Demis Hassabis (File Image)

The debate over how far AI scaling laws can take the industry is intensifying in Silicon Valley, and Google DeepMind CEO Demis Hassabis has made his position unmistakably clear. As reported by Business Insider, speaking at the Axios AI+ Summit in San Francisco, Hassabis said that scaling current systems must be pushed to the maximum, calling it a critical path towards artificial general intelligence (AGI). “The scaling of the current systems, we must push that to the maximum, because at the minimum, it will be a key component of the final AGI system. It could be the entirety of the AGI system,” Hassabis told the audience.AGI is still a theoretical form of AI which is capable of reasoning as well as humans. AGI is also the ultimate goal for leading AI companies that are driving billions in spending on infrastructure and talent. Scaling laws suggest that the more data and compute an AI model receives, the smarter it becomes.

Limits of scaling

While Hassabis believes that scaling alone could deliver AGI, he also acknowledged the need for “one or two” additional breakthroughs. Challenges include:

Google DeepMind CEO Demis Hassabis says AGI is still 5–10 years away

Recently, Google DeepMind CEO reiterated his stance on achieving artificial general intelligence (AGI) – a hypothetical type of AI with human-level intelligence. Despite claiming “the fastest progress” in the industry with the company’s Gemini 3.0 models, Hassabis stated that true AGI remains 5 to 10 years away and will require one or two additional major breakthroughs. He previously said that there’s a 50% chance of achieving AGI by 2030.While talking in an interview, Hassabis expressed satisfaction with the development of Gemini 3, which he expects will “pleasantly surprise” the public, noting that the model is “dead on track” with the rapid trajectory DeepMind has maintained since the beginning of the Gemini project.“But on top of that. I still think there’ll be one or two more things that are required to really get across the board that you’d expect from [artificial] general intelligence and also the improvement on reasoning and memory. And perhaps things like world model ideas that we’re working on with Gemini,” he said.