Big Technology Podcast · Alex Kantrowitz

Best of Big Technology: Demis Hassabis On AGI, Deceptive AIs, Building a Virtual Cell

·58 min·5 clips

The show opens with a clear setup: Demis Hassabis joins from Google DeepMind to talk about AGI, Google's AI roadmap, and the way AI research is pushing scientific discovery. The interview starts with the blunt question everyone keeps circling: where are we on the path to human-level intelligence, and how far off is it. Hassabis does not try to make the answer neat. He says the field still needs systems with richer world models and more capable agent-based behavior. He also argues that AGI should be treated with scientific integrity, not bent to fit commercial incentives or hype. The host presses on the word itself and the conversation shifts to what people actually mean when they say intelligence. Hassabis says the best humans are still astonishing in every domain, from sports to music to games, and that AI work has not changed that view. Instead, he frames human cognition as remarkably adaptable, both individually and as a society. The interview keeps coming back to what current systems can and cannot do. Hassabis suggests the next wave will be built around agent-based systems rather than simple pattern matching. He draws a line between useful near-term products and the deeper scientific claim behind AGI. That distinction matters to him. The conversation also touches the idea that current models may still need a more developed world model for deeper reasoning. The host keeps pushing for plain English, and Hassabis keeps translating the jargon back into something more concrete. That gives the discussion a useful edge. Rather than treating AI progress as a blur of breakthrough language, the episode keeps asking what changes in practice. It points toward systems that can reason, act, and help with scientific discovery, not just generate outputs. The back half of the interview opens wider into philosophy and the human condition. Hassabis says society will need serious thinkers to help navigate what comes next. He even gestures toward the need for philosophers with the kind of stature people once associated with Kant, Wittgenstein, or Aristotle. The episode ends on a long-horizon view of what AGI and artificial superintelligence could mean for civilization. It is ambitious, but the show keeps the framing grounded in the actual research conversation rather than letting it drift into pure futurism.

As heard by us

A brisk AGI conversation that keeps turning grand claims into concrete questions.

This episode tracks Demis Hassabis on AGI timelines, what current systems still miss, and why the bar should stay scientific instead of bending to commercial convenience.

Read the full review in PlayNext →

Why you'd press play

You want the AGI argument without the usual lab-slogan fog.

Read the full recommendation in PlayNext →

Listen to the show on

Apple Podcasts Spotify

Play clips from this episode More from Big Technology Podcast