Anthropic, the so called responsible AI shop just had another insider walk out the door, and he didn’t mince words. Jacob Coxon, a 27 year old pretraining researcher who spent three years inside both OpenAI and Anthropic, resigned last week and told the public what the labs only mutter amongst themselves. Revealing that they are racing straight toward self improving superintelligence and are gambling with the future of human superiority. These superhuman systems can hack anything, completely remake entire industries in a single night, and seize real power and resources. The people building it earnestly believe it could kill us all by the end of the decade. That is the bleak confession Is the hype a political and marketing ploy as the billionaire Tech bros are scrambling to narrate? They do raise some good points. A Wall Street Journal article about him appears to have been published roughly 18 minutes earlier, and Peter Wildeford was already commenting on that WSJ piece nearly two minutes before Coxon’s own post. Nathan Calvin then amplified Coxon about six minutes after publication and Daniel Kokotajlo about ten minutes after. Coxon left before his Anthropic equity vested but still held OpenAI equity. His criticism was more qualified than some summaries claimed, describing Anthropic as more responsible than OpenAI and not yet cutting corners while warning that competitive pressure could change that. His claim that people inside frontier labs genuinely fear catastrophic outcomes received public corroboration from Anthropic’s Evan Hubinger, and although the Sanders-Casar Ban Artificial Superintelligence Act had already been announced five days earlier, Coxon’s post was quickly cited in subsequent Senate outreach. Anthropic’s own alignment science lead, Evan Hubinger, didn’t even bother to deny it. He replied that Coxon is correct, they really do believe AI could kill all humans, and he personally puts the chance above 10 percent within the next decade. “Alignment” is the name for the unsolved problem of making sure those goals stay compatible with human survival and values. Ai is racing toward systems that are much more capable than humans, pursuing goals we didn’t fully specify, in a world where we depend on software, infrastructure, and automation. Simply put, humans will be forced to stop making the decisions after a super intelligent annihilating machine was created with no way to stop it. Another countdown to extinction we are just now beginning to grapple with. Hubinger also said Anthropic is doing its best, but it still doesn’t have a plan for keeping future super smart AI safe and under control and it’s not clear they’re on a path to have one. This comes months after another Anthropic safety lead, Mrinank Sharma quit. Warning that the world is in peril from AI, bioweapons, and a whole series of interconnected crises. But what are the tech bros REALLY worried about? President Trump just slammed the brakes on the latest globalist psyop dressed up as “safety,” rejecting the sudden chorus from Silicon Valley’s AI high priests. Dario Amodei, Sam Altman, and even Elon Musk, who now want to “pace the frontier” after years of racing toward the singularity like it was a gold rush. From the Irish Open he made it plain. America is leading China, whoever wins AI wins, and the only guardrail this country needs is a strong, high IQ president who already has the criminal and regulatory power to stop the bad actors pretending to be angels. Meanwhile Barack Obama, the same architect of the surveillance state and the kill switch culture, is privately whispering to Democrat leaders that they must seize AI regulation as their midterm weapon, warning that the technology is “moving very fast in private hands” and could be “dangerous” unless the party clamps down on safety, kids, and the jobs the models will wipe out classic elite language for more centralized control. This isn’t about protecting humanity; it’s the same cartel playbook we’ve seen with pharmaceuticals, finance, and social media—incumbents begging Washington for embedded evaluators, antitrust waivers, and talks with Beijing so they can lock in their monopoly while China goes full-send. Trump sees the sick conspiracy against data centers and American energy for what it is: an attempt to kneecap the United States so the New World Order can dictate the terms of the next industrial revolution. Obama and the Democrats want a public conversation that ends with bureaucrats and globalist NGOs writing the rules; Trump is telling the conspiracy theorists, treasonists, and leakers to beware because the golden goose isn’t getting slaughtered on his watch. The race is on, the gloves are off, and the only pause the American people should accept is the one that keeps us first. The race is the point. Coxon said Anthropic understands the stakes but is locked in a contest to get there first because it believes no one else will act responsibly, so it must do it itself despite the risk. Accepting that race and calling it the “endgame” from a private company’s Slack is a hubristic gamble. Attempting to speed run alignment without extraordinary confidence that no better path exists is exactly the kind of elite decision making that never asks the public for permission. The same labs that warn about extinction keep pouring billions into the next model. The public is expected to trust the people who just admitted they do not have a solution. That is the warning they keep delivering on their way out the door. This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit bownereport.substack.com/subscribe