Former Anthropic researcher Jacob Coxon said at a November 9 Stanford event that he still held OpenAI equity, agreed it was “blood money” and offered a blunt explanation for keeping it: “I like money, bro.” Source
Coxon spoke with Stanford professor Rob Reich in a conversation hosted by Stanford HAI and the AI Governance Lab, uploaded to Stanford HAI’s YouTube channel on November 13. He also questioned whether leaving Anthropic had been the right decision and floated returning to lab work. Source
Reich wasn’t buying the answer
Reich asked how people should view Coxon’s financial interest in OpenAI doing well. When Reich suggested he could take the moral high ground by giving up that interest, Coxon agreed he could—and acknowledged he wasn’t doing so. Source
Coxon said he would donate if he believed doing so would change anything. Reich pressed a different point: giving it away might improve Coxon’s moral credibility. Source
“I agree, but you could also stop asking the question,” Coxon replied. He then suggested that, without the question, his moral credibility wouldn’t have been challenged. Source
The question, inconveniently, survived the answer.
“That’s a real dodge, Jacob. That’s a real dodge,” Reich said. Source
Coxon didn’t say how much OpenAI equity he held or what it was worth. Source
Second thoughts aren’t a safety all-clear
Asked why he hesitated to encourage other researchers to leave, he said: “I’m not sure I made the right decision in leaving.” He questioned whether staying and doing safety research might have been more effective. Source
He floated the possibility that “maybe I’ll go back,” but announced no employer or return arrangement. He also described himself as confused about the situation. Source
Earlier in the conversation, Coxon said he was “genuinely quite scared” about what AI would look like in 2030. He described his departure as a cumulative decision, rather than a response to one immediate incident. Source
He rejected a blanket call for researchers to resign without international coordination, arguing that development could otherwise continue with people less concerned about the technology’s impact. He said countries ultimately needed to coordinate on regulation. Quitting a lab, in his account, doesn’t by itself stop the race.
Comments