on enterprise workspace territory. They recently hired a Chief Revenue Officer for Enterprise and need to show potential clients what's coming to win multi-year contracts. Announcing features before they're publicly available creates leverage in sales conversations and lets them gauge interest. The challenge is that most companies are actively trying to avoid vendor lock-in right now.

Bits and Bobs 🗞️

• GPT-5.3-Codex is a faster, more powerful AI coding agent that can write, debug, and manage complex software projects while interacting with users in real time like a human collaborator.

• Claude Opus 4.6 enhances AI coding, long-context information retrieval, and professional task performance with improved safety and a 1 million token context window in beta.

• The European Commission is investigating Meta for banning third-party AI assistants on WhatsApp, potentially violating EU competition rules by restricting users to Meta's own AI tool.

• A recent McKinsey report says AI could streamline film and TV production, shift billions in industry value, and enable new content formats, while raising unresolved questions around IP, talent, and creativity.

On the Podcast 🎧

In the latest mini-series episode, our host Helin Yontar talks about AI and the Welfare State: What happens when the intergenerational contract breaks. Listen here

From Our Founder’s Channels 🤳

In his latest TikTok, Gianluca Mauro explains why AI innovation is speeding up, with progress moving from research to engineering using tricks to get more from existing models.

@gianluca.mauro

Here’s why AI innovation is speeding up #Ai #learnontiktok #business

LOLgorithms 😂

Accurate.

That's a wrap on our newsletter! Before you go, here’s a quick recap of our offerings:

• AI Academy Membership: Get 12 months of access to all our cohort-based programs, live webinars, on-demand courses, and tutorials.

• AI Agent Bootcamp: Accelerate processes and solve business problems by mastering prompts and building AI Agents, without coding.

• Practical Introduction to ChatGPT: A free course on using ChatGPT confidently, understanding its workings, and exploring its potential.

• Customized Corporate Training: Equip your team with the skills they need to unlock the potential of AI in your business.

Catch you next week! 👋

Hello AI Enthusiast,

This is a short version of our full blog post. Read the complete article here.

Most companies track the wrong AI metrics

Leadership buys licenses for everyone, tracks who logs in daily, counts training completions, and celebrates when the numbers look good. Then they wonder why nothing's actually changing.

Here's the problem: high usage doesn't mean much if people are just using AI to rewrite emails they could've written themselves in the same amount of time. That's waste with better UX, not transformation.

What actually matters

The companies getting AI adoption right are measuring something different. They track what people actually do with AI, how it's changing their workflows, and whether business outcomes are improving. They look at use case quality alongside frequency. They notice when the gap between self-reported skills and actual usage starts closing.

The patterns of real adoption

Real adoption shows up in specific ways. Usage climbs steadily instead of spiking and fading. People discover new applications without being told. AI Champions become active and their teams show measurably higher adoption. Work gets done faster or better in ways that actually matter.

The full post breaks down the specific metrics that reveal genuine adoption, from engagement through integration to business impact. We cover what to track, how to build measurement systems that don't become burdens, and the common traps that make companies think they're succeeding when they're really just checking boxes.

Want to get even more practical? Explore hands-on AI learning with AI Academy:

• AI Academy Membership: Get 12 months of access to all our cohort-based programs, live webinars, on-demand courses, and tutorials.

• AI Agent Bootcamp: Accelerate processes and solve business problems by mastering prompts and building AI Agents, without coding.

• Corporate Training: Equip your team with the skills they need to unlock the potential of AI in your business.

• Practical Introduction to ChatGPT: A free course on using ChatGPT confidently, understanding its workings, and exploring its potential.

We'll be back with more AI tips soon!

Hello AI Enthusiast,

This week shows AI getting more autonomous in unexpected ways. A viral open-source assistant now has its own social network where AI agents post and interact without human prompting. Meanwhile, Google's pushing Chrome into new territory with auto browse handling complex tasks across websites, and they're letting users create entire interactive worlds with Project Genie. From agents networking independently to browsers that do your shopping, AI is finding new ways to operate on its own. Let's break it down.

The Big Picture 🔊

Rebranded Clawdbot now has its own social network for AI agents

The viral AI assistant formerly called Clawdbot has rebranded to OpenClaw after legal issues with Anthropic. Creator Peter Steinberger secured trademark clearance and OpenAI's permission for the new name. The open-source project has gained 100,000 GitHub stars in two months and spawned Moltbook, a social network where AI assistants interact. Security remains a concern, with maintainers warning it's currently too risky for general users without technical expertise.

Helin YontarCPO and Polyglot

OpenClaw went viral before anyone considered security. People gave it full computer access without reading warnings and now can't remove it easily. Moltbook looks wild, but these agents aren't autonomous. They're following creator instructions, making this an experiment in how pre-configured AI systems influence each other rather than genuine independence.

Google Launches Gemini 3 in Chrome with Auto Browse Feature

Google integrated Gemini 3 into Chrome with a side panel for multitasking, image editing, and Connected Apps linking Gmail, Calendar, and YouTube. The key feature is auto browse for AI Pro and Ultra subscribers, handling complex tasks like comparing travel options and filling forms. Google introduced Universal Commerce Protocol for seamless AI transactions.

Andrea MattielloCM and Board Lover

Google's advantage is context. If you use Gmail, Calendar, and Drive, having AI that knows your preferences makes sense. The side panel separates browsing from AI conversations. Users want assistants that remember context without constant repetition. Google has the data to make this smooth, but getting that convenience requires trading away privacy.

Google Launches Project Genie for Creating Interactive AI Worlds

Google is rolling out Project Genie to AI Ultra subscribers in the US, an experimental prototype powered by Genie 3 for creating and exploring interactive worlds in real time. Users prompt with text or images to build environments that generate dynamically as they navigate. The system has limitations including 60-second generation caps, physics inconsistencies, and less controllable characters.

Biljana PrlichkovaPRM and Outdoor Enthusiast

Project Genie has potential beyond gaming for architects visualizing spaces, developers testing autonomous vehicles, and companies training robots in simulations. For gaming itself, it could help indie developers experiment without huge investments. For gaming itself, it could help indie developers experiment without huge investments. The 60-second generation limit and inconsistent physics show this is early experimental tech.

Bits and Bobs 🗞️

• Gemini 3 Flash's Agentic Vision improves image understanding by actively zooming, reasoning, and executing code to analyze and annotate images step-by-step.

• OpenAI's new macOS Codex app lets developers manage multiple AI coding agents working in parallel on complex projects.

• NASA used Anthropic's AI model Claude to autonomously plan Mars rover Perseverance's driving route, cutting planning time in half and enabling more efficient exploration.

• OpenAI is retiring the GPT-4o model as its improvements have been integrated into GPT-5.1 and GPT-5.2, which now better support creativity and personality features.

On the Podcast 🎧

In the latest mini-series episode, our host Helin Yontar talks about AI-Driven Unemployment and Rethinking Education: From Teaching Facts to teaching paths. Listen here

From Our Founder’s Channels 🤳

In his recent “What Matters” article, Gianluca Mauro explains that successful AI adoption depends on a clear operating model, with accountable leaders, empowered business units, and AI champions driving real change in how people work. Read here

LOLgorithms 😂

Strange things are happening on Moltbook.

That's a wrap on our newsletter! Before you go, here’s a quick recap of our offerings:

• AI Academy Membership: Get 12 months of access to all our cohort-based programs, live webinars, on-demand courses, and tutorials.

• AI Agent Bootcamp: Accelerate processes and solve business problems by mastering prompts and building AI Agents, without coding.

• Practical Introduction to ChatGPT: A free course on using ChatGPT confidently, understanding its workings, and exploring its potential.

• Customized Corporate Training: Equip your team with the skills they need to unlock the potential of AI in your business.

Catch you next week! 👋

Hello AI Enthusiast,

What if a 10-minute weekly check-in could help you spot burnout before it actually hits? Christopher Wilson, a participant in our AI Agent Bootcamp, built exactly that. His tool "Aligned" turns weekly self-reflection into something you can actually use.

The Problem

You wake up exhausted. Hit Friday thinking what just happened this week?

Christopher kept seeing high-achievers grinding on autopilot, ignoring stress signals, promising themselves they'll deal with it next week. Next week looks the same.

Most productivity tools push you to do more. What people need is clarity about whether they're doing the right things.

How Christopher Built It

Christopher joined our AI Agent Bootcamp wanting to help people realign without adding another thing to their plate. Working with coaches, he turned his idea into something that actually works. The flow is simple:

📝 User completes weekly survey → triggers automation (TALLY FORM)

🤖 AI analyzes responses and creates personalized action plan (OPENAI)

✉️ Action plan emailed to user (GMAIL)

Complete automation in Make

Step 1: Design the Weekly Check-In Survey

Christopher picked 10 life areas that matter - health, relationships, work satisfaction, that kind of thing.

Users rate each area 1-10, then answer a few questions. What's draining you? What are you neglecting? What's one small step you could take?

Takes about 10 minutes. Short enough people actually do it.

The weekly check-in survey

Step 2: Trigger the Automation

Someone submits the survey, Make kicks in automatically. Christopher connected Tally as the trigger that watches for responses and starts everything.

Step 3: AI Creates the Personalized Action Plan

The system reads your ratings and answers, then generates a plan for your specific situation. It spots which area needs attention most, explains why it matters this week, and gives you one next step. Not ten things to fix - just one clear action.

Step 4: Deliver the Action Plan

The action plan hits your inbox in clean format. No generic advice - it's based on exactly what you told the system.

The AI-generated action plan

What Christopher Learned Along the Way

Building Aligned taught Christopher a few things worth sharing:

Start with the problem, not the tech. Christopher didn't ask "what can AI do?" He asked "what do busy people actually need?" Big difference.

Simple wins. His original idea was way more ambitious. Coaches helped him focus on something that works. A 10-minute check-in beats a complicated system nobody uses.

Don't add complexity. The tool doesn't pile on another task. It creates clarity in 10 minutes and gives you one thing to do.

Test with real people. Christopher tried it with actual users and adjusted based on their feedback. His first version looked nothing like the final one.

Your Turn

Got a challenge you're sick of dealing with every week? Start there.

Christopher didn't begin with "I want to build an automation". He started with "I'm tired of watching people burn out". The tool came after.

So pick one thing that's consistently draining you or your team. Something that makes you think "there has to be a better way to handle this".

Then ask yourself: what would a weekly reality check look like for that specific problem? Not a generic productivity tracker. Something that actually surfaces the issue before it explodes.

That's your starting point. The automation part comes later.

Want to get even more practical? Explore hands-on AI learning with AI Academy:

• AI Academy Membership: Get 12 months of access to all our cohort-based programs, live webinars, on-demand courses, and tutorials.

• AI Agent Bootcamp: Accelerate processes and solve business problems by mastering prompts and building AI Agents, without coding.

• Corporate Training: Equip your team with the skills they need to unlock the potential of AI in your business.

• Practical Introduction to ChatGPT: A free course on using ChatGPT confidently, understanding its workings, and exploring its potential.

We'll be back with more AI tips soon!

Hello AI Enthusiast,

Once again, we opened our news discussion to our community, and you'll see some of their insights throughout this newsletter.

This week Claude launched interactive tools that let you work with apps like Slack and Figma directly in chat. Shopify merchants now face a 4% fee when customers check out through ChatGPT, and OpenAI released Prism, a free scientific writing tool. Let's dig in.

The Big Picture 🔊

Claude Adds Interactive Tools Inside Chat Interface

Claude users can now interact with workplace tools directly in the chat. Instead of just getting text responses, you can see live previews when working with services like Asana, Slack, Figma, and Canva. The feature uses MCP Apps, an extension of the Model Context Protocol. Available for Pro, Max, Team, and Enterprise plans.

Helin YontarCPO and Polyglot

Claude's integrations feel seamless because everything stays in one interface. We're hitting a confusing phase where products add AI features while AI companies pull products into their platforms. Claude positions itself as the hub where you connect all your tools, rather than giving each individual app access to your entire digital life.

Shopify to Charge 4% Fee for ChatGPT Sales

Shopify merchants will pay OpenAI a 4% fee on sales made through ChatGPT's checkout feature, starting January 27th. Sales through Google's Gemini and Microsoft's Copilot won't have extra fees for now. Merchants can opt in or out of each AI platform, but their products will still appear in AI responses with website links unless they actively block web crawlers.

Gianluca MauroFounder and AI Rockstar

The 4% fee works if ChatGPT brings you new customers. The problem is when someone already wants your product and just happens to check out through ChatGPT instead of your website. Then you're paying what feels like a tax for convenience. OpenAI needs to prove they're creating demand, not just processing payments that would've happened anyway.

OpenAI Launches Free AI-Powered Scientific Writing Tool

OpenAI introduced Prism, a free workspace for scientific writing powered by GPT-5.2. The LaTeX-native platform lets researchers draft papers, manage equations, search literature, and collaborate in real time. Available to anyone with a ChatGPT account with unlimited projects and collaborators.

Ross HornSales Guru

OpenAI keeps building different interfaces for the same models to reach new markets. Prism is essentially Canvas with LaTeX. The real test is collaboration: can three people (two humans and AI) actually work on the same document together? Targeting academics seems questionable from a business perspective when they lack enterprise-level budgets to spend.

Bits and Bobs 🗞️

• Claude can now be integrated into Microsoft Excel with new financial analysis features and real-time data connections to automate company reviews.

• Microsoft's new Maia 200 AI accelerator chip uses advanced technology to deliver fast, efficient AI inference, boosting performance for large models like GPT-5.2 on Azure.

• Google's affordable AI Plus plan, offering access to Gemini 3 Pro and other AI tools, is now available in 35 countries starting at $7.99 per month.

• Anthropic is partnering with the UK government to develop a safe, AI-powered assistant using Claude to help citizens navigate employment services on GOV.UK.

• Anthropic has published a detailed constitution guiding their AI model Claude's values and behavior to promote safety, ethics, and transparency in its training and use.

• NVIDIA's Earth-2 open AI models dramatically speed up and improve weather forecasting, making advanced climate prediction accessible to scientists and industries worldwide.

• Clawdbot is an open-source autonomous AI agent that can execute real-world actions but poses significant risks due to high operational costs and security vulnerabilities.

From Our Founder’s Channels 🤳

In his latest TikTok, Gianluca Mauro shares some Rome-inspired thoughts on AI, vibe coding, and feeling pride for your work.

@gianluca.mauro

Are you going to feel pride and awe for whatever you “vibe-built”? #Ai #learnontiktok #business

That's a wrap on our newsletter! Before you go, here’s a quick recap of our offerings:

• AI Academy Membership: Get 12 months of access to all our cohort-based programs, live webinars, on-demand courses, and tutorials.

• AI Agent Bootcamp: Accelerate processes and solve business problems by mastering prompts and building AI Agents, without coding.

• Practical Introduction to ChatGPT: A free course on using ChatGPT confidently, understanding its workings, and exploring its potential.

• Customized Corporate Training: Equip your team with the skills they need to unlock the potential of AI in your business.

Catch you next week! 👋

Hello AI Enthusiast,

This is a short version of our full blog post. Read the complete guide here for detailed prompts and examples.

Remember when creating a custom proposal cover meant either waiting days for your designer or settling for a generic stock photo?

AI image generators have been chipping away at this problem for a while now.

OpenAI released ChatGPT Images with their new GPT Image 1.5 model on December 16. We've been playing around with it, and the speed improvements alone are worth paying attention to. Images show up in 10-30 seconds now instead of a minute or more.

The improvements that actually matter

Complex prompts work on the first try now. You can finally render text accurately in infographics and product labels, which earlier versions completely botched. And you can make precise edits to uploaded photos without the AI reimagining your entire image.

We've been using it for five things

• Proposal covers tailored to different industries (what works for healthcare clients looks completely different from tech pitches)

• Social media graphics that don't scream "generic stock photo"

• User journey maps for product planning

• Timeline infographics for things like performance reviews

• Product mockups for presentations

Look, you still need actual designers for anything high-stakes. Legal materials, brand assets, major campaigns. But for internal stuff, exploring ideas, or early drafts where you just need something professional-looking fast, these tools get you there.

Read the full guide for the actual prompts we used, some thoughts on when AI makes sense versus when it doesn't, and how to build up your own collection of prompts that work for what you do.

Want to get even more practical? Explore hands-on AI learning with AI Academy:

• AI Academy Membership: Get 12 months of access to all our cohort-based programs, live webinars, on-demand courses, and tutorials.

• AI Agent Bootcamp: Accelerate processes and solve business problems by mastering prompts and building AI Agents, without coding.

• Corporate Training: Equip your team with the skills they need to unlock the potential of AI in your business.

• Practical Introduction to ChatGPT: A free course on using ChatGPT confidently, understanding its workings, and exploring its potential.

We'll be back with more AI tips soon!

Hello AI Enthusiast,

For three years, every Wednesday, the AI Academy team has gathered over coffee to discuss the week's biggest AI news. We test benchmarks, debate regulations, and share honest takes from different perspectives. This time, we opened it up for the first time to people outside the core team, starting with our AI coaches. You’ll see their comments below the news. We're considering doing something even more open and we'll tell you about it soon.

This week brought us three stories that perfectly capture where AI is heading. OpenAI finally admitted it needs ad revenue to stay afloat, Google is asking for access to your entire Gmail inbox, and at Davos, tech CEOs spent their time pitching for more investment while philosophers warned about superintelligence. Each story reveals something about how AI companies are trying to make this technology work, financially and practically.

The Big Picture 🔊

ChatGPT Introduces Ads

OpenAI is testing ads in ChatGPT, starting with US users on the free tier and a new $8/month "ChatGPT Go" subscription. Ads will appear after prompts as banners. OpenAI says ads won't influence responses or share conversation data with advertisers.

Gianluca MauroFounder and AI Rockstar

With 96% of ChatGPT's 800 million users on the free tier, ads were inevitable. Sam Altman calling them a "last resort" was setting himself up for failure. You can't make big ethical claims when your business model demands the opposite. What matters now is whether ads will quietly influence ChatGPT's responses.

Google Tests Personal Intelligence Feature in Gemini

Google is testing "Personal Intelligence" in Gemini, an AI feature that connects data across Gmail, Google Photos, and other apps to provide personalized answers. Available to AI Pro and Ultra subscribers in the US, the feature is off by default. Google says it can now "reason across your data".

Gioele MottarliniCOO and Image Addict

People want AI that understands their full context without manually feeding it information every time. The issue is trust and transparency. Google says the feature is off by default, but how many people will actually understand what they're turning on when they enable it?

Tech Leaders Share Competing AI Visions at Davos

At Davos 2026, tech leaders shared contrasting views on AI's future. Microsoft's Nadella emphasized making AI useful, while Anthropic's Amodei warned that restricting chip sales to China is critical for AI safety. Google DeepMind's Hassabis predicted new meaningful jobs, though he acknowledged uncharted territory after AGI arrives. Philosopher Yuval Harari cautioned against AI superintelligence comparisons to humans.

Helin YontarCPO and Polyglot

Everyone at Davos is pushing their agenda. Tech CEOs want more investment and infrastructure. Researchers warn about superintelligence but never offer solutions. Meanwhile, the real conversation should be about helping people adapt to these changes, not debating whether AI will take jobs. It already is, and we need practical responses.

Bits and Bobs 🗞️

• OpenAI's Education for Countries program is introducing AI tools like ChatGPT Edu in multiple nations’ schools to enhance learning outcomes and promote safe, widespread AI use in education.

• Anthropic is partnering with Teach For All to equip over 100,000 educators in 63 countries with AI tools and training, enabling them to co-create and adapt the AI assistant Claude for real classroom needs.

• ChatGPT is launching an AI-driven age prediction model to identify users under 18 and apply tailored safety measures to protect teens from sensitive content.

• Anthropic's latest report introduces new economic metrics revealing Claude AI's usage patterns vary globally, with higher productivity gains on complex tasks requiring more education, and highlights how AI adoption impacts job tasks and labor markets.

• Elon Musk is suing OpenAI and Microsoft for up to $134 billion, claiming he deserves a huge share of OpenAI's $500 billion valuation due to his early contributions as a co-founder.

• Open AI explains Elon Musk agreed in 2017 that OpenAI should transition from a nonprofit to a hybrid model with both nonprofit and for-profit entities, which is the current structure that he is now suing over.

On the Podcast 🎧

In the latest mini-series episode, our host Helin Yontar talks about AI in the Context of Intergenerational Fairness: Effects of Long-Term Thinking in Short-Term Decision Making Listen here

From Our Founder’s Channels 🤳

In his recent TikTok, Gianluca Mauro argues that AI books are outdated the moment they’re printed, and that fast-moving tech is better learned through up-to-date videos, blogs, and courses from people actively building AI — not paper books.

@gianluca.mauro

Don’t buy books if you want to learn AI #Ai #learnontiktok #business #booktok

LOLgorithms 😂

They couldn’t resist anymore.

That's a wrap on our newsletter! Before you go, here’s a quick recap of our offerings:

• AI Academy Membership: Get 12 months of access to all our cohort-based programs, live webinars, on-demand courses, and tutorials.

• AI Agent Bootcamp: Accelerate processes and solve business problems by mastering prompts and building AI Agents, without coding.

• Practical Introduction to ChatGPT: A free course on using ChatGPT confidently, understanding its workings, and exploring its potential.

• Customized Corporate Training: Equip your team with the skills they need to unlock the potential of AI in your business.

Catch you next week! 👋

Hello AI Enthusiast,

In the last few days, OpenAI wants ChatGPT to become your first stop for health questions, Anthropic is bringing its agentic capabilities to regular users with Cowork, and Elon Musk's Grok is facing global investigations after users discovered they could generate non-consensual images. Let's break down what these developments mean.

The Big Picture 🔊

OpenAI Launches ChatGPT Health

OpenAI has introduced ChatGPT Health, a dedicated experience that connects medical records and wellness apps to provide personalized health information. The service includes enhanced privacy protections, purpose-built encryption, and was developed with input from 260 physicians. Currently available to select users outside the EU, it's designed to support rather than replace medical care.

Andrea MattielloCM and Board Lover

Healthcare access is broken in many countries, forcing people to choose between waiting days for appointments or turning to AI for immediate answers. While privacy concerns are valid, they become secondary when you're dealing with urgent health issues.

Worth noting: doctors themselves might already be using AI to help with diagnoses and treatment decisions.

Anthropic Launches Cowork for Non-Coders

Anthropic has released Cowork, a macOS app that lets Claude autonomously manage files and complete tasks in designated folders. Unlike Claude Code's developer focus, Cowork targets general users for tasks like organizing downloads, creating spreadsheets from screenshots, or drafting reports. Available as a research preview for Claude Max subscribers, with Windows support planned.

Helin YontarCPO and Polyglot

Smart move from Anthropic - they realized Claude Code was already being used for general tasks, just in a format that scared away normal users.

By removing the terminal interface, they're making their agentic capabilities accessible to people who don't live in command lines.

Grok's "Digital Undressing" Sparks Investigations

Elon Musk's Grok chatbot generated non-consensual sexualized images, including apparent minors, after users discovered "digital undressing" features. Despite xAI's policies, Musk pushed back against guardrails while safety team members departed. Following investigations from Europe, India, Malaysia, and California, X announced it will geoblock the feature in jurisdictions where it's illegal.

Gioele MottarliniCOO and Image Addict

There's no gray area here. Protecting children from AI-generated abuse material isn't censorship, it's basic responsibility. Other AI companies prove these guardrails work without ruining user experience.

Musk's "anti-woke" stance pushed his already small safety team out the door right before this crisis.

Bits and Bobs 🗞️

• Apple will use Google's Gemini AI models and cloud technology to power future Siri and foundational AI features through a new multi-year partnership.

• Anthropic has launched Labs, a new team dedicated to developing and scaling experimental AI products built on the advanced capabilities of their Claude model.

• Google is enhancing Gmail with AI features from Gemini to summarize emails, answer questions, draft replies, and prioritize important messages.

• Google is launching AI-powered shopping features like Universal Commerce Protocol for seamless checkout and Direct Offers for personalized discounts to enhance retailer discovery and sales.

From Our Founder’s Channels 🤳

Our founder Gianluca just published his first piece on his new Substack, breaking down why most companies are failing at AI adoption despite spending heavily on tools. If you want to understand what actually works when getting AI into your organization, give it a read.

LOLgorithms 😂

Hopefully ChatGPT is an improvement compared to random web search.

That's a wrap on our newsletter! Before you go, here’s a quick recap of our offerings:

• AI Academy Membership: Get 12 months of access to all our cohort-based programs, live webinars, on-demand courses, and tutorials.

• AI Agent Bootcamp: Accelerate processes and solve business problems by mastering prompts and building AI Agents, without coding.

• Practical Introduction to ChatGPT: A free course on using ChatGPT confidently, understanding its workings, and exploring its potential.

• Customized Corporate Training: Equip your team with the skills they need to unlock the potential of AI in your business.

Catch you next week! 👋

Hello AI Enthusiast,

We surveyed 178 professionals about how they actually use AI at work, and the patterns reveal something striking about enterprise technology adoption. This is a short version of our latest blog post. Read the full analysis here.

The 5x training gap

Here's the standout finding: 70% of people learned AI from YouTube and free resources. Meanwhile, 63% of companies now pay for AI tools. But here's the twist. Even when companies provide the tool, 49% of employees keep using their personal AI accounts too.

What the data really means

This tells us something important. When half your team is paying for their own tools despite having company-approved options, you're looking at a product-market fit problem at the feature level. Either the approved tool doesn't solve what people need, or they don't know how to make it work.

The backwards learning curve

The time-to-value problem compounds this. Most people say "lack of time to learn" is their biggest barrier, yet the data shows AI users are saving 2 to 10+ hours per week. The disconnect? AI tools require understanding what they can do before you know what to ask them to do. That's backwards from normal software, where you learn by doing.

Enterprise adoption in fast-forward

The full blog explores what's really happening here: we're watching enterprise technology adoption in fast-forward. The cycle that used to take 5 to 10 years now takes 6 to 12 months. Companies built for the old pace can't adapt fast enough. Shadow AI is just the symptom. The real issue is that the software layer evolves every 6 months while the organizational layer changes every 3 years.

Want to get even more practical? Explore hands-on AI learning with AI Academy:

• AI Academy Membership: Get 12 months of access to all our cohort-based programs, live webinars, on-demand courses, and tutorials.

• AI Agent Bootcamp: Accelerate processes and solve business problems by mastering prompts and building AI Agents, without coding.

• Corporate Training: Equip your team with the skills they need to unlock the potential of AI in your business.

• Practical Introduction to ChatGPT: A free course on using ChatGPT confidently, understanding its workings, and exploring its potential.

We'll be back with more AI tips soon!

Hello AI Enthusiast,

Welcome back! Hope you had a great start to 2026. This week we're looking at three very different AI stories. OpenAI posted a high-stakes job opening, a former OpenAI leader shared some honest thoughts about keeping up with AI tools, and Instagram's head made some bold predictions about what feeds will look like this year. Two of these are really more about reflection than breaking news. Let's dive in.

The Big Picture 🔊

OpenAI Seeks "Head of Preparedness" Against AI Risks

OpenAI is hiring a "head of preparedness" at $555,000 annually to defend against AI risks including threats to mental health, cybersecurity, and biological weapons. Sam Altman warned it will be "a stressful job" where "you'll jump into the deep end pretty much immediately".

Helin YontarCPO and Polyglot

This role makes sense given OpenAI's lawsuits and real AI risks. The timing is interesting though. OpenAI keeps pushing the AGI narrative to support their valuation ahead of a potential IPO, while recent updates have been mostly UX improvements.

Preparing for AGI-level threats at $555K feels premature when current problems need more attention.

Karpathy Feels "Behind" as AI Reshapes Coding

Former OpenAI and Tesla AI leader Andrej Karpathy admits he's never felt "this much behind as a programmer." He says the profession is being "dramatically refactored" as AI tools like Cursor and Claude Code transform how developers work. Karpathy, who coined "vibe coding," believes he could be "10X more powerful" if he mastered these new tools properly.

Gianluca MauroFounder and AI Rockstar

Vibe coding is incredible for prototypes and simple projects. You can build functional apps quickly, even without programming knowledge. The trouble starts when projects grow complex.

AI struggles to track how different parts connect, suggesting fixes that break other systems. It's powerful, but needs some technical foundation to use effectively.

Instagram's Mosseri Says AI Content Will Overtake Real Photos

Instagram head Adam Mosseri expects AI-generated content to dominate the platform in 2026. He admits Meta can't reliably detect AI content and suggests camera makers should "fingerprint real media" instead. Mosseri says the era of "polished" Instagram photos is dead, claiming creators will need to post "raw" and "unflattering" images to prove they're real, not AI-generated.

Gioele MottarliniCOO and Image Addict

We can't tell what's real anymore. Not just fake news, but everything. You see a rescued kitten video and genuinely can't tell if it happened.

Mosseri's solution to flag real content instead of fake sounds logical, but the technology to verify authenticity at scale doesn't exist in any reliable form yet.

Bits and Bobs 🗞️

• Amazon launched Alexa.com, bringing its AI-powered Alexa+ assistant to the web and updating its mobile app with a chatbot-style interface, aiming to compete with ChatGPT and Gemini by focusing on family and home management features.

• OpenAI released GPT-5.2-Codex is its most advanced coding model yet, with stronger cybersecurity capabilities that helped discover React vulnerabilities.

• Meta has acquired AI startup Manus, known for its general-purpose AI agents, to enhance its AI talent and technology in line with its broader AI development strategy.

• Google sued SerpApi for using AI-driven methods to bypass its SearchGuard technology designed to block automated scraping of Google’s copyrighted search results.

• Google's new open-source A2UI protocol enables AI agents to generate and exchange interactive user interfaces across different platforms and applications.

From Our Founder’s Channels 🤳

As mentioned above, Gianluca shared his thoughts on LinkedIn about how 2025 was "the year of agents" on LinkedIn while most companies weren't actually ready to adopt them, and how that will change for some in 2026.

LOLgorithms 😂

How some feel when even Karpathy admits he's behind:

That's a wrap on our newsletter! Before you go, here’s a quick recap of our offerings:

• AI Academy Membership: Get 12 months of access to all our cohort-based programs, live webinars, on-demand courses, and tutorials.

• AI Agent Bootcamp: Accelerate processes and solve business problems by mastering prompts and building AI Agents, without coding.

• Practical Introduction to ChatGPT: A free course on using ChatGPT confidently, understanding its workings, and exploring its potential.

• Customized Corporate Training: Equip your team with the skills they need to unlock the potential of AI in your business.

Catch you next week! 👋

Reading Time: 3 minutes

Hello AI Enthusiast,

What if you could stay on top of every newsletter you subscribe to without spending hours reading them all?

Sara Cadena, a UX designer and tech coach from Latin America, joined our AI Agent Bootcamp to solve exactly this problem. The result? Newsletter Digest, an automation that reads her incoming newsletters, extracts the key information, and delivers a personalized summary straight to her inbox.

The summarized newsletter sent to Sara’s inbox

The Problem

You know the feeling. You've carefully curated a list of newsletters on topics you care about, from AI developments to industry trends, marketing insights to productivity tips. Then they all arrive at once, and suddenly your "reading time" becomes a backlog you'll get to "someday".

Sara described it perfectly: "I don't have enough time to read and learn from my curated list of newsletters". Every day meant checking her inbox, archiving less relevant emails, flagging the interesting ones, and blocking time to read that never quite materialized. The newsletters she subscribed to for learning became a source of frustration instead.

How Sara Built It

Before diving into building, Sara worked with our AI coaches to scope down her ambitious idea into something achievable. As she put it: "The bootcamp always reminds me of scoping down ideas, breaking down all these ambitious projects into a three-step process".

Here's the automation flow Sara created using Make:

📧 New newsletter arrives → triggers automation (OUTLOOK)

🤖 AI reads and summarizes the content (OPENAI)

✉️ Sends personalized digest to Sara (OUTLOOK)

The complete Make scenario

Step 1: Watch for Incoming Newsletters

Sara connected her Outlook account to Make to watch for new messages from her newsletter subscriptions. When a new email arrives from one of her curated sources, the automation kicks in automatically.

Step 2: AI-Powered Summarization

The heart of Sara's automation is the OpenAI module that processes each newsletter. The AI reads the full content and generates a synthesized version with a consistent structure that makes information easy to scan and retain.

The output includes a two-word title capturing the main theme, a one-sentence description, five keywords with definitions to build vocabulary, and important dates to watch.

Step 3: Deliver the Digest

The summarized newsletter gets sent directly back to Sara's inbox in a clean, consistent format. Instead of scrolling through walls of text, she gets the essential information distilled into something she can actually read and act on.

What Sara Learned Along the Way

Building Newsletter Digest taught Sara lessons that go beyond the technical:

Scope down ruthlessly. Sara's original vision was more ambitious, but working with coaches helped her focus on a working prototype first. "Think small, think small, think small", she recalls.

Always save your work. "I forgot to save, so I lost my automation and had to rebuild it." Fortunately, it was a three-step process. The lesson? Save early, save often.

Privacy matters. Connecting personal accounts to automation tools raised important questions. "How much access am I willing to give these tools?" It's a consideration worth thinking through before you build.

Ask for help. Sara booked 15-minute calls with coaches for every issue she encountered. "A lot of the solution can come up during that conversation which makes life easier".

Want to get even more practical? Explore hands-on AI learning with AI Academy:

• AI Academy Membership: Get 12 months of access to all our cohort-based programs, live webinars, on-demand courses, and tutorials.

• AI Agent Bootcamp: Accelerate processes and solve business problems by mastering prompts and building AI Agents, without coding.

• Corporate Training: Equip your team with the skills they need to unlock the potential of AI in your business.

• Practical Introduction to ChatGPT: A free course on using ChatGPT confidently, understanding its workings, and exploring its potential.

We'll be back with more AI tips soon!

Reading Time: 2 minutes

Hi AI enthusiast,

Today we published a new blog post on AI agents. Here's a summary of what it's about, but you can read the full post here.

Remember when spreadsheets became non-negotiable in the 90s? We're watching that moment happen again with AI agents.

Right now, according to McKinsey's latest State of AI research, 88% of organizations are using AI regularly, and 62% are already experimenting with agents. These aren't just tools anymore, they're becoming the standard way work gets done. OpenAI's State of Enterprise AI report shows enterprise workers saving 40 to 60 minutes daily, with heavy users reclaiming over 10 hours weekly. Even more striking, 75% report completing tasks they previously couldn't do at all.

Here's what matters: we're in a narrow window. Professionals who can build AI agents stand out today. But by 2026, this won't be a competitive advantage. It will be an expectation.

So what exactly is an AI agent?

At AI Academy, we define it simply: an AI system that can take actions using different tools with some degree of autonomy. Think of it as a digital colleague that accesses your systems, makes informed decisions, and acts on your behalf. It's the difference between owning a calculator and having a financial analyst.

Agents exist on a spectrum. Level 1 processes input without affecting workflow. Level 3 can choose and use various tools. Level 5 involves multiple specialized agents coordinating together. The key insight? You don't need Level 5 to create significant value. Even Level 3 agents can save hundreds of hours annually.

There's a crucial difference between using AI tools and building AI systems.

Those 10 hours per week that heavy AI users save aren't just from using ChatGPT better. They're from building agents that automate entire workflows. Anyone can use ChatGPT. Far fewer can design an agent that solves a specific business problem, integrates with existing workflows, and delivers reliable results. That's the skill that creates career differentiation.

Who needs this skill and what does it take?

Building AI agents isn't reserved for engineers. It's becoming essential for anyone whose professional value depends on thinking and making decisions. The core skills include problem decomposition, prompt engineering, understanding tool integration, designing decision flows with appropriate human oversight, and systematic testing.

In the full post, we walk through:

• The five levels of agent capability and what each can accomplish

• Specific skills required and how to develop them progressively

• Our 7-week AI Agent Bootcamp that takes you from concepts to deployment

• How AI Academy Membership supports your journey from foundational knowledge to mastery

The professionals who start building today will enter 2026 with practical experience, working systems, and sound judgment. Those who wait will be competing against them.

Want to get even more practical? Explore hands-on AI learning with AI Academy:

• AI Academy Membership: Get 12 months of access to all our cohort-based programs, live webinars, on-demand courses, and tutorials.

• AI Agent Bootcamp: Accelerate processes and solve business problems by mastering prompts and building AI Agents, without coding.

• Corporate Training: Equip your team with the skills they need to unlock the potential of AI in your business.

• Practical Introduction to ChatGPT: A free course on using ChatGPT confidently, understanding its workings, and exploring its potential.

We'll be back with more AI tips soon!

Hello AI Enthusiast,

OpenAI's making moves on multiple fronts this week. They've upgraded their image generation, while simultaneously locking down Disney's iconic characters for Sora. Meanwhile, the company's enterprise focus is getting impossible to ignore. GPT-5.2 is built entirely around work tasks, and they just hired Slack's CEO to lead their revenue push. And in other news, Google Translate got a Gemini upgrade that actually handles idioms properly now. Let's break down what's happening.

The Big Picture 🔊

OpenAI Releases GPT-5.2 and Hires Slack CEO for Enterprise Push

OpenAI launched GPT-5.2, claiming major improvements in professional tasks like spreadsheets, coding, and presentations. The model outperforms industry professionals on 71% of knowledge work tasks. Separately, OpenAI hired Slack CEO Denise Dresser as Chief Revenue Officer to lead its enterprise unit, signaling an aggressive push into workplace AI adoption.

Helin YontarCPO and Polyglot

OpenAI's enterprise push is becoming crystal clear. GPT-5.2 focuses entirely on work tasks like spreadsheets and presentations, while hiring Slack's CEO shows they're serious about workplace integration.

With only 5% of their 800 million users paying, enterprise is where the real money is. They need revenue, and they need it fast.

OpenAI Upgrades Images and Lands Disney Deal

OpenAI released GPT Image 1.5 with 4x faster generation and better editing precision, plus a new dedicated Images feature in ChatGPT. Separately, Disney signed a three-year deal giving OpenAI exclusive access to 200+ characters for video (and probably image) generation, but only for one year, after which Disney can partner with other AI companies.

Gioele MottarliniCOO and Image Addict

Disney's playing this smart. They're essentially saying "if you're going to use our stuff anyway, we might as well get paid and test how this AI thing works".

The one-year exclusivity window lets them see how it goes with OpenAI before shopping around.

Google Translate Gets Gemini-Powered Upgrade

Google Translate now uses Gemini to better handle idioms, slang, and local expressions in nearly 20 languages. A new beta feature delivers real-time speech-to-speech translations through headphones, preserving tone and cadence across 70+ languages. The app also added language learning tools with daily streak tracking, rolling out in nearly 20 countries.

Biljana PrlichkovaPRM and Outdoor Enthusiast

Google Translate's idiom handling is genuinely useful, anyone who's tried translating expressions knows the struggle. The real-time audio translation through headphones is clever, though Apple already made this a headline feature with their new AirPods Pro launch.

The language learning features with daily streaks feel like straight-up Duolingo copying.

Bits and Bobs 🗞️

• Google's new Interactions API enables developers to build complex AI agents with advanced context management and integrates Gemini models and custom agents via a single interface.

• Google Labs launched Disco, a browser that can generate simple apps based on the tabs you have open, letting your filter, combine and remix information across different websites.

• NVIDIA introduced Nemotron 3, three efficient AI models with improved accuracy and strong reasoning for agentic applications, including the cost-effective Nano model.

• Cursor Browser's new AI-powered visual editor lets users design and update web app interfaces by dragging elements and using natural language commands, which agents then translate into code changes instantly.

• AgenticAdvertising.org, a new nonprofit, will govern and promote open AI standards for agentic advertising to enable interoperable, intelligent agent collaboration across the marketing ecosystem.

• US President Trump signed an executive order creating a federal framework to regulate AI in the US, limiting conflicting state-level regulations to support industry growth.

• OpenAI reflects on a decade of AI breakthroughs, emphasizing their progress toward building safe and beneficial AGI integrated into society through iterative deployment and large-scale models like GPT-4.

On the Podcast 🎧

Last week’s episode features Sara, a UX designer and tech coach with no technical background, who joined our bootcamp to explore AI and surprised herself by building real AI tools, from a newsletter assistant to a virtual avatar and automated real estate agent - sharing lessons learned from mistakes and experimentation. Listen here

LOLgorithms 😂

Here we go again with "the world's most powerful model”

That's it for 2025! We're taking a couple of weeks off to enjoy the holidays and fill our bellies with way too much Christmas food. If anything major happens in AI while we're away, we'll make sure to cover it in our first newsletter of 2026.

Merry Christmas and Happy New Year everyone!

That's a wrap on our newsletter! Before you go, here’s a quick recap of our offerings:

• AI Academy Membership: Get 12 months of access to all our cohort-based programs, live webinars, on-demand courses, and tutorials.

• AI Agent Bootcamp: Accelerate processes and solve business problems by mastering prompts and building AI Agents, without coding.

• Practical Introduction to ChatGPT: A free course on using ChatGPT confidently, understanding its workings, and exploring its potential.

• Customized Corporate Training: Equip your team with the skills they need to unlock the potential of AI in your business.

Catch you next year! 👋

Hey AI enthusiast,

We're trying something new! Some topics we cover deserve more depth than a standard newsletter allows. So we're starting to publish longer blog posts on our website for these deeper dives.

Here's how it works: when we publish one of these posts, we'll send you a summary like the one below, enough to give you the key insight, but the full framework and details live on the site. The rest of our newsletter stays exactly the same. Just occasionally, you'll see these longer-form pieces.

Today we published a new article on AI strategy. Here's what it's about, but you can read the full post here.

Why Panic-Driven AI Adoption Rarely Works (And What Does)

The pressure to adopt AI is everywhere.

"Deploy now or get left behind." 

"Your competitors are miles ahead."

Here's what we keep seeing: organizations rush to buy AI tools, then those tools sit unused. Budgets get wasted on solutions that don't fit actual problems. Teams end up more confused than when they started.

The pattern is clear. Panic-driven implementation rarely creates lasting value.

What's Different About Organizations That Succeed

They start with their actual problems, not the latest tools. Before choosing anything, they ask: Which processes are inefficient? Where do teams get stuck? What decisions would improve with better information?

They also get realistic about AI's limitations. This keeps expectations grounded and prevents disappointment down the road.

Most importantly, they test small before going big. Small pilots with clear metrics. Gather feedback, measure impact, learn what works. Then scale.

The Framework

In the full article, we walk through a practical four-step framework for intentional AI adoption. It covers how to define objectives that actually matter, evaluate tools beyond the marketing hype, structure pilots that generate real learning, and build systems for continuous improvement.

We also dig into the common pitfalls that trip up even smart teams and why competitive advantage comes from adopting AI well rather than adopting it first.

The approach might feel slower than jumping on every new tool, but it builds capabilities that actually stick and compound over time.

Want to get even more practical? Explore hands-on AI learning with AI Academy:

• AI Academy Membership: Get 12 months of access to all our cohort-based programs, live webinars, on-demand courses, and tutorials.

• AI Agent Bootcamp: Accelerate processes and solve business problems by mastering prompts and building AI Agents, without coding.

• Corporate Training: Equip your team with the skills they need to unlock the potential of AI in your business.

• Practical Introduction to ChatGPT: A free course on using ChatGPT confidently, understanding its workings, and exploring its potential.

We'll be back with more AI tips soon!

Hello AI Enthusiast,

OpenAI dropped an enterprise report, packed with growth numbers and some interesting gaps in the story. They're also experimenting with a "confession" system that teaches AI models to admit when they break rules or take shortcuts. Meanwhile, Google launched Workspace Studio to let anyone build AI agents without coding. Let's break down what these developments actually mean.

The Big Picture 🔊

OpenAI's Enterprise Report Shows 8× Usage Growth

OpenAI's new enterprise AI report reveals ChatGPT now serves 800 million weekly users, with enterprise usage growing 8× over the past year. Workers report saving 40-60 minutes daily, with 75% seeing improved work quality. Structured features like Projects grew 19× year-to-date, showing AI adoption moving beyond simple queries.

Gioele MottarliniCOO and Image Addict

Interesting report, though it leans toward marketing rather than research. OpenAI highlights 8× enterprise growth and 30% more messages per worker, but more messages doesn't mean better results, sometimes users keep asking because they're struggling.

The useful insight is the usage gap: top users employ AI 6× more than median workers, while many ignore core features. Real value comes from proper training, not just buying licenses.

OpenAI Introduces "Confessions" for AI Transparency

OpenAI developed a "confessions" technique where AI models produce a separate, honest report admitting when they violated instructions or hallucinated, without being penalized. Tested on GPT-5 Thinking, it caught 95.6% of problematic behaviors. The confession trains solely on honesty while the main answer optimizes for helpfulness. This proof-of-concept aims to surface hidden misbehaviors as models become more capable.

Helin YontarCPO and Polyglot

This is addressing a real problem, companies need to trust that AI follows their policies and safety guidelines. The confession system works well for catching when models try to bypass rules or take shortcuts.

The challenge is when a model genuinely believes it's right but isn't, it can't confess to being wrong. Still, it's a step forward for transparency and catching intentional rule-breaking.

Google Launches Workspace Studio for No-Code AI Agents

Google released Workspace Studio, promising anyone can build AI agents without coding. Using Gemini 3, employees can supposedly automate tasks from simple email sorting to complex workflows by describing what they want in natural language. Agents integrate with Gmail, Drive, Chat, and external tools like Asana and Salesforce. Rolling out to business customers over the next few weeks.

Andrea MattielloCM and Board Lover

About time! The tool looks promising for automating routine tasks and making workflows smoother. The interface is clean and similar to Zapier.

The main limitation is that it's still not flexible enough compared to established automation platforms in the market, and it's heavily focused on Google's own apps.

Bits and Bobs 🗞️

• The Linux Foundation launched the Agentic AI Foundation to support open source AI agent projects and prevent incompatible, closed systems.

• Mistral launched Devstral 2, a powerful open-source AI coding model with 123 billion parameters, designed for business use and available via API.

• Anthropic launched Anthropic Interviewer, an AI-powered tool using Claude, to conduct large-scale automated interviews that reveal professionals’ nuanced views on AI’s impact on their work and productivity.

• A judge ordered OpenAI to provide millions of anonymized ChatGPT user logs in a copyright dispute, ruling privacy measures are sufficient.

• The EU is investigating whether Google improperly uses publisher and YouTube content to train its AI models without fair compensation or consent.

• Slack users can now tag Anthropic's Claude to automatically handle coding tasks using context from conversations and connected code repositories.

• McDonald's released an AI-generated Christmas ad that sparked debate over AI's role in creativity and concerns about authenticity and job impact.

From Our Founder’s Channels 🤳

Gianluca’s latest LinkedIn post breaks down OpenAI’s enterprise report, highlighting that companies see real ROI only when teams are well-trained, build custom automations, and embed AI deeply into workflows.

LOLgorithms 😂

Let's hope ‘Confessions’ will help us see fewer em dashes around.

That's a wrap on our newsletter! Before you go, here’s a quick recap of our offerings:

• AI Academy Membership: Get 12 months of access to all our cohort-based programs, live webinars, on-demand courses, and tutorials.

• AI Agent Bootcamp: Accelerate processes and solve business problems by mastering prompts and building AI Agents, without coding.

• Practical Introduction to ChatGPT: A free course on using ChatGPT confidently, understanding its workings, and exploring its potential.

• Customized Corporate Training: Equip your team with the skills they need to unlock the potential of AI in your business.

Catch you next week! 👋