@nickcammarata all this talk about pausing research, but isn't it a prisoner's dilemma? So it'll be like climate change, how do you get everyone to pause it?, it's a bit different than enriched uranium
@GainMaxxing@doodlestein robotics but its intwined with the current ai bubble, so perhaps after a correction. although i think the internet and algorithms has made alpha so widespread that boom and bust cycles i feel will go faster than before, an eternal bubble if you will, kinda scary
@chrislakin@mattgoldenberg isnt this survivorship bias? what if the consequence of going too far below the playing it safe is irrecoverable damage or injury, like a uncurable disease, loss of loved one, or death? recoverable but also bad outcomes include unpayable debt, mental illness, estrangement, etc
@zeroxkyle you do realize that the latest models have been trapped only making optimization gains on hardware thats already 2 years old. B300. Theres a roughly 2 year lag between labs designing and releasing a scaled up and optimized model that can push the latest generation of Nvda gpus
@rennyzucker they want to maintain their status and privileges without having to compete for it either because they lack the purpose for it or because they know they will be outcompeted