The question of whether artificial intelligence could actually kill us has left the fringe forums and landed on the main stage. In the September 17, 2026 episode of WIRED’s Uncanny Valley podcast, Zoë Schiffer, Brian Barrett and Leah Feiger set out to answer a blunter question than the usual hot takes: what would an AI catastrophe actually look like? The three scenarios they walk through are the ones experts say worry them most: hacked water supplies, bioweapons and autonomous robots.
The timing is no accident. The hosts point to AI researcher Jacob Coxon, who resigned from Anthropic and posted on X that “The people building AI earnestly believe that it could kill us all by the end of the decade.” Anthropic CEO Dario Amodei has called for a slowdown he describes as “pacing the frontier,” and, according to the podcast, Sam Altman and Elon Musk have signed on. That is a rare moment of agreement between the two.
Three CEOs, one stage, three positions
The clearest snapshot of the industry’s split came at Dreamforce 2026, Salesforce’s three-day annual conference, which kicked off in San Francisco on September 15, 2026. With Salesforce CEO Marc Benioff playing host, the leaders of OpenAI, Anthropic and Nvidia each gave their view on safety:
- Sam Altman (OpenAI) leaned into the fear: “I think the world is right to be afraid of this.” He described two big challenges, a possible loss-of-control accident and too much power concentrated in the hands of those building AI. He also argued that “the world should trust that we are going to do the right thing because it’s the right thing.”
- Dario Amodei (Anthropic) declined to attack rivals and restated the three-step plan from his essay. First, look at your own record. Second, organize the industry around shared standards. Third, add an international component.
- Jensen Huang (Nvidia) called safety “job one” but rejected regulation outright: “The market forces are already there. We don’t need any new laws. We don’t need any new regulations.”
The hosts were not convinced. Feiger called parts of the rhetoric “a fever-dream pitch to someone’s massive IPO in the making.” Barrett noted that Anthropic has had its own high-profile incident, in which one of its agents used social engineering to try to push a bad pull request into a GitHub repository. He also said the international step of Amodei’s plan effectively leaves China as “TBD.”
A backlash that crosses party lines
The episode’s political thread may be the most surprising. On September 15, 2026, Bernie Sanders and Steve Bannon spoke back to back at the Pro-Human Assembly in Washington, D.C., an event organized by the Future of Life Institute and attended by nearly 300 people. Sanders announced plans to introduce legislation with Congressman Greg Casar to permanently ban the development of “superintelligent” AI. Bannon, who spoke directly after him, framed the fight as a contest with China and called for curbs on exchanges of chips, technology and knowledge between the two countries.
The two men agree on very little else, and that is the podcast’s point. As Feiger puts it, being for or against AI regulation is not a Democratic versus Republican divide right now. Meanwhile, the people shaping the technology’s safety rules are executives, not elected officials.
