Table of Contents
ToggleAt 3 a.m., the streets of Russian Hill in San Francisco were as quiet as a postcard. After all, this is an upscale residential area.
No one saw a young man walk past; only the surveillance cameras showed a 20-year-old approaching the metal door of a house on the north side of Chestnut Street and throwing a bottle he was holding. It was a simple Molotov cocktail filled with kerosene and stuffed with cotton; the laws of physics caused it to bounce off the door and ignite. It was just a small fire; no one was injured.
But the owner of this house is Sam Altman.
Crime scene
The court files reconstructed the suspect's movements that night, and read like the narration of a low-budget thriller.
At 3:37 a.m., a Molotov cocktail was thrown out, and a small fire started outside the door. Altman's residence was not further damaged.
Eighty-four minutes later, the same person appeared at the entrance of OpenAI headquarters, several kilometers away.
Around 5:00 AM, surveillance footage showed him picking up a chair and smashing it against the building's glass door. When security arrived, he said one sentence: He was going to burn down the building and kill everyone inside.
Police arrested him at the scene, and what they found on him elevated the incident from a "midnight intrusion with a mental problem" to another level, as he had a kerosene canister, a blue lighter, additional incendiary devices, and an unregistered handgun.
There is also a paper file.
A Manifesto for Altman
The federal indictment cited this two-part file, the first part of which is titled "Your Last Warning".
It claims the author has "killed or attempted to kill" Sam Altman and admits to the intent. This is followed by a list of names and addresses of several AI company CEOs, board members, and investors.
The indictment uses the term "hit list."
This document calls on others to join his "movement," while the remaining second part discusses the threat of AI to human extinction, concluding with a sentence directly addressed to Altman:
"If you miraculously survive, I will take it as a sign from God that you are being redeemed."
This was an assassination on near-religious grounds.
Who is the suspect?
Daniel Alejandro Moreno-Gama, 20, is from Spring, a suburb of Houston, a place called The Woodlands.
At 20, federal prosecutor Craig Missakian filed two counts of attempted murder (of Sam Altman himself and a security guard) plus attempted arson. Federal charges include intentional property damage with explosives and possession of an unregistered firearm.
On Monday morning, April 13, the FBI conducted a search of the suspect's home and left after several hours. Acting Special Coordinator Matt Cobo of the FBI said at a press conference:
This is not an impulsive crime. This is premeditated, targeted, and extremely serious.
FBI Director Kash Patel personally oversaw the operation in Texas. Prosecutor Missakian added, with a rare directness among law enforcement officials: "If evidence shows that Moreno-Gama carried out these attacks to change public policy or coerce government officials, we will prosecute them for domestic terrorism."
Domestic terrorism – this is the first time the term has been used to describe “disrupting the development of AI.”
Another shot on Sunday morning
The story didn't end on Friday.
On Sunday morning, April 12, a Honda sedan slowly drove along Chestnut Street in Russian Hill, made a U-turn in front of Altman's residence, and then fired a shot at the house.
The two suspects were subsequently arrested: Amanda Tom, 25, and Muhamad Tarik Hussein, 23.
The police statement was cautious and brief: "There is currently no evidence to suggest that the two incidents are related."
This is legal language, but it also implies another possibility: perhaps this isn't complicity, or perhaps it's worse. Perhaps it means that the Molotov cocktail incident itself has become a "model case," being interpreted by others in their own ways.
The copycat began to replicate itself.
Sam Altman's sleepless night
On the afternoon or evening of April 10, the same day as the Molotov cocktail incident, Sam Altman posted on his personal blog, along with photos of his family.
He began like this:
"We usually try to keep a low profile, but this time I shared photos of my family in the hope that it will deter the next person who wants to throw a petrol bomb at my house, no matter what they think of me."
…Now I wake up in the middle of the night, angry, and starting to think: I underestimated the power of words and narrative.
He did not name names, but the article is traceable; a recent sharp critique in The New Yorker described Sam Altman as a dangerous technological utopian evangelist.
That article circulated in anti-AI communities, was screenshotted, taken out of context, and used as a quote in various posts advocating that "AI must be stopped."
Altman called for "less confrontational rhetoric" on his blog, welcoming "good-faith criticism and debate," and concluded with:
"I understand the anti-technology sentiment, and I know that technology isn't always good for everyone. But overall, I believe that technological progress can make the future incredibly wonderful, for your family and my family as well."
As we continue to debate these issues, we should reduce the escalation of rhetoric and tactics, and strive to minimize the number of explosions, both metaphorically and in reality.
A man who woke up in the middle of the night, with his husband and children at home, has just learned that someone wants to burn down his house and is trying to persuade this generation with words.
Narrative as a weapon
Now we need to talk about something even more uncomfortable.
Sam Altman's brief discussion of "confrontational debate" on his blog is worth serious consideration: how narrative can become a weapon.
In the blueprint for an AI-based society, and against AI leading humanity... well... will everything become part of AI in the future? Will there still be humanity left?
This confrontation is dangerous and could create two opposing camps: one that supports AI and the other that opposes it.
If federal prosecutors ultimately indict Moreno-Gama on domestic terrorism charges, it will set an unprecedented precedent in U.S. legal history, marking the first criminal case motivated by "anti-AI ideology" to be classified as terrorism.
The legal threshold for this charge is very high; it must be proven that the action was intended to change public policy or coerce government agencies.
The most crucial piece of evidence was the passage in the "Your Last Warning" manifesto carried by the suspect that called on others to join the "movement." This wasn't an angry individual; this was someone trying to launch a broader campaign.
The prosecutor stated, "We will not tolerate any attempt to alter the way Americans live and work through fear or violence."
The fear of AI is likely to become a narrative enemy, and then weaponized. This confrontation could be even more intense if the suspect is actually charged with terrorism.
The development direction of the AI industry will not be changed by a Molotov cocktail. But we can speculate that in the future, those who try to question the development of AI will be thrown to the other side of the crowd and labeled as "confrontational elements."
I hope we will all be well in the future.
