Skip to content
Founder Atlas

A conversation with Google DeepMind's Demis Hassabis

Key ideas

Hassabis explains how DeepMind funded and ran a 20-year deep tech bet: interim milestones, a London talent advantage, hard early fundraising, and planning for success responsibly.

From a machine transcript of the audio. Quotes link to the moment they are said.

  1. 1

    Break a long bet into milestones

    Hassabis tells founders of deep tech companies that an ambitious 20-year goal has to be broken down. He says choosing the right interim milestones proves progress to backers, the team and yourself, and names Atari games and protein folding as DeepMind's.

  2. 2

    Start when the ingredients line up

    Hassabis says that around 2010 he could see the pieces arriving: early deep learning papers, widely available GPUs and some understanding of the brain's algorithms. He says the early bets on learning and generality held up and that the roadmap was more or less on track. He still guesses that a small handful of breakthroughs beyond scaling are needed for human-level AI.

  3. 3

    Expect early funding to be hard

    Hassabis describes getting the first seed round done as almost impossible, and says he took no salary for the first couple of years. He says UK investors had little understanding of or appetite for so long a payoff. In the US, he recalls, the backers were mostly individual entrepreneurs who understood such a journey, not VCs.

  4. 4

    Build where talent lacks options

    Hassabis says he stayed in London partly because strong UK and European graduates who did not want finance had few other options. He says this gave DeepMind an advantage for its first five to seven years, until multinationals and VCs caught on.

  5. 5

    Plan for what success would mean

    Hassabis says that because they believed AGI was possible within 20 years, they treated ethics and safety as necessary from the start. Under high uncertainty he favors cautious optimism and what he calls being bold and responsible over moving fast and breaking things. He acknowledges this gets hard in a commercial frenzy.

In Demis’s words

“You have this sort of enormous 20-year ambitious goal, but you've got to be able to break that down.”
Demis Hassabis·A conversation with Google DeepMind's Demis Hassabis·Watch at 18:55(opens the original in a new tab)
“We were just trying to scrape together, frankly, you know, our first seed round. It was almost impossible. You know, I didn't pay myself a salary for the first couple of years, things like that.”
Demis Hassabis·A conversation with Google DeepMind's Demis Hassabis·Watch at 2:57(opens the original in a new tab)
“So I felt there was a sort of competitive advantage by being here and actually making the most of that talent pool almost unopposed in the UK.”
Demis Hassabis·A conversation with Google DeepMind's Demis Hassabis·Watch at 23:05(opens the original in a new tab)

About this entry

Published
Jun 7, 2024
Channel
Stripe
Type
Video
Length
37:47
Language
English
Added
Nov 2, 2026
Source last checked
Nov 2, 2026

What’s available here

  • Original source

    Linked. Opens on YouTube.

  • Player on this page

    Available. YouTube’s own player, loaded only when you press play.

  • Full text or transcript

    Not reviewed. No license to show a text version here.

  • Our summary

    Draft. Written by us with AI from a machine transcript of the video’s audio (Whisper) and checked against it. Machine transcripts can mishear; each quote links to the moment so you can hear it. Not yet rights-reviewed.

  • Ask questions with AI

    Not reviewed. Off until there is licensed text to work from.