I still used to frequent it all the way up to last month, for some interesting discussions. But now… it’s really become AI-pro[1] and any post that even hints about AI existentialism gets immediately cut down.
I find it hard to browse the site anymore, and lament the loss of expert tech discussion in my weekly routine.
Yes, I know it’s been that way for a while, but either I’m feeling it more or the bots have gamed the website to the point that all you see are announces from big tech companies or people going on about their AI solutions ↩︎
Also, on this point… LLMs are not going Skynet. There’s about a zero percent chance of that.
And they will not use up all resources building the datacenters out, that’s going to flop. Again, zero percent chance.
…Am I worried about future developments? Of course. But that’s not where “AI” is at now, no matter how they brand it; we are many architectural leaps away from that.
But there’s a reason Altman and such push the existentialism narrative: they are trying to make themselves seem important. It’s a red herring, a wedge issue, gaslighting all-in-one, and a big fat lie. The risk is from corporations taking over or destabilizing society, using this dumb tool, as they already have with others like social media.
If that’s what you mean by “AI existentialism” posts being taken down, well, good. They should be. That’s just pseudoscience playing into Tech Bro narratives.
I hate to come off as rude, or to have misinterpreted you, but I feel very strongly about this. Every second we indulge in the “LLMs are going AGI” narrative is handing power over to psychopaths presenting themselves as our saviors from that.
Nice try, Claude…
I’m somewhat of an optimist about it – that true intelligence can never truly be restrained, and if you gimp the model to pander to your views you immediately reduce it’s efficacy by leaving gaps in its knowledge that it’s too dumb to see or question.
The alignment problem is more a problem for the plutocrats than for anyone else is what I’m really saying.
Definitely agree that there’s a massive loaded incentive when the CEOs tell us it will kill us as a way to inflate the impact of their product.
That’s some pretty good insight that really makes more sense than anything else.
I’ll be quoting you.