The loudest voices warning that artificial intelligence could end human civilization are coming from the very companies best positioned to profit from tighter rules on the technology — and that coincidence is starting to draw sharp scrutiny. In a Daily Wire column, Ben Shapiro argues that the industry’s doomsday rhetoric may function less as a public service and more as a competitive weapon.
The piece opens with a blunt thesis: “We are all being played.” Shapiro frames the central question as a choice between two readings of the AI industry’s alarmism. Are the founders pushing for regulation genuinely terrified of what they’re building? Or are they “magnifying public angst about AI in order to achieve regulatory capture” — rewriting the rules of the road to box out new competitors, enable collusion, and lock in monopolistic control?
Shapiro doesn’t leave much suspense about where he lands. “I might be a little cynical,” he writes. “I think the latter is more likely.”
The Coxon Flashpoint
The debate burst into public view, according to the column, after Jacob Coxon — a researcher who spent time at both Anthropic and OpenAI — warned online that many people in the field believe AI will kill everyone. He then took that argument to the BBC, saying there’s a strong chance humanity could die in the near future.
Shapiro compares the framing to past climate apocalypse predictions — “All the ice caps are going to be gone! The polar bears are all going to drown!” — and faults Coxon for never explaining the mechanism. “He hasn’t explained in detail, or even in relative detail, how we get from ‘AI is outracing us’ to ‘We all die,'” Shapiro writes, noting that if the risks can be gamed out, they can presumably be constrained.
He also flags Coxon’s background: the researcher worked for Newspeak in 2021, described as a left-wing British NGO focused on “democracy tech projects,” “collective transformation,” prison abolition and extinction rebellion. Coxon lasted only six weeks at Anthropic, per the column. Nvidia’s Jensen Huang called his claims outlandish, deeply untrue and arrogant, while Elon Musk labeled him “sus.”
A Horse Race, Frozen in Place
Shapiro’s core argument rests on the structure of the AI industry itself, which he splits into two camps with opposing incentives. On one side are hardware players like Nvidia and chipmakers such as AMD, who make the infrastructure AI runs on and have every reason to want development to accelerate. “If there’s an AI pause, Nvidia is in serious trouble,” he writes.
On the other side are software developers — OpenAI with ChatGPT, Anthropic with Claude, Meta with Llama, Google with Gemini — who build models in two flavors: closed-source and open-source.
The distinction matters, because the business models diverge sharply. Closed-source, proprietary AI keeps the inner design hidden; users subscribe, feed in data, and pay per token. Open-weight models publish the core mathematical files online for free — not the full formula, but the general ingredients — allowing anyone to modify and build on top of them.
Shapiro illustrates the difference with a restaurant metaphor. Open source is the chef handing over the cookbook, the ingredient list, and a cooking video so you can adjust the recipe yourself. Open weight is the chef making you the meal and letting you take it home to add your own ingredients — Meta’s approach. Closed source is ordering off the menu with no substitutions: the chef decides how it’s made, you can’t take it home, and you have to eat it inside a protected area of the restaurant.
Under that framing, the closed-source labs face an existential fear: that an open-weight competitor produces a better, portable, modifiable product and renders the walled-garden restaurant irrelevant. “You’re dead in the water,” Shapiro writes.
That dynamic, he argues, explains the pattern of who sounds the loudest alarms. “The closed-source models — Microsoft, Google, Anthropic, OpenAI — tend to be the ones arguing the most that AI is going to kill all of us.” If the public can be convinced that the takeout restaurant might serve poisoned food, the argument goes, the closed restaurant survives — not because its food is better, but because the alternative has been regulated out of reach.
Apocalypse Camp vs. Stagnation Camp
Shapiro divides the broader AI debate into two schools of thought. The “apocalypse camp” warns that a black-swan event — superintelligence going rogue, launching nuclear missiles, ending humanity — is the paramount danger, and that government restriction is the only safeguard. The “stagnation camp” counters that the real threat is falling behind: if development stalls, China wins, the economy suffers, and productivity declines.
Shapiro acknowledges it’s possible to hold both views simultaneously — to believe restrictions invite stagnation while also conceding that unrestricted development leaves room for catastrophe. But he’s skeptical of the motives behind the apocalypse camp, suggesting its members may want to restrict their opposition, and may also want to limit their own legal exposure. If the government regulates an industry, companies can claim they bear no liability because they checked the government’s boxes.
Anthropic, which Shapiro describes as leaning left, is leading that camp. He points to founder Dario Amodei, who compared President Trump to a feudal warlord when explaining his 2024 vote for Kamala Harris, and who went on CNN to say he largely agreed with Coxon’s warnings.
That, for Shapiro, is the tell. “If somebody came out from my company and said, ‘Ben Shapiro’s company is about to end the world and kill all of humanity,’ I’d be angry,” he writes. Instead, Amodei seemed “somewhat happy” about the assertion that his own company might be responsible for an apocalypse. “I find that suspicious,” Shapiro concludes. “Call me a cynic.”
The column lands on a broader warning about extraordinary claims: predictions of catastrophe, he notes, have an extremely poor track record — “a lot of people out there have predicted the last 25 recessions.” An extinction-level claim, in his view, demands extraordinary evidence. What the public is getting instead, he argues, is a set of warnings that conveniently align with the commercial interests of the companies issuing them.
Source: www.dailywire.com — https://www.dailywire.com/news/the-doomsday-warning-thats-starting-to-look-awfully-convenient