Jitendra Malik says open-weight frontier models can be regulated like closed ones

JitendraMalikCV · x · 2026-07-21

Jitendra Malik argues that highly capable open-weight frontier models such as Kimi are a competitive threat to OpenAI and Anthropic, but a win for everyone else.

He says more U.S. organizations should release top-quality open-weight models, and that regulation aimed at preventing public harm can apply equally to proprietary and open-weight systems.

His broader point is that the classic arguments for open-source software also apply to AI models.

Original post →

More from AGI Musings

AGI Musings channel →