Google AI CEO Demis Hassabis wants everyone to take note on AI: Humans only have a few years left to prepare for …

Demis Hassabis, CEO of Google DeepMind recently issued one of the starkest AI warnings. According to a report by Axios, speaking at the company’s annual developer conference, Hassabis said that humanity is standing in the ‘foothills of the singularity’ and has only a few years left to prepare for artificial general intelligence (AGI). Speaking to Axios after this keynote, Hassabis said that his prediction that AGI could arrive in as little as four years reflects growing confidence that the industry has found the right technical path. Hassabis also noted that the next wave of AI agents should be seen as a societal stress test for far more powerful systems to come. “You can imagine the agentic era in this next year is a little bit like a practice run,” he said. He pointed to Anthropic’s Mythos model as a warning shot, showing how quickly businesses and governments can be caught off guard by accelerating AI capabilities.

Urgency for governments and economists

Google AI CEO also admitted that he deliberately chose provocative language to spark urgency among policymakers, economists and the public. “You’ve got to take this seriously,” Hassabis said. Along with this, he also welcomed tentative steps by the US government towards mandating safety testing before new models are released but stressed that safety efforts must be accelerated. He also revealed ongoing discussions with leaders at other top AI labs about possible safeguards, though he declined to share specifics.

Risks of recursive self-improvement

One looming milestone Hassabis highlighted is recursive self-improvement — systems capable of accelerating their own development. While he said we are not yet at the point where AI systems improve themselves independently, the pace of progress is clearly quickening. “I think what we’re seeing is soft self-improvement, in the sense of these coding agents are making engineers much more productive,” he explained.

Demis Hassabis has a message for laid off engineers

In related news, recently Demis Hassbis revealed that he thinks companies cutting engineers to ride the AI wave have got it backwards. The Google DeepMind CEO told WIRED that when developers become three or four times more productive, the smart move is to do three or four times more work, not march people out the door. To him, treating AI gains as a reason to shrink headcount is a failure of nerve dressed up as strategy.

And he made clear he would happily take the talent everyone else is shedding. "I have a million ideas, from lab drug discovery to game design," he said. "I'd love to have some free engineers to go and do those kinds of things." His pitch arrives in the middle of a brutal stretch for tech workers, with 2026 layoffs already past 142,000 and the cuts still coming.