Post #1904952
2026-04-30 06:38 UTC
https://archive.ph/hHTtf
> Anthropic warned that illicitly distilled models lacked necessary safeguards, creating significant national security risks. If these models are open-sourced, the risk multiplies as capabilities spread freely beyond any single government's control
"Illicit destilled models", "if these models are open-sourced" - I stg these US LLM labs just sound like supervillians at this point
Replies (1)
-
@pojntfx@mastodon.social 2026-04-30 06:40
> Anthropic, which raised $30 billion in its latest funding round and is now valued at $380 billion, said that distillation attacks support the case for export controls: chip access restrictions reduce both direct model training capabilities and the extent of improper distillation. It's literally just "wdym you can just the models locally? How can we possibly seek rent on that?"