causal

RL's Sparse Updates Protect Existing Knowledge

RL concentrates gradient updates only on what is relevant to getting the outcome right, making its updates sparse, which matters for continual learning because on-the-job learning must not overwrite everything else the base model knows.

causalpending

Speaker

Dwarkesh Patel

Evidence Quote

“you don't want to overwrite and forget all the other things the base model knows.”

Source

Dwarkesh Patel Podcast - The next big breakthrough will be AIs learning on the job— Dwarkesh Patel Podcast
Created: 11/17/2026, 6:25:50 AM

Your notes

Loading notes...