How persistent is the inference cost burden? (opens in new tab)

Covered by XDADiscussed on Substack

RL scaling might be better than it looks, and inference costs to reach a capability level fall fast

Sign in to keep reading the full article.

Covered in 1 article