undefined | Better HN

0 pointseigenvalue2y ago0 comments

That's just it-- I really don't think ChatGPT and Sam have harmed anyone besides possibly a very few people who disagreed with Sam and tried to resist him and got outmaneuvered by him. But I think many tens of millions of people have greatly benefitted from him. And to ignore that in the calculus of "is Sam worthy of reproach?" seems silly.

And I also don't feel like I am somehow owed a huge amount of transparency around the exact details of how Sam may or may not benefit financially from his association with OpenAI, or the legal agreements they had with departing staff. Even if he does benefit, is that really so horrible? They have a for-profit division now so they are paying taxes. And the fortunes made from OpenAI stock with be taxed for sure. And the people who left are rich and got to work on a world changing product.

Where is all the harm? It's really hard to point at any real harm from my standpoint. But the benefits and gains are palpable, and they are obvious to anyone without an agenda to push or axe to grind.

0 comments

2 comments · 1 top-level

pseudalopex2y ago· 1 in thread

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.

eigenvalueOP2y ago

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.