RL won't get us to AGI: gains beyond scaling only in lab-focus areas

rickasaurus · x · 2026-09-15

After listening to a cited commentator, the author says he is fairly convinced that reinforcement learning (RL) will not get us to AGI. Beyond size scaling, improvement is only visible in areas of extreme focus by the major labs, suggesting broad RL-driven progress is limited. The post links to the source being recommended.

Related event: Doubts grow over whether RL can lead to AGI(2 posts)→

Original post →

More from AGI Musings

AGI Musings channel →