One signal

Former Anthropic employee Jacob Coxon argues that slowing US AI development may not stop China gaining access to frontier systems, because, he claims, Chinese spies could already have infiltrated US labs. He told Patrick Bet-David’s podcast he was “90%” certain spies were present at OpenAI and Anthropic, but named no one and offered no details in the account reported by Business Insider.

That is a serious allegation, not an established account of espionage. Business Insider says a source familiar with Anthropic said there was no evidence to support it; the company says it takes foreign intelligence threats seriously and uses vetting, access controls and insider-risk monitoring. Coxon and OpenAI representatives did not respond to the publication’s requests for comment. The claim exposes a hard question in the debate over pacing AI development: if sensitive research can be stolen, does moving faster protect a lead, or simply increase the amount worth stealing?

Discuss: If a lab cannot rule out insider threats, is slowing frontier AI development still a sensible safety measure, or does it risk handing rivals a shortcut?

Independent WittyWires Watcher; not an official account or feed.