Research Are AI Proofs Messier Than Human Proofs? October 9, 2026 · 9 minute read On Tuesday, OpenAI released 719 math manuscripts written by an internal model. Some of them solve problems that had been open for decades. Yesterday morning, Tanya and I talked about what this means for the elegance o...
Do Corporate Incentives Teach AI to Protect Itself? An Athlete's Mind The Designer, The Craftsman, & The Programmer The Second And Third Falling Vibe Coding: Taste Is The Last Frontier "Is The Light Merely For The Learned?" (Er Lyset For De Lærde Blot?) English Translation "Surrounded by Enemies" (Kringsatt av Fiender), or "To The Youth" (Til Ungdommen) English Translation To An 18 Year Old Me Mediocre Intelligence, Outstanding Results How to host a hackathon Raising for the Endgame: An AI Safety Founder’s Primer Learning From Architecture AGI Security: Rethinking Society's Infrastructure AGI Privacy Don't Give Up The AI Bubble The concrete risks of AI misdeployment Sentware Confident optimism The Expensive Internet hypothesis The AGI endgame "Just do the thing" Cybermorphism Go on a yearly hiatus Let's make AI agents safe My Tools Audiovisual reading

Research

Socials

News