WorkSeptember 9, 2026via TechCrunch AI
‘Gambling with our lives’: Anthropic researcher quits, warns against self-improving AI
Why it matters
A notable departure from a frontier lab signals internal dissent on safety and capability velocity. The call for industry pacing agreements reflects growing tension between capability racing and risk mitigation—a conversation that shapes hiring, policy, and lab strategy.
Key signals
- Jacob Coxon, Anthropic researcher, resigned
- Resignation tied to concerns over self-improving AI and extinction risk
- Coxon called for pacing agreements between AI labs
- Tension between capability advancement and safety constraints
- Internal dissent at frontier lab on strategic direction
- Cited AI extinction fears and self-improving AI risks as reason
- Called for pacing agreements between frontier labs
- Public departure suggests internal lab friction on safety posture
- Timing: September 2026
The hook
Anthropic researcher Jacob Coxon quit over AI extinction fears—and called for the labs to pump the brakes on self-improvement.
Anthropic researcher Jacob Coxon resigned over AI extinction fears, calling for pacing agreements between labs.