The Great Molotov Bounce: Sam Altman’s Unofficial AGI Stress Test
At 3:45 AM Pacific Time on April 10, 2026, while most of us were dreaming of robots taking our jobs (but hopefully not our lunch), a window in San Francisco performed a feat of physics that deserves its own Nobel Prize: it bounced a Molotov cocktail.
"Through some divine comedy of material science, a bottle destined to ignite history instead performed a 'Boing!' that would make a rubber ball jealous."
The Darwin Award for Escape Plans
We have to talk about the suspect. After failing his audition for professional arsonist, he didn't disappear into the night. Instead, he made his way to OpenAI Headquarters to announce his plans to burn that down too.
It’s like robbing a bank and then deciding the best place to hide is the police station's lobby while asking for a rewards card. Thanks to the prompt work of the SFPD, this tactical genius is now enjoying a stay in pre-paid, low-compute housing.
The Sleepless CEO
Sam Altman’s reaction was... uniquely Sam. He posted a blog on Friday night (April 11) that was equal parts heartfelt and "I’m so done with this week."
- Family Photo Defense: Sharing a picture of his loved ones to counter a literal firebomb.
- Giving Credit to Physics: Thanking the heavens nobody was hurt.
- The New Yorker Shade: He blamed a recent "incendiary" article (pun intended?) for stirring up the vibes that led to the attack.
Sam admitted he used to "brush aside" the narrative power of words, but now he’s "awake in the middle of the night and pissed." It turns out, even when you're building a god-in-a-box, human-written sentences can still ruin your sleep.
Wisdom from the Blog:
"Once you see AGI you can’t unsee it. It has a real 'ring of power' dynamic to it, and makes people do crazy things."
— Sam Altman, clearly mid-LOTR marathon
Verdict: Glass is More Reliable than AGI
Lessons learned: First, if you're going to target a tech titan, research their double-glazing specifications first. Second, if your escape route leads to the victim's office, your own intelligence might need some reinforcement learning.
Sam concluded by calling for a "de-escalation of rhetoric" and "fewer explosions in fewer homes." We heartily agree. Let's keep the fire for the prompts, not the property.