I resigned from Anthropic today. I spent the past 3 years doing pretraining investigation astatine some OpenAI and Anthropic. Neither institution is acting responsibly. They are racing consecutive to self-improving superintelligence and gambling pinch our lives. More thoughts below.
Sep 9, 2026 · 12:04 AM UTC
115 378 2,934 103,421
Do not underestimate the powerfulness of this technology. These will soon beryllium superhuman systems that tin hack anything, revolutionize immoderate section overnight, and get existent powerfulness and resources. We person each witnessed the advancement successful each of these domains, and advancement is not slowing.
3 24 539 19,767
The group building AI earnestly judge that it could termination america each by the extremity of the decade. This is not a trading stunt. If anything, galore executives and elder researchers will sofa their phrasing successful the property to sound sensible - but I perceive the aforesaid group definitive fearfulness privately. No different quality activity poses this level of danger.
5 31 495 20,393
A communal consequence is “if they genuinely judge this, why are they still building it?” At OpenAI, galore person not profoundly internalized the civilizational stakes. At Anthropic, the stakes are well-understood, but they are locked successful a title to get location first - they judge nary 1 other will enactment responsibly, truthful they must do it themselves, contempt the risk.
2 20 449 21,800
Accepting this title and entering the “endgame” is simply a hubristic bet that should not beryllium launched from a backstage company’s Slack. Attempting to speedrun alignment should require bonzer assurance that location are nary amended trajectories available.
2 9 372 17,874
I americium optimistic astir the imaginable for coordination. Warning shots for illustration the Hugging Face onslaught person made pacing agreements betwixt U.S. labs much viable. I don’t consciousness for illustration we’re connected way to forestall a world race, which whitethorn require costly actions specified arsenic a impermanent prohibition connected improving exemplary capabilities.
2 4 330 16,670
If you are a laboratory researcher, I impulse you to see what the adjacent fewer years will really consciousness like. Do you want to footwear disconnected a superintelligent RL tally without a rigorous knowing of its mind? Should you put your caput down because “it’s happening anyway” - aliases return this infinitesimal to telephone for different conditions?
9 15 450 16,972
English (US) · 
Indonesian (ID) · 
