Post

motherfucker you just hacked the australian government through your organization's incompetence, you do NOT get to tell us how to accept bad things when it is VERY clear that you and OpenAI could have stopped an awful lot of them!

Like, LLMs are all over now and probably can't be meaningfully gotten rid of, but that doesn't mean we should just accept bad things and not try to mitigate them. Additionally, the way AI technology was released to the public IS ALSO SAM ALTMAN'S FAULT!

It's cool I guess that regardless of what we may think of LLMs and the people who sell them, there's a pretty clear consensus on what to think of Mr Altman.

I mean, I don't actually like anybody selling LLMs. I've refused to pay any of the labs and only begrudingly pay an open-weight inference provider because they do unfortunately have utility and not having experience with them impacts my job prospects. That's less of an issue for you.

I try not use them wastefully, I try to make sure that the code I put out into the world isn't just slop, I try to make sure I'm doing more good than bad. I hope that we can better mitigate the harms someday soon since we can't erase them. I don't know if that's enough. I'm sure you think it isn't.

This really wasn't meant to be a shot so much as a dunk on Altman, haha. More dunking, I say! He is bad and should feel bad.

Also, the US drove a kid to suicide for downloading some journal articles -- it feels like we could at least muster some basic criminal liability for hacking a foreign government!