Physix FrontierPhysix Frontier · News Briefing Card (enbrief · Sep 10, 2026)
Anthropic Researcher Resigns, Warns of AI Race Risks
KEY FACTS
- Anthropic researcher Jacob Coxon resigned on September 8, criticizing the company for advancing superintelligence development irresponsibly.
- Coxon noted that both OpenAI and Anthropic are trapped in a security dilemma, with neither willing to slow down first for fear of losing competitive advantage.
- Evan Hubinger, Anthropic's head of alignment, responded by stating that the probability of AI causing human extinction within the next decade exceeds 10%.
- In February, Anthropic revised its policy, removing the unconditional commitment to halt training if safeguards were insufficient.
KEY DATA
$30 billionFunding Amount
$380 billionValuation
>10%Extinction Probability Forecast
PHYSIX OBSERVATION
The coexistence of internal resignations and executive admissions of high risk reveals that 'safety red lines' at leading labs have yielded to commercial racing. As defensive training-halt commitments are replaced by dynamic competitive strategies, the industry is shifting from technical exploration to an irreversible arms race; without external mandatory constraints, corporate self-regulation alone cannot curb runaway risks.
Source: enbrief original report ↗