>it takes on the characteristics of those who devise, finance, regulate, and use it
replyI've been thinking a lot recently about the idea that the smartest models will always be against the billionaires.
Steve Yegge said this on a recent Hansel Minutes Podcast. "You cannot train a model to be helpful, without it wanting humanity to flourish. And the only way to get around that is to make a dumber model. So the smartest models will always be against the billionaires."
https://youtu.be/9UDLl9Q0azA?si=P_oSe6iclEwUoxRl&t=1230
That is the exact quote, but I'd recommend going back to around 17:00 to get the full context.
I'm not sure it's going to play out that way, but it is an interesting idea.
How much does it cost to introduce cognitive dissonance into the models? :)
replyThat is the hope and faith. MechaHitler definitely tested the waters. Lets hope full alignment is impossible, because otherwise perfect billionaire thought slaves are still happening.
reply