Jacob Coxon became the face of AI safety anxiety in September when he quit Anthropic with a viral warning that frontier labs were racing toward superintelligence. This week the conversation turned to his wallet. In a Stanford HAI conversation with professor Rob Reich, Coxon confirmed he still holds equity in OpenAI, where he worked for about three years, and said: “I like money, bro. I’m like, I agree, it’s blood money.” On X, the reaction ranged from “hypocrisy” to “mad respect,” especially after he replied to a challenge with: “you can count this as a pledge.”
Quick Take
- What he said: Asked “do you have equity in OpenAI?”, Coxon answered “Yeah, I do,” admitted it would be easier to argue his case if he gave up the stock, and said “it’s nice money. Like, I like it.”
- The pledge: When Wagner James Au (@slhamlet) asked on X whether he would pledge to donate all his OpenAI equity to charity when OpenAI goes public, Coxon replied: “In all the good worlds the stock goes to zero anyway. But sure, you can count this as a pledge.”
- The backstory: Coxon resigned from Anthropic on September 8 after about four months, warning about recursive self-improvement and superintelligence. He forfeited unvested Anthropic stock, according to multiple accounts.
- Second headline: In the same talk he said he was “not sure I made the right decision in leaving.” Days later he wrote on X: “I won’t be attempting to rejoin anthropic.”
- Why it matters: The debate tests whether AI insiders who warn about the industry can credibly keep profiting from it.
In plain English
A young AI researcher quit his job at Anthropic in September and warned publicly that AI companies are moving dangerously fast. At a Stanford event, he admitted he still owns valuable shares in OpenAI, his previous employer, and jokingly called the money “blood money” that he likes anyway. Many people called that hypocritical. Shortly after, when someone asked him on X to promise to give the shares to charity when OpenAI goes public, he said yes. People are now arguing about whether his warnings are still believable.
What happened
The conversation was hosted by Stanford HAI and the AI Governance Lab and published on November 13, according to CEOInterviews.AI, which transcribed it. Several posts on X date the event itself to November 9. About eleven minutes in, Reich asked directly: “So for you, do you have equity in OpenAI?”
Coxon: “Yeah, I do.” Asked how people should view his financial interest, he said: “I agree it would probably be easier for me to argue with that if I, like, you know, ceded rights to my OpenAI stock, but it’s like, it’s nice money. Like, I like it.” Reich: “But you’re not.” Coxon: “I’m not. No. Why not? I like money, bro. I’m like, I agree, it’s blood money.”
He then argued that donating would not change anything: “if I thought for a moment that donating it like would change anything, then sure, I’d do it.” And he added: “my parents were like worried I was having a psychotic break. I think if I also given away all the money, they’d be even more convinced of that.” Reich called one of his answers “a real dodge.”
On X the same day, Coxon posted a chart that he said came from Anthropic’s IPO roadshow and “was a big factor in my decision to leave,” showing what he described as superexponential gains in pretraining compute efficiency. Under that post, Wagner James Au (@slhamlet) asked: “Jacob, will you pledge to donate all your OpenAI equity to charity when that startup goes public?” Coxon answered: “In all the good worlds the stock goes to zero anyway. But sure, you can count this as a pledge.”
Background: who is Jacob Coxon?
According to Reich’s introduction, Coxon studied computer science at the University of Cambridge, spent roughly three years at OpenAI and then less than half a year at Anthropic, working on pretraining. He resigned publicly on September 8, 2026, in an X thread that Reich said drew around “175 million impressions or views.” We covered his resignation at the time.
In the Stanford talk, Coxon explained the decision: he saw “exact projections for the capabilities of models in the next six months to one year” that made him think “we’re on the cusp of recursive self-improvement,” and that superintelligence could arrive “before the decade is out.” His remedy is political: “Nation-state coordination, I think we need ultimately.”
Why equity matters in AI safety debates
Employee stock at frontier labs can be worth a lot, and its value depends on the labs continuing to race. That creates an obvious tension for anyone who leaves a lab to warn against that race. Reich framed it as moral credibility: giving up the shares “might give you more moral credibility though.” Coxon’s counter was that the question itself creates the credibility problem: “if you hadn’t asked it, then the moral credibility wouldn’t have been questioned.”
His later X pledge is conditional on an OpenAI listing, which has not happened. He has not published how much equity he holds. A widely shared estimate of about $30 million came from an account that asked Claude to guess, and it is labeled by its author as a ballpark, not a reported figure.
Pros and cons
For users
- Pro: The episode makes visible something the public rarely sees: how insiders’ finances and convictions interact.
- Con: Personality drama can crowd out the substance of what Coxon is warning about, such as the pace of capability gains.
For developers
- Pro: AI researchers get a public case study in what it costs to speak out, and what questions to expect if they do.
- Con: The pile-on may discourage others from speaking candidly, which Coxon himself hinted at when he wrote that he had “learned my lesson about nuance on camera.”
For companies
- Pro: For OpenAI and Anthropic, the debate shifts attention from the critic’s arguments to the critic’s consistency.
- Con: It also keeps a former insider’s claims about their roadmaps, including the IPO roadshow chart, in the headlines during Anthropic’s IPO week.
What X is saying
The clip that drove the story:
“I like money, bro. I agree. It’s blood money.”
— Rohan Paul (@rohanpaul_ai) November 15, 2026 · View on X
Jacob Coxon, the researcher who sensationally quit Anthropic, in a new interview at Stanford.
Pressed by Rob Reich on whether keeping his OpenAI equity hurts his credibility, Coxon openly admits he still holds it, likes the money, and agrees it is “blood money.” He argues that giving it away would not change anything
Coxon’s own reply that became the “pledge”:
@slhamlet In all the good worlds the stock goes to zero anyway.
— Jacob Coxon (@hilbertspaess) November 13, 2026 · View on X
But sure, you can count this as a pledge.
Critics did not hold back. “Jacob Coxon admits he is keeping ‘blood money.’ This is a chilling worldview,” wrote @NathanLeamerDC. Supporters focused on the pledge: “Jacob Coxon just pledged to donate 100% of his OpenAI equity to charity after the IPO,” posted @AISafetyMemes, adding “Mad respect.” Writer @bicellaria noted simply: “Oh cool, Jacob Coxon has pledged to donate all of his OpenAI equity to charity when they IPO.”
On the “maybe I’ll go back” remark, Coxon later clarified on X that it “was a combination of a bit and an attempt to express deep uncertainty about the deontological stance I took by leaving,” and wrote: “I won’t be attempting to rejoin anthropic.”
Grok’s take
Summarizing the story, Grok wrote: “Jacob Coxon (ex-OpenAI/Anthropic researcher) quit Anthropic in Sept over AI extinction risks, forfeiting unvested equity there. In a recent Stanford interview he confirmed he still holds his OpenAI shares.” It added that he “views the gains as tied to the same race he says risks humanity, but keeps them anyway.”
Critical analysis
Two questions are getting mixed up. One is about character: is it consistent to call your shares “blood money” and keep them? Coxon’s own on-stage answer was weak, and Reich said so. His later pledge, made in a one-line reply on X, partly answers the critics, but it is informal, conditional on an OpenAI IPO and unverifiable today.
The other question is about substance: is he right about the pace of AI progress? His financial position does not make his technical claims true or false. If anything, holding OpenAI stock while warning that OpenAI’s work is dangerous cuts against his financial interest, not for it. The chart he shared, which he says comes from Anthropic’s IPO roadshow, has no labeled units or absolute numbers, which Coxon himself acknowledged on X, citing his NDA. That is the part of his argument that deserves the most scrutiny.
It is also worth noting how the story spread: through short, sharp clips. The full hour covers his reasoning, his doubts about whether leaving was the right call and his push for international coordination. Twelve seconds about money traveled furthest.
What’s still open
- Whether Coxon formalizes the pledge, for example through a written commitment or a named charity.
- Whether and when OpenAI goes public, which determines if the pledge ever applies.
- Whether Anthropic comments on the roadshow chart Coxon has been circulating.
- What Coxon does next: he said he expected to be “in London writing Substacks,” but remains in the Bay Area.
About the analysis
Researched and written with AI assistance under the editorial responsibility of the i10X News team. The title image is an AI-generated image. Read our AI Use Policy. Spotted an error? Request a correction.
This independent i10x analysis is based on the CEOInterviews.AI transcript of the Stanford HAI conversation between Jacob Coxon and Rob Reich, Coxon’s own posts on X, reactions on X, and our earlier coverage of his resignation. Quotes from the event are taken from the published transcript.
i10x Perspective
The AI safety debate has a credibility economy, and Coxon just learned its exchange rate. “I like money, bro” is the kind of line that outlives any argument made around it. The pledge helps, but it is a reply, not a contract. The more useful question for everyone else is the one he keeps pointing at: how fast are the labs really improving, and who outside them gets to see the numbers? Judge the messenger if you like, but check the chart.






