Nice paper that is basically an empirical survey of all the recent RL algorithm improvements.

my matrix multiplication kernels for magnetron now beat PyTorch's performance on my CPU: Magnetron matmul: 1955.2 GFLOP/s Torch matmul: 1752.3 GFLOP/s check out the kernel code: github.com/MarioSieg/magn… magnetron detects cache sizes and uses state of the art block tuning

i usually power limit my gpus down by 30% during inference the loss in performance is negligible, and a huge temperature improvements benchmark everything btw, if you don't measure it you cannot make it better

[episode 123 of frontier lab gossip: OH on X dms] > yo did you hear about the MSL packages and lmsys scores? > apparently folks have $XXX mil in comp tied to however many weeks they can get their models to stay at #1 on lmsys > they seem to still think that's a flex > when its

AIが人間社会を統治することも想定して、AIエージェントによる民主主義政治のシミュレーションを行ったところ、AIだからといって清い政治を行うわけではなく、制約と仲裁が必要だと示唆されています。 研究者たちは17体のAIエージェントに、人間のような複雑な心理プロファイルを与えました。