A 20‑year‑old identified as Daniel Moreno‑Gama threw a Molotov cocktail at Sam Altman's residence at 3:45 a.m., then walked three miles to the OpenAI headquarters and threatened to set it ablaze. Police booked him on suspicion of attempted murder. Investigators found he was not acting alone; he held six roles in the online community PauseAI, where his Discord handle read “Butlerian Jihadist.”

Moreno‑Gama’s social feeds were saturated with doomer rhetoric: Instagram posts paired AI capability curves with captions warning that humanity would die without immediate action, and a Substack essay titled “AI Existential Risk” claimed extinction from AI was “nearly certain.” He repeatedly cited Yudkowsky and Soares’ pamphlet *If Anyone Builds It, Everyone Dies*, urging readers to “deal with the threat first.”

Community members responded by deleting his messages and distancing themselves, yet the episode illustrates how a certainty‑driven risk narrative can motivate violent extremism. PauseAI later removed his Discord posts, and its spokesperson has denied endorsing any illegal action. The incident forces the AI safety discourse to confront the practical hazards of rhetoric that treats existential risk as a justification for real‑world attacks.