Open Weights Won't Necessarily Slow Down AGI

natolambert · x · 2026-07-19

The author replies that if open weights ultimately dominate, it will squeeze lab profit margins and subsequently reduce available R&D funding, potentially delaying AGI. He believes this concern has merit, but the US is already far ahead in AI investment. Thus, even if open models continue to advance, an "AI crash" isn't guaranteed. If labs do lose their edge, it won't be because AI lacks value; rather, it will likely be because **compute** remains a bottleneck, ultimately driving more value toward compute providers.

Related event: Professor Challenges Notion That Open-Weight Dominance Will Collapse AI(2 posts)→

Original post →

More from AGI Musings

AGI Musings channel →