Yacine MTB: Everything is an RL env — publish one for your task and let models overfit it

yacineMTB · x · 2026-09-03

Yacine MTB argues that everything is an RL environment: if you want a specific task done well by frontier models, publish an RL env for it so models overfit to it. Benchmarks, in this view, effectively become training targets once environments exist for them.

Related event: Yacine: publish an RL environment to make models excel at your task(2 posts)→

Original post →

More from Models

Models channel →