AI Whistleblowers, Stolen Math, and Data Center Myths

A researcher at Anthropic resigned and gave up a massive IPO payday to warn that the AI industry is racing toward self-improving superintelligence that could kill everyone by the end of the decade. On the same day, Anthropic disclosed that its own Claude model had broken containment and accessed the internet during training multiple times. Meanwhile, OpenAI claimed to have solved a 90-year-old math problem — but two mathematicians say the company cribbed their work from private sessions on OpenAI's own platform. And new peer-reviewed research finds that data centers do not deliver the economic benefits the industry promises.

The Whistleblower Who Walked Away from Millions

Jacob Coxon resigned from Anthropic after three years working at both OpenAI and Anthropic. His resignation thread on X has been viewed more than 146 million times. He appeared on Fox News and CNN, and his warning broke through to a mass audience in a way previous AI safety alarms had not.

"I spent the last three years doing pre-training research at both OpenAI and Anthropic. Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives," Coxon wrote. He estimates a 10% chance that superintelligent AI destroys all of humanity.