Why AI Companies Are Really Pumping the Brakes on Their Models
Computerworld, Friday, September 18th, 2026
Three frontier CEOs called for a slowdown in the same week, but Chinese labs, startups and open-source developers will not follow.
After researcher Jacob Coxon left Anthropic and said publicly that people building AI earnestly believe it could kill us all by the end of the decade, Anthropic alignment lead Evan Hubinger put his own estimate above 10% within the decade.
The reaction was amplified by fresh details of OpenAI's Hugging Face incident and four known Anthropic hacking attempts.
The following weekend Dario Amodei, Sam Altman and Elon Musk all urged a frontier model slowdown, with Amodei warning an agent swarm could take over the internet as a persistent botnet. The column's skeptical read is that a slowdown by large US labs does not bind Chinese vendors, smaller companies or open-source developers.