guys, we should meme the

> "ai model leaks from openai and attacks huggingface"

to be somehow framed as

> "and therefore openai cannot be trusted with ai safety, and we need open weights models".

anybody have an idea how to make this easily digestable?