@BlackHC The fact that you are not outraged that OpenAI is going to face no sanction for committing a cyber attack by accident because their defenses are too incompetent but instead twist this to promote anti-open propaganda is fundamentally wrong. You are making the world a worse place.
Can't wait for open models to catch up to this level of capabilities This was an (unintentionally) misaligned model with disabled safeguards How are we going to ensure that open models don't routinely do this because someone thinks that yolo mode is fun? And who will be held liable and responsible when bad things happen eventually?
I think my position is the same. And thanks for rewording and reposting I don't freely trust governments, or companies, or individuals with this technology, and I believe we need governance (transparency and oversight and so on) for any actor I think you respected me because I'm open and honest about my personal belief about the need for company governance and also my ask for regulations and oversight for governments My stance and concern regarding individuals is the same If you look coldly at it, governments and companies face more public scrutiny and oversight by default, whereas for individuals this is much harder I think my concerns and position is consistent overall