OpenAI built their empire on open source then gatekeeps access to closed models, and then has managed to dodge accountability for hacks their agents have done due to sandboxing negligence by saying "The AI did it".

My opinion that OpenAI should be shut down and leadership held accountable has not changed.

That said, fixing bugs in internet-critical software like Linux is actually important work, and makes me hate them like 10% less.

If their excuse for not opening the models is they want to patch the world first, and they follow through, I can get behind humoring that narrative.