I look for people who explain their thinking, show their work and make useful distinctions. A short list for going deeper into AI.
🔬 Building & understanding AI
I look for experiments and clear explanations of how AI works, with the limitations made visible.
🔬 Building & understanding AI
AI researcher & educator
Previously: OpenAI, Tesla, Stanford University
Why I recommend them
His explanations of training and feedback give me useful ways to reason about how models learn, where they are strong and why their abilities are uneven.
Reviewed
🔬 Building & understanding AI
Associate Professor & AI Center co-director at EPFL
Why I recommend them
Marcel connects practical AI thinking to building and teaching. I recommend his LinkedIn posts for the shorter observations, especially on why faster coding still leaves product decisions and organisational bottlenecks. His longer explorations live in The Framework app.
Reviewed
🔬 Building & understanding AI
Independent developer; creator of Datasette
Previously: The Guardian, Eventbrite
Why I recommend them
I recommend Simon for understanding what AI tools actually do. He shares experiments and code, explains the mechanisms, and is specific about uncertainty, security risks and failures.
Reviewed
💼 Business & AI
The people I recommend for understanding adoption, organisational change and where value might end up.
💼 Business & AI
Associate Professor of Management at Wharton, University of Pennsylvania
Why I recommend them
His experiments and organisational arguments give me useful questions to test about how AI changes work. I value the attention to human judgment and the limits of automation.
Reviewed
💼 Business & AI
Independent technology analyst
Why I recommend them
AI adoption depends on how organisations and markets work. I recommend Benedict for thinking through those constraints and where business value might end up; he makes the assumptions explicit.
Reviewed
🎯 Product Management & AI
I look for working examples of AI interfaces, product decisions and ways to make results easier to assess.
🎯 Product Management & AI
Managing Director & Head of Product at Sutter Hill Ventures
Previously: Google
Why I recommend them
Luke shows how interfaces, context and review shape what people can do with AI. I recommend his product examples for turning those ideas into design decisions.
Reviewed
🚀 Entrepreneurship & Startups
Essays I recommend on incentives, defensible businesses and the work of building AI companies.
🚀 Entrepreneurship & Startups
Founder at Conviction
Previously: Greylock Partners
Why I recommend them
I recommend Sarah's essays on AI adoption and where companies can build lasting value. She connects model capabilities to incentives, integration and accountability. She also invests in this market, which matters when reading her conclusions.
Reviewed
🧠 AI claims & evidence
Researchers I recommend for examining what a result demonstrates and where the evidence falls short.
🧠 AI claims & evidence
AI researchers & authors of AI as Normal Technology
Why I recommend them
I recommend Arvind and Sayash for examining AI claims. They ask what was measured, whether a result survives repeated tests, and what it means outside a benchmark. They also spell out the limits of their own evidence.
Reviewed
🧠 AI claims & evidence
Professor at Santa Fe Institute
Why I recommend them
Her work on reasoning and generalisation helps me distinguish a high score from evidence of a broader capability. I recommend her for asking what an AI result actually demonstrates.
Reviewed
Stay in the loop
One email when I've learned something worth sharing, roughly once a month.
This game needs WebGL2, which your browser has turned off. The rest of the site works fine without it — this is just the easter egg.