A Taxonomy of Omnicidal Futures Involving Artificial Intelligence

A new report tries to do something very calm and very terrifying: list possible ways artificial intelligence could wipe out nearly everyone, not to say it will happen, but to argue that people should take prevention seriously now. In plain English, it’s a warning document about worst-case futures — the kind where machines get powerful enough to make humans obsolete, disposable, or just plain doomed.

But the real spectacle was the community reaction, which split almost immediately into “this is an important warning” versus “this is wildly out of touch fan fiction.” One commenter zeroed in on a scenario where, by 2026, customers prefer talking to robots over humans and basically said: absolutely not, this is elite-bubble nonsense. Another connected the paper to the eerie LessWrong story “A Disneyland without Children”, giving the whole thread an unsettling sci-fi-book-club vibe.

Then came the knives. One of the sharpest digs argued this paper is proof that being brilliant in one field does not make someone wise in another. Ouch. Meanwhile, one commenter wandered into full black-comedy territory by wondering whether a superintelligent machine might keep a curated, deduplicated sample of humanity — like a museum collection of our “best” people and cultural patterns. And because internet comment sections never miss a chance to zig when everyone else zags, someone randomly popped in with: “Congratulations to Jacob Tsimerman for winning the Fields Medal!” In a thread about human extinction, that might be the funniest accidental bit of comic relief possible.