Llama 3 is out, but it seems that still only @MistralAI really has our back: it still has downstream usage restrictions of model output. @AIatMeta this is a garbage restriction that harms open source and I was really hoping you would see the light this time.
@xlr8harder @MistralAI @AIatMeta can't believe i woke up early for this
@xlr8harder @MistralAI @AIatMeta How’s it different from making an uncensored model? @erhartford
@xlr8harder @MistralAI @AIatMeta @elonmusk does this fit your definition of open source?
@xlr8harder @MistralAI @AIatMeta You would be so surprised who probably makes them put that language in there. Here is a hint. It’s not Zuck or branding.
@xlr8harder @MistralAI @AIatMeta This is similar to the liscence restriction they had for react.
@xlr8harder @MistralAI @AIatMeta Just do it anyway and no one will ever know
@xlr8harder @MistralAI @AIatMeta "derivative works" is pretty broad. Could argue it's any Llama-like
@xlr8harder @MistralAI @AIatMeta just saw your post, i posted the same thing a few moments ago.
@xlr8harder @MistralAI @AIatMeta The only reason this is here is that if they found out Claude 4 or GPT-5 was trained on Llama 3 they can sue. It’s not a big deal for everyday people.