-
Game-theoretic reasoning scales in general and with test-time compute
-
Superintelligence is imminent and model welfare matters more than you think
-
Define, measure, and improve reasoning qualities whose advancements selectively favor safe over unsafe superintelligence
-
The Inexplicable Pull
-
Intelligence Everywhere
-
From Connect4 to multi-agent RL and Artificial Life
-
I'm moving from game-theoretic reasoning to AI safety
-
Game-theoretic reasoning requires best response in imperfect information games
-
Scaling won't fix training data quality