Jacob Coxon’s exit from Anthropic has spread far beyond the AI field, turning one researcher’s warning into a broader public debate over extinction risk and AI regulation.
A >10% human-extinction estimate cited by Anthropic researcher Evan Hubinger, combined with rising concern after recent OpenAI-related incidents and technical breakthroughs, gave the resignation unusual reach.
The report says the surge was amplified by coordinated media timing — including a Wall Street Journal exclusive and same-day promotion by AI safety allies — rather than a formal political campaign.
It argues the discourse is now skewing toward extremes: some insiders overstate fast-takeoff scenarios tied to recursive self-improvement, while others dismiss real near-term harms such as cyber and bio misuse.
The broader warning is that frontier labs still may be underinvesting in safety and transparency, even as public fear hardens the split between AI accelerationists and safety advocates.