[pseudalopex]

Altman aims to be trusted to say what regulations should and should not be made. It should not surprise you people consider evidence of dishonesty and suspicious coincidences relevant to trust.

People have lost jobs and likely careers to AI models trained on their works. You could assert in the long run all individuals will be better off. You could assert the benefits to others made the harms virtuous. You could assert they deserved it. I don't know how you could deny they were harmed. You could assert it was inevitable. But this would negate credit if it would negate blame. This is a distraction from the question of trust however.

[eigenvalue]

Unfortunately, I think this was totally inevitable, particularly now that there are powerful open models that can be fine-tuned Llama3. At this point, you can't stop it any more than you can stop piracy of books and movies. And I'm not even so sure that "access to their copyrighted works" was the primary reason for anyone being disrupted from AI.