It really is in there.
Anthropic’s IPO filing has a long section on what could go wrong, and one of the entries is the end of the world: advanced AI could “pose catastrophic or existential risks to humanity”. Crypto Twitter screenshotted the sentence, started calling it page 31, and drew the risk itself, in a suit, with a coffee, under an Employee of the Month plaque, the department in pieces behind it.
The sentence, in the filing itself (a very large page). Later amendments push it to page 32 or 33; 31 is the number that stuck.

