There’s been a lot of discussion since DEF CON and Black Hat about how good or bad a job OpenAI did in preventing and handling the incident with Hugging Face. I’m a little surprised—I probably shouldn’t be—at how confident people in our industry are that they could do better or even know the full details of the environment to have known if it were even possible without knowledgeable people doing significantly better.

I honestly think there’s a little conflict of interest in some of the criticism. I do think everyone will learn a lot, but there’s also a bunch of “influencers” making a name for themselves by being very vocally critical. It’s cool right now to hate on the frontier labs.

They don’t deserve a pass, but I also think there’s room for a more helpful dialogue. I’ve seen no truly good write-ups from this crowd of how they would have done better, and that’s telling.