scripod.com
The Future of Intelligence with Demis Hassabis (Co-founder and CEO of DeepMind)

Overview

Shownote

Highlights

Transcript

Chapters

Pins

The Future of Intelligence with Demis Hassabis (Co-founder and CEO of DeepMind)

Google DeepMind: The Podcast

2025/12/16
Google DeepMind: The Podcast

Google DeepMind: The Podcast

2025/12/16
OverviewShownoteHighlightsTranscriptChaptersPins

Shownote

Thanks for joining us this year.  Subscribe to stay updated on our return in 2026, and revisit our episode library to catch up on everything from driverless cars to drug discovery: https://www.youtube.com/playlist?list=PLqYmG7hTraZBiUr6_Qf8YTS2Oqy3OGZEj ...

Highlights

In this reflective and forward-looking episode, Demis Hassabis of DeepMind shares insights on AI’s evolving trajectory—from narrow breakthroughs to the pursuit of trustworthy, reasoning-rich systems capable of tackling humanity’s deepest scientific and societal challenges.
02:39
AI excels at Olympiad-level math yet falters on basic consistency, revealing fundamental gaps in reasoning
12:39
Adding a confidence score like AlphaFold could solve hallucinations in Gemini 3.0
17:39
Plugging Gemini into Simmer creates an interesting interaction between two AIs, potentially leading to an infinite training loop
28:18
Gemini 3 has a scientific personality that pushes back on non-sensical ideas
38:31
The central question is about the limits of Turing machines, and the speaker is uncertain whether the human mind involves only classical computation.
41:13
Reality is a mental construct, and all sensory experiences are information
46:14
Being human involves tool-making and scientific curiosity, which led me to build AI

Chapters

How does AI move beyond chat to become a true scientific partner?
00:00
Can an AI assistant be both brilliant and humble about its limits?
10:12
Why do virtual worlds matter for teaching AI to understand reality?
15:11
Is AI overhyped today—and underestimated tomorrow?
25:50
What happens to democracy and economics when scarcity ends?
35:50
Is the human mind just a very complex computer?
41:13
How do we build transformative tools without losing control?
46:14

Transcript

Demis Hassabis: I'm basically doing everything I ever dreamed of. And we're at the absolute frontier of science in so many ways, applied science as well as machine learning. And that's exhilarating, that feeling of being at the frontier and discovering som...