3:40 AM, April 10
In San Francisco's Russian Hill neighborhood, someone threw a Molotov cocktail onto the doorstep of OpenAI CEO Sam Altman's home.
The arrested individual is 20-year-old Daniel Alejandro Moreno-Gama. Surveillance footage captured the attack, and security personnel quickly extinguished the fire. No one was injured. Moreno-Gama then went to OpenAI headquarters, threatened to burn down the building, and was arrested there.
His Discord handle was "Butlerian Jihadist," a reference to the Dune series — a story in which humanity destroys all machines.
What He Was Thinking
Moreno-Gama was a member of the PauseAI community, a group calling for an immediate halt to AI development due to the risk of human extinction. He published six articles on Substack between January and March 2026, including one titled "Eulogy for Humanity."
In December, he wrote on Discord: "We are close to midnight. It's time for real action." Community administrators warned him that advocating violence would result in a ban.
PauseAI has explicitly condemned the attack, stating that violence is completely contrary to their position. They noted that Moreno-Gama had posted only 34 messages on Discord and had almost no substantive engagement with the community.
But regardless of organizational ties, the incident shows that extreme anxiety about AI is no longer confined to the internet.
Same Week, The New Yorker
Around the same time, The New Yorker published a deep investigation by Ronan Farrow, titled "Sam Altman Might Control Our Future — Can He Be Trusted?"
Based on hundreds of interviews and internal documents, the report accuses Altman of systematic deception, excessive concentration of power, and a series of controversial decisions. The phrase "compulsive liar" appears in the article.
Altman responded on his blog, calling the piece a "hit job" and acknowledging that he had underestimated "the power of language and narrative." He urged the tech industry to reduce inflammatory rhetoric.
He said he was warned that the article, "published in an era of intense AI anxiety, could make the situation more dangerous for him."
Two Events, One Picture
The coincidence of these two events is not accidental; it is a symptom.
AI has moved incredibly fast in recent years: GPT-5.4, Mythos discovering thousands of zero-day vulnerabilities, AI models helping people write code, make decisions, and read medical images. Some see efficiency and possibility; others see loss of control and threat.
Groups like PauseAI are not new. There have been petitions, protests, and letters to Congress. But this is the first time someone has turned to violence — even if only a Molotov cocktail.
Moreno-Gama does not represent the entire AI-skeptic community, just as every extremist does not represent the larger group they come from. But he represents an escalation of emotion: when a person feels that "nothing works, only action," he wrote "close to midnight."
What Altman Said
Altman's blog response was measured. He did not condemn the direction of The New Yorker's investigation but focused on urging industry figures to be mindful of their language. He said he hoped for "fewer explosions at fewer doorsteps."
It was a darkly humorous line, but the underlying message was serious: the way the AI industry talks is itself influencing the public's emotions about the field.
Whether Altman is truly as The New Yorker portrayed him is a question that requires factual verification. But his judgment about "AI anxiety and public narrative" proved accurate at this moment.
This Is Not the End
Moreno-Gama was arrested, PauseAI condemned him, the media reported it, and the story will continue.
The anxiety that drove his decision will not disappear. AI systems are rapidly surpassing ordinary people's understanding of them, and the gap between them is fertile ground for fear. How large model companies communicate what they are doing to the public, and how regulators race to keep pace, will influence many more people's choices.
Someone wrote "we are close to midnight" and threw a Molotov cocktail. This is not evidence of an AI apocalypse. But it is a very real signal of the level of social anxiety.
Sources: Sam Altman responds to 'incendiary' New Yorker article after attack on his home (TechCrunch); Man who firebombed Sam Altman's home was likely driven by AI extinction fears (The Decoder); OpenAI published a New Deal for AI. Days later, CocoLoop, someone firebombed Sam Altman's house. (SF Standard); Sam Altman Confirms Molotov Cocktail Incident and Responds to Incendiary New Yorker Investigation (Hollywood Reporter)