Jacob Coxon, a researcher who spent three years working on neural network pretraining at OpenAI and Anthropic, announced his resignation from Anthropic on September 9, 2026, warning that leading AI laboratories are engaged in an irresponsible race toward self-improving superintelligence.
In a resignation manifesto that quickly spread across technology and investment circles, Coxon stated: 'I resigned from Anthropic today. I spent the last three years doing pretraining research at both OpenAI and Anthropic. Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives.'
The post drew a public reaction from Emi Yoshikawa, former Ripple vice president of global operations and now a venture capital figure. Yoshikawa quoted an insider and said the manifesto was gaining traction, adding that developers inside Anthropic genuinely believe artificial intelligence could destroy humanity within a few years.
Coxon highlighted a cultural difference between the two labs: many OpenAI employees still do not fully grasp the scale of the risk, while Anthropic employees recognize the danger but continue accelerating development out of fear of losing the competitive race. He called on engineers to reflect on their work and urged regulators to create international agreements governing the pace of AI development, including a temporary ban on advancing core AI capabilities.
Meanwhile, Wall Street is placing an enormous financial bet on AI infrastructure. Hyperscalers are targeting roughly $700 billion in AI infrastructure investments, increasingly financed through bonds, private credit, and complex deal structures backed by about $1.65 trillion in debt. The contrast between existential AI safety warnings and large leveraged investments has raised concerns among analysts that weak AI returns could trigger a broader financial shock, with spillover risks for crypto markets.
Fintech investors and venture capital funds are beginning to treat internal turmoil at AI laboratories as a direct financial risk. For Anthropic and OpenAI, Coxon's revelations represent not only a reputational blow but also a potential threat to valuations ahead of new funding rounds.
