VOYTEX: museli by mit uplne jinou, tajnou architekturu nez jsou ty aktualne popularni transformery, protoze na trenovani tech modelu potrebujes tolik GW ze to nejde jen tak utajit, takova spotreba. coz jako ... nevim no. vedelo by se, ze nejlepsi AI mozky zahadne mizi nekam na projekt manhattan. coz se nutne stane, ale jeste se to nedeje
nezda se mi ze by US do 2030 stihlo projekty ve smyslu nize, Cina v pohode, maji jiny styl prace. Z hlediska emisi se to asi bude hodne zhorsovat. Myslet si, ze AI je k nicemu a neprosadi se, to se bojim ze jako Vaclav Klaus rikaval ze internet je jen nahrazka Zlutych stranek a on by se tim nevzrusoval (to ma byt na Shefika ze to prece nic moc neumi)
According to Aschenbrenner, by 2028 the most advanced models will run on 10 gigawatts of power at a cost of several hundred billion dollars. By 2030, they’ll run at 100 gigawatts at a cost of a trillion dollars.
For context, a typical power plant delivers something in the range of 1 gigawatt or so. So that means building 10 power plants in addition to the supercomputer cluster by 2028. What would all those power stations run on? According to Aschenbrenner, on natural gas. “Even the 100 [gigawatt] cluster is surprisingly doable,” he writes, because that would take only about 1,200 or so new natural gas wells. And if that doesn’t work, I guess they can just go the Sam Altman-way and switch to nuclear fusion power.