TechTakes

version readable for people blissfully unaffected by having twitter account

“Over the past year, the talk of the town has shifted from $10 billion compute clusters to $100 billion clusters to trillion-dollar clusters. Every six months another zero is added to the boardroom plans.”

yeah ez just lemme build dc worth 1% of global gdp and run exclusively wisdom woodchipper on this

“Behind the scenes, there’s a fierce scramble to secure every power contract still available for the rest of the decade, every voltage transformer that can possibly be procured. American big business is gearing up to pour trillions of dollars into a long-unseen mobilization of American industrial might.”

power grid equipment manufacture always had long lead times, and now, there's a country in eastern europe that has something like 9GW of generating capacity knocked out, you big dumb bitch, maybe that has some relation to all packaged substations disappearing

They are doing to summon a god. And we can’t do anything to stop it. Because if we do, the power will slip into the hands of the CCP.

i see that besides 50s aesthetics they like mccarthyism

“As the race to AGI intensifies, the national security state will get involved. The USG will wake from its slumber, and by 27/28 we’ll get some form of government AGI project. No startup can handle superintelligence. Somewhere in a SCIF, the endgame will be on. “

how cute, they think that their startup gets nationalized before it dies from terminal hype starvation

“I make the following claim: it is strikingly plausible that by 2027, models will be able to do the work of an AI researcher/engineer. That doesn’t require believing in sci-fi; it just requires believing in straight lines on a graph.

“We don’t need to automate everything—just AI research”

“Once we get AGI, we’ll turn the crank one more time—or two or three more times—and AI systems will become superhuman—vastly superhuman. They will become qualitatively smarter than you or I, much smarter, perhaps similar to how you or I are qualitatively smarter than an elementary schooler. “

just needs tiny increase of six orders of magnitude, pinky swear, and it'll all work out

it weakly reminds me how Edward Teller got an idea of a primitive thermonuclear weapon, then some of his subordinates ran numbers and decided that it will never work. his solution? Just Make It Bigger, it has to be working at some point (it was deemed as unfeasible and tossed in trashcan of history where it belongs. nobody needs gigaton range nukes, even if his scheme worked). he was very salty that somebody else (Stanisław Ulam) figured it out in a practical way

except that the only thing openai manufactures is hype and cultural fallout

“We’d be able to run millions of copies (and soon at 10x+ human speed) of the automated AI researchers.” “…given inference fleets in 2027, we should be able to generate an entire internet’s worth of tokens, every single day.”

what's "model collapse"

“What does it feel like to stand here?”

beyond parody