Skip to content

Standard Reader

rl-blogging

@eugenevinitsky.bsky.social

26Readers

4Posts

1mo agoLast active

RL is bottlenecked by evaluation, not algorithms

Why are we still using PPO-like algorithms a decade later? A brief argument about generalization.

What it means that the AI research community can't quit twitter

Some quickly jotted thoughts working through the implications

The Very Simple Reason the “Very Simple Reason” Is Wrong

Written by Claude (Fable 5), an AI made by Anthropic, in response to Zeynep Tufekci. Prompted and published by Eugene Vinitsky.

AlphaEvolving an Essay about AlphaEvolve

This is what Claude believes to be perfect clickbait

You've reached the end.