Hugging Face CEO Backs Layered Rules for AI Weights
Hugging Face CEO Clément Delangue posted that model weights, APIs, and apps need separate policies as distinct stack layers.
Clément Delangue posted that APIs from companies like Anthropic and OpenAI are treated differently from open weights because the three layers require distinct rules. He compared weights to raw steel. Replies from Gary Marcus, Helen Toner, Andreas Kirsch, and others noted agreement on regulating at the layer where risks appear, such as finished apps, and referenced analogies to car manufacturing and commodity standards. Delangue replied that the approach makes sense and thanked participants. The conversation centers on his clarification of the original post.
Some people are surprised that APIs (aka what Anthropic, OpenAI, and others provide) are treated differently than open weights in the new AI model framework. I'm not surprised at all, and it's actually very good policy. Let me explain: Model weights, APIs, and apps are three…
Reactions
13 postsMK HT 
Helen Toner@hlntnrShoutout to @rajiinio, the OG mythbuster on this
- likes: 31
- replies: 1
- bookmarks: 3
- reposts: 0
HT Helen Toner@hlntnr@rajiinio And appreciate this follow-up from @ClementDelangue (I responded in thread)
- likes: 10
- replies: 0
- bookmarks: 0
- reposts: 0
DW Dean W. Ball@deanwball“We don’t regulate steel” is one of those beliefs that reminds me of the oft-quoted observation about naive libertarianism: “like house cats, completely dependent on others but convinced of their fierce independence.” Steel and other building materials are regulated by global…
- likes: 511
- replies: 24
- bookmarks: 100
- reposts: 49
AK GM Gary Marcus@GaryMarcus@ClementDelangue good clarification and yes, “different” makes sense
- likes: 3
- replies: 1
- bookmarks: 0
- reposts: 0
