The resignation this week of Anthropic researcher Jacob Coxon over concerns that artificial intelligence labs are “gambling with our lives” lit up the long-smoldering argument about whether AI is heading in a direction hazardous to human life.
Not only is he not alone at Anthropic — whose alignment science lead agreed with the possibility that AI has a more than 10% chance to “kill all humans” in the next decade — but OpenAI chimed in, as its chief scientist called for an AI research slowdown. And CEO Sam Altman reportedly told staff that the company is considering slowing down leading-edge AI work and wants other companies to do the same — if the government says that’s not an antitrust problem.
We’ll have to see if any of this existential worry actually changes anything, but I’ll believe it when I see it. Just this week, as almost every week, there was a flurry of new models, agents and apps galore — from Meta, OpenAI (twice), China’s DeepSeek and more. I understand why, and I’m still a bit skeptical of the AI doomers given the lack of specific reasons why AI can’t be controlled, but there’s definitely a disconnect here.
One thing that may slow things down regardless is the rising opposition to data centers. In a new feature story, Paul Gillin digs into exactly why AI requires so much compute — and why there isn’t a short-term fix. Tech giants have a lot of work ahead of them persuading communities not to shut down the next era of tech innovation.
