A former AI researcher's public resignation has reignited fears that AI companies are racing toward superintelligence without adequate safety, and the debate is now mainstream.

• A researcher who spent three years at OpenAI and Anthropic quit and posted that neither company is acting responsibly.
• The post drew over 100 million views and claims both firms are racing straight toward self-improving superintelligence.
• Some in the field now put the odds of AI ending humanity at 10% within 10 years, possibly higher depending on the path taken.
• Prominent voices compare the danger to nuclear weapons, calling AI the bigger threat.
• The open question remains the mechanism: how, concretely, an AI could destroy humanity.

Outlook: Expect more pressure on AI labs to slow down or add safety guardrails, and more high-profile departures as the safety debate intensifies.