Neural chameleons can(‘t) hide from activation oracles
Published on January 22, 2026 1:47 AM GMT[epistemic status - vibe coded, but first-pass sanity-checked…
Category: Opinion
When should we train against a scheming monitor?
Published on January 21, 2026 8:48 PM GMTAs we develop new techniques for detecting deceptive…
Claude Codes #3
Published on January 21, 2026 7:50 PM GMTWe’re back with all the Claude that’s fit…
Claude’s new constitution
Published on January 21, 2026 7:37 PM GMTRead the constitution. Previously: 'soul document' discussion here.…
Crimes of the Future, Solutions of the Past
Published on January 21, 2026 7:20 PM GMTThree hundred million years ago, plants evolved lignin—a…
On visions of a “good future” for humanity in a world with artificial superintelligence
Published on January 21, 2026 6:27 PM GMTLet us imagine a world with artificial superintelligence,…
The case for AGI safety products
Published on January 21, 2026 5:23 PM GMTThis is a personal post and does not…
Vibing with Claude, January 2026 Edition
Published on January 21, 2026 4:00 PM GMTNB: Last week I teased a follow-up that…
Kredit Grant
Published on January 21, 2026 12:56 AM GMTI have $5,710 in OpenAI/Anthropic credits which I…
Money Can’t Buy the Smile on a Child’s Face As They Look at A Beautiful Sunset… but it also can’t buy a malaria free world: my current understanding of how Effective Altruism has failed
Published on January 20, 2026 11:28 PM GMTI've read a lot of Ben Hoffman's work…