scripod.com
20VC: DeepMind's Demis Hassabis on Why AGI is Bigger than the Industrial Revolution | Why LLMs Will Not Commoditise & We Have Not Hit Scaling Laws | Bottlenecks in AI & The Energy Crisis Caused By AI | Whether AI Will Do More to Harm or Help Inequality

Overview

Shownote

Highlights

Transcript

Chapters

Pins

20VC: DeepMind's Demis Hassabis on Why AGI is Bigger than the Industrial Revolution | Why LLMs Will Not Commoditise & We Have Not Hit Scaling Laws | Bottlenecks in AI & The Energy Crisis Caused By AI | Whether AI Will Do More to Harm or Help Inequality

The Twenty Minute VC (20VC): Venture Capital | Startup Funding | The Pitch

Apr 07
The Twenty Minute VC (20VC): Venture Capital | Startup Funding | The Pitch

The Twenty Minute VC (20VC): Venture Capital | Startup Funding | The Pitch

Apr 07
OverviewShownoteHighlightsTranscriptChaptersPins

Shownote

Demis Hassabis is the Co-Founder & CEO of Google DeepMind - working on AGI, responsible for AI breakthroughs such as AlphaGo, the first program to beat the world champion at the game of Go; and AlphaFold, which cracked the 50-year grand challenge of protei...

Highlights

In this insightful conversation, Demis Hassabis—co-founder and CEO of Google DeepMind—offers a candid, forward-looking perspective on the state and trajectory of artificial intelligence, grounded in decades of pioneering research and real-world impact.
00:00
90% of modern AI industry breakthroughs were by Google Brain, Google Research, or DeepMind
04:05
The human brain is the only known proof of general intelligence
05:01
There's a good chance of AGI emerging within the next five years
06:00
Compute is the biggest bottleneck for scaling AI systems and conducting experiments
07:07
Frontier labs are getting good returns from compute expansion, and the returns are still substantial, though less than at the start of scaling.
07:36
Continuous learning is still lacking in current AI systems
08:30
DeepMind's rapid progress stems from organizational changes, combining talent and resources, and a startup-like focus
11:06
General intelligence shouldn't have 'jagged' performance
12:02
Labs with the ability to invent new algorithmic ideas will have an advantage in the next few years as existing ideas are exhausted
13:03
Foundation models won't be replaced but built upon—they're successful and here to stay.
14:47
The real revolution will come when several AI drugs pass the whole process, allowing regulators to trust model predictions and potentially skip animal testing
17:01
The right regulation could set minimum standards for leading AI providers
18:01
AI systems must be certified to prevent deception and ensure authenticity of information
19:34
Preventing AI from outputting non-human-readable tokens is a key safeguard proposed by the international AI Safety Agency
21:05
Independent testing and auditing of powerful AI systems is essential to build public confidence
22:06
AGI will be 10 times the industrial revolution in scale and speed
23:06
We question if we overestimate short-term and underestimate long-term AI progress, or if it's coming faster than expected
23:30
AI is currently overhyped in the short term but underappreciated in the long term (about 10 years)
24:30
AI will pay for itself by optimizing infrastructure, modeling climate and weather, and enabling new technologies like fusion
26:09
Fusion energy could help with climate, environment, and space exploration
28:02
Unlocking pension fund investments is critical to scale European tech companies beyond the startup phase
29:33
They hit it off due to shared ambition and love for sci-fi
31:00
Building a general-purpose drug design platform at Isomorphic applicable to any therapeutic area
31:32
The interviewee's desired legacy is advancing science and curing diseases

Chapters

AI Breakthroughs Are Dominated by Google Labs
00:00
What Actually Counts as AGI; and Where Are We Today?
04:00
What Are the Biggest Bottlenecks Holding AI Back Today?
05:00
Have We Hit the Limits of Scaling Laws?
06:00
Where Is AI Ahead of Expectations; and What's Still Missing?
07:00
Why Can't AI Systems Learn Continuously Like Humans?
07:30
How Did DeepMind Go from Behind to Leading the Pack?
08:30
Are We Heading Toward Model Commoditization; or Winner-Takes-All?
11:00
What Does the Future of Open Source Really Look Like?
12:00
What Does a Post LLM World Look Like?
13:00
Can AI Really Fix Drug Discovery—and Cut the 10-Year Timeline?
14:45
What Does "Good" AI Regulation Actually Look Like?
17:00
Who Should Be the Ultimate Arbiter of Truth in an AI World?
18:00
If Demis Had One Shot to Fix AI Safety, What Would He Do?
19:30
Is This Time Different for Jobs; or Will History Repeat Itself?
21:00
Is AGI Bigger Than the Industrial Revolution; and Faster?
22:00
Are We Underestimating AI Despite All the Hype?
23:00
Does AI Lead to Massive Inequality; or Universal Prosperity?
23:30
How Do We Solve the Energy Crisis Created by AI?
24:30
Why Stay in the UK Instead of Moving to Silicon Valley?
26:00
Will Europe Ever Build a Trillion-Dollar Tech Giant?
28:00
Meeting Elon Musk for the First Time?
29:30
What Big Questions About AI Is No One Talking About?
31:00
What Does Demis Want His Legacy to Be?
31:30

Transcript

Demis Hassabis: The returns are still very substantial, although they're a bit less than they were obviously at the start of all of this scaling. I would say about 90% of the breakthroughs that underpin the modern AI industry were done either by Google Bra...