An opinionated guide to which AI to use to do stuff
Simon Willison · Simon Willison · 2026-07-27
Simon Willison comments on the evolution of Ethan Mollick's AI usage guide, noting its shift from chat interfaces to agentic systems over the past year, Gemini's disappearance from the list, and the confusing naming conventions for computer-use modes across ChatGPT and Claude.
Appears in
Extraction
Topics: agentic-aillm-usabilitychatgptclaudeai-products
Claims
- Ethan Mollick's AI guide has shifted in one year from recommending chat interfaces to emphasizing agentic systems capable of performing multi-hour autonomous work.
- Gemini has dropped from Mollick's recommended list because Google has not established a competitive entry in the agentic coding and computer-use category.
- Both ChatGPT and Claude offer computer-use modes whose naming conventions do not map onto each other and differ between mobile and desktop contexts in non-intuitive ways.
- Enabling ChatGPT's 'Work' mode on mobile removes the internet restriction from its Code Interpreter container, a behavior Willison describes as 'spectacularly unintuitive.'
Key quotes
Today it's much more about agentic systems - 'where the AI is capable of doing the equivalent of many hours of real human work in one go'.
I think the difference between ChatGPT Work on a mobile device and ChatGPT Work inside the desktop app (where it's effectively a less intimidating skin on top of Codex) is spectacularly unintuitive.