Google DeepMind’s Demis Hassabis & The Paradox of AI Progress - Semafor Tech

Abstract

It’s not just AI safety advocates and data-center opponents that threaten to curb AI’s momentum. The industry is running into limits on how fast it can actually accelerate, Google DeepMind founder Demis Hassabis said in an interview with Semafor.

Shortages of critical components like high-bandwidth memory and a pullback in open research have constrained AI’s ability to scale quickly — and that might provide natural guardrails for the evolving technology.

“Look, it may be a good thing that it’s not as fast,” Hassabis said. “There’s a whole bunch of other things that we need to think through with this technology,” from commercial issues to philosophical ones. “We don’t have a lot of time to sort out before we get to [Artificial General Intelligence].”

Hassabis spoke to Semafor’s Reed Albergotti ahead of his visit to the World Economic Forum in Davos, where he planned to help business and government leaders prepare for AI’s impact, and get a clearer sense of how the international community is thinking about the technology.