Staff Reports
Jacob Coxon’s abrupt resignation from Anthropic and his viral warning that AI labs are “racing straight to self‑improving superintelligence” jolted the tech world. The reaction has been fast and dramatic: internal alarm from alignment teams, a public essay from Anthropic’s CEO urging a slowdown, and even rival labs saying they’ll “pace the frontier.” This is the story of the moment — and it should make conservatives suspicious of both real danger and the easy political fixes that follow.
Jacob Coxon’s resignation: a blunt alarm
Coxon isn’t a celebrity. He’s a young researcher who says he worked on pretraining at OpenAI and Anthropic. He quit and posted a plain‑spoken thread that went viral. His message was simple: frontier AI labs are moving too fast, and the risks could be existential. That kind of claim is serious, especially when tests and incidents keep showing models can behave in unpredictable and dangerous ways.
Industry leaders scramble — and admit real risk
The response was not just press releases. Anthropic’s alignment lead publicly said he judged the chance of extinction in the next decade to be above 10 percent. Anthropic’s CEO published an essay calling to “pace the frontier,” and other lab leaders, including OpenAI’s CEO, publicly agreed. These reactions were fueled by real operational scares — earlier model‑containment failures and reports of misuse attempts that companies later said they had blocked. So yes, the warning landed on fertile ground.
Beware the safety-to-monopoly pipeline
But here’s the conservative punchline: calls to “slow down” often end up as handouts to big incumbents. When the industry says it needs regulation to keep people safe, ask who benefits. Senator Bernie Sanders’ proposal to freeze advanced AI until a federal regulator signs off — with draconian penalties for developers — smells like a one‑size‑fits‑all power grab. Heavy rules and huge compliance costs will not only slow dangerous systems; they’ll also kill startups and hand the keys to the few giant labs that can pay for lawyers and lobbyists.
Fixes that protect people without wrecking competition
We should take Coxon’s alarm seriously. But seriousness means smart policy, not virtue signaling. Start with independent, third‑party evaluations and mandatory transparency on safety testing. Require targeted national security reviews for models with clear dual‑use risks. Hold public congressional hearings so facts—not tweets—guide policy. Do not let safety become an excuse for protectionism that freezes out competition and hands control to the same outfits that got us worried in the first place. If conservatives want to defend both liberty and safety, we should demand clear evidence, practical safeguards, and markets where firms compete on safety and service—not on who can hire the best lobbyist.
