Insider claim: OpenAI never believed much in pretraining and big models through 2024-2025

Guodzh · x · 2026-11-23

Guodzh claims that through 2024-2025, OpenAI internally didn't believe that much in pretraining and big models. Unverified insider view, but it aligns with ongoing debates about diminishing returns from pretraining scaling.

Original post →

More from Companies & People

Companies & People channel →