You can’t check out any time you like.
1. AI Is Nowhere Near Its Peak, and It’s Already This Good
AMD’s CEO Lisa Su was asked about the AI bubble on Bloomberg’s Wall Street Week. Su runs one of the companies that makes the chips AI runs on, and her answer was this:
“What I’m here to tell you is that it’s nowhere near its peak capability.”
Yes, she has a financial incentive to tell you that AI has future growth. But it doesn’t mean that she’s wrong.
Now put that next to what the tools can already do. Just in the last few weeks I’ve had an AI refactoring an app in the background while I co-wrote newsletters with AI in another window. I can run several apps and client projects at once, each with an agent working inside its own repository. Working this way, which I call hypertasking, I produce three to five times the output that I did before. And this is not anything unusual. Students of my courses and other people using AI and AI agents are achieving similar outputs. And that is with tools that, by Su’s account, are an early version of what is coming.
What Su is largely talking about is capability. I think the more interesting point is people.
Every large human achievement came from many people cooperating and building on each other’s work. AI lowers the cost of building everything, which means that 8 billion people on the planet can potentially now build.
Yet most of them haven’t started. Microsoft’s latest Global AI Diffusion Report found that only 18.8% of the world’s working-age population used a generative AI product even once in June 2026, so four in five people haven’t touched it. With most of humanity still on the outside, you can see why I say that AI is nowhere near its potential.
So work out which rung you are on: never used it, a free chatbot now and then, paying for it and using it weekly, using it daily, or agents that work inside your files. Then take one step up this month.
2. The Walled Gardens Are Starting in AI
For the past two years we have been treating AI models as interchangeable parts. You chose a model and used it in whatever tool you liked. And as the model race continued between Anthropic and OpenAI, most users went back and forth between their models to get work done. Now that is starting to change, and we are seeing the start of walled gardens in AI.
Anthropic launched Claude Code Mods on 1 October, which are plugins that can rewrite what the agent does and draw their own panes, and they only work in Claude Code with Claude models. OpenAI introduced Intelligent UI on 7 October, where ChatGPT answers with interactive buttons, forms, charts and diagrams inside ChatGPT, and its Codex and ChatGPT apps are merging with functionality designed for its own apps.
Both vendors want to encourage you to use THEIR app, not the CLI, and not their model in a different app or AI harness.
Mods, skills, interactive answers and saved work all live in one vendor’s app, and the more you build there, the more it costs you to leave, even if a rival model becomes better or cheaper.
This leads to some questions.
Will subsidised subscription plans disappear for direct model access? I’m not sure.
Will the labs go API-only for direct model access? Maybe.
And what happens post IPO? A public company reports its margins every quarter, and plans that lose money on heavy users are harder to defend to shareholders.
What does this mean for us?
For now, subsidised subscriptions continue. So use each garden’s features, but try to keep the parts you own portable. I keep the same instructions-file structure in every repository and I store my skills in a neutral form (AGENTS.md and .agents/ with CLAUDE.md and .claude/ symlinked), so that I can plan with one model and build with another.
3. Courses
I have a range of online courses that teach business people what they need to know about productivity and AI:
1. The Complete Claude, Claude Code & AI for Work Productivity
A comprehensive course taking you from beginner AI concepts like prompt and context engineering, to cutting-edge AI productivity using terminal-based AI tools like Claude Code for non-coding office work. Usable with ChatGPT, Claude, Gemini and other current LLMs.
2. The AI Playbook
This is my longer premium course on how businesses can deploy AI tools and technology across their processes and teams.
3. The AI Agents, Automations and Agentic Workflows Guide
This non-technical course shows business people and non-coders how they can build and use AI agents in ChatGPT and Zapier.
4. Next Level Productivity
A practical, straightforward course that teaches you how to achieve elite-level personal productivity in today’s constantly interrupted world.
5. ChatGPT for Managers
See how AI can solve complex management challenges in less than 30 seconds. Full prompt library and examples included.
That’s it for this week!
— Aaron