Skip to content

Comment on The model is the productparent

Comments

That doesn’t make sense though if scaling is actually stalling. The reason so much compute goes into training now is scaling, which keeps base model lifetime short.

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.