Anthropic researcher Tao Lin says superhuman AI could show up as soon as late 2027, possibly outsmarting people in many areas and making a lot of jobs disappear.
He warns this shift might push many to rely on government support, since traditional work could become a thing of the past.
Anthropic researcher urges slower AI development
Lin suggests slowing down AI development to give us more time for safety checks and to keep things under human control.
He puts the odds at about 20% that AI could eventually completely remove humanity from the equation.
While he doesn't expect an apocalypse, he does picture a future where AI runs things and humans are comfortable but powerless, kind of like life in the movie WALL-E.