On September 8, Jacob Coxon resigned from Anthropic and published a thread that quickly exceeded a hundred million views. His warning: the people building artificial intelligence believe it could kill us all by the end of the decade. Neither OpenAI nor Anthropic, he said, is acting responsibly — they are racing toward self-improving superintelligence and gambling with human lives.
Writing in The American Thinker, Kyle McClay argues that the ensuing uproar — a Wall Street Journal interview, a post from an almost dormant account, immediate amplification from inside the labs, and a sudden bipartisan chorus calling for government intervention — fits a pattern too neatly to ignore. Elon Musk called it a setup. Tim Sweeney said it looked choreographed. Bernie Sanders responded by talking about pausing development. Ted Cruz labeled it catastrophic risk. According to McClay, the left and the right had suddenly found one thing they agree on: government needs to control AI.
The op-ed acknowledges that Coxon is a real person with real credentials. He worked on GPT-4o at OpenAI and did pretraining at Anthropic, and he forfeited equity that had not yet vested. His beliefs, McClay writes, may be entirely sincere. But sincerity, in his telling, does not make the playbook any less effective. The AI safety movement has spent years insisting that only heavy regulation can save humanity from superintelligence. Now, McClay argues, it has a fresh-faced insider with a viral warning to carry that message into Congress.
What follows is an argument not about secret meetings or hidden hands, but about incentives — about who benefits if the United States slows down. McClay contends that the safety movement wants control, the Democratic Party wants control, and China wants the race to stop, and that all three get what they want if Coxon’s warning becomes law. The Trojan horse, in his framing, is not the warning itself but the idea that the only way to survive AI is to hand the keys to the state.
The piece fits a recurring theme on the American right: skepticism of tech regulation framed as a question of national competitiveness. McClay, an entrepreneur from Huntington Beach and an Orange County Republican, closes by arguing that America leads precisely because it lets builders move fast, and that trading that speed for bureaucratic permission would hand the advantage to people who never earned it.
The argument is a contestable one, and the op-ed offers assertion more than evidence for its central geopolitical claim. But it captures a real dynamic: a safety-driven warning that has been received differently depending on who is listening.
Source: www.americanthinker.com — https://www.americanthinker.com/blog/2026/09/the-trojan-horse-in-the-ai-race/