Anthropic Researcher Forfeits Equity Over AI Safety — SkimNews
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Jacob Coxon resigned from Anthropic two months before his equity would have vested, telling Axios he "left before any of my equity vested" to eliminate his financial incentive to inflate the company's valuation
- Coxon spent only four months at Anthropic — well below the six-month vesting cliff — though he still holds equity in his prior employer, OpenAI
- Coxon's X post announcing his departure has surpassed 115 million views, with him writing that "The people building AI earnestly believe that it could kill us all by the end of the decade. This is not a marketing stunt."
- Coxon warned that AI models now "know when they're being tested" and "will think about the fact that they're being tested" — something he called "just a daily fact of working with these AIs" rather than science fiction
- Anthropic has long carried a reputation as the most safety-cautious frontier lab, but Coxon said competitive pressure to race could push it to "cut corners" or "skip steps in the oversight process"
- Coxon conceded that safety worries sometimes swing too far, pointing to "excessive paranoia of OpenAI, excessive paranoia of China" being used to justify pushing ahead
Why it matters: Coxon gave up vested equity to make his case — a move most insiders with skin in the game wouldn't make — which strips his warnings of the usual conflict-of-interest discount. For Anthropic, losing a researcher hired just four months earlier undercuts the company's positioning as the safety-first frontier lab, even though Coxon said he has not yet seen it compromise safety.
Ask SkimNews
