Matt Clifford: The Bull & Bear Case for China's Ability to Challenge the US' AI Capabilities | E1172

Matt Clifford is the Co-Founder of Entrepreneur First (EF), the leading global talent investor and incubator. EF has incubated startups worth over $10bn, including Cleo, Tractable and Aztec Protocol. Matt is also Chair of ARIA, the UK’s Advanced Research and Invention Agency, and advises the UK government on AI and in 2023 served as the Prime Minister’s Representative for the AI Safety Summit at Bletchley Park.
-----------------------------------------------
Timestamps:
(00:00) Intro
(01:06) Impactful Childhood Hardship
(04:34) The Future of AI & LLM Commoditization
(10:28) Is AI Facing a Hype Cycle Decline?
(15:47) Is China Two Years behind in AI Race?
(24:46) Vertical Integration in AI & Chip Industries
(26:59) Is the World Ready for Autonomous AI Models?
(29:42) Content Automation & Socioeconomic Inequality
(31:19) European Regulation: Overreach or Appropriate Response?
(39:23) The Impact of AI on Future Warfare
(49:19) Is Entrepreneurship Accessible to Everyone?
(53:03) Contrasting Entrepreneurial Ambitions: US vs. Europe
(59:14) Quick-Fire Round
-----------------------------------------------
In Today’s Episode with Matt Clifford We Discuss:
1. The Most Important Questions in AI:
Are we seeing diminishing returns where more compute does not lead to a significant increase in performance?
What is required to reach a new S curve? What do we need to see in GPT 5?
Why does Matt believe that search is one of the biggest opportunities in AI today?
2. The Biggest Opportunities in AI Today:
How does Matt see the future for society with a world of autonomous agents?
What is the single biggest opportunity around agents that no one has solved?
Is society ready for agentic behaviours to replace the core of human labour?
How does warfare change in a world of AI?
Does AI favour states and good actors or criminals and bad actors more favourably when it comes to offence and defence?
3. China and the Race to Win the AI War:
Does Matt believe that China are two years behind the US in terms of AI capability?
What are Matt’s biggest lessons from spending time with the CPP in China working on AI policy?
In what way is the CCP more sophisticated in their thinking on AI than people think?
What is the bull and the bear case for China in the race for AI?
What is the core impact of US export controls on chips for China’s ability to build in AI?
Does a Trump vs a Biden election change the playing field with China?
4. What Makes Truly Great Founders:
Does Matt agree that the best founders always start an entrepreneurial activity when they are young?
What is more important the biggest strength of one of the founders or the combined skills of the founding team?
What did EF believe about founders and founder chemistry that they no longer believe?
Does Matt believe that everyone can be a founder? What are the two core traits required?
-----------------------------------------------
Subscribe on Spotify:
open.spotify.com/show/3j2KMcZ...
Subscribe on Apple Podcasts:
podcasts.apple.com/us/podcast...
Follow Harry Stebbings on Twitter:
/ harrystebbings
Follow Matt Clifford on Twitter:
/ matthewclifford
Follow 20VC on Instagram:
/ 20vchq
Follow 20VC on TikTok:
/ 20vc_tok
Visit our Website:
www.20vc.com
Subscribe to our Newsletter:
www.thetwentyminutevc.com/con...
-----------------------------------------------
#20vc #harrystebbings #mattclifford #entrepreneur #venturecapital #founder #ceo #ai

[@20VC]

Subscribe to the 20VC KZbin channel for more great interviews: www.youtube.com/@20VC

[@benchpress200]

old joke: how far is China behind the US in tech? about 10 hours -- they just wake up the next day and see what's been done

[@joaoavf]

Super thoughtful and clear thinking, amazing to hear it!

[@20VC]

So glad you enjoyed it!

[@pringlecapital350]

Matt mentioned a book but it's not linked in the notes. Which book was he referring to?

[@barriehep]

He’s impressive. (And thank you Matt - one of my best investments is an EF alum ✌️😀✌️)

[@20VC]

He really is impressive!

[@sumanthmurthy1642]

Very interesting conversation! One of the reasons new grads want to work in banking or hedge funds is that the tech/scientist salaries are low in the UK compared to the US. I definitely believe higher salaries will draw more people into tech and I know this well since I live in Cambridge! I regularly meet brilliant people. If we just paid more the talent flow will automatically increase

[@christopheraaron2412]

7:31 there we go one and a half minutes after I posted well then he mentions using the techniques they gave us alpha go and alpha fold.

[@christopheraaron2412]

6:00 The flattening off of adding more compute that's very interesting. So now the real value is going to get the systems to be modified so you get the same output with less compute and less energy and then we can add in there some recursive self-improvement with human supervision similar to what you got with systems that gave you alphago and alpha fold. Those systems gave us narrow superhuman intelligence and we should use that tool for the improvement for one thing when you have improvement in software it's almost immediate whereas if you need improvement and hardware will then you got to order it you got to pay for it you have to bill the data center then you got to hook up someplace where you're not going to be bringing the grid down or you might have to build a nuclear reactor to power it so there is no doubt that the top of that s curve is not that far off.

[@alessiograncini9564]

how come mistral comes from France tho

[@christopheraaron2412]

31:25 well if they try to regulate it then it's all going to move somewhere where they don't.

[@christopheraaron2412]

9:29 wouldn't a good source of data perhaps be the self play between the computer agents that alpha fold and alpha Go used, and also there is Dr Steven Wolfram and his company called Wolfram alpha and digital physics where he uses what we know about physics today and then see if running simulations gives us something similar that mirrors the real universe. One could imagine that the amount of data that could be available by running those simulations and feeding that back into the systems would be basically the limits of the universe.

[@dharmaone77]

EF is not a fantastic deal for founders (you give up a fair amount of equity for not very much in return - I guess you’re forced to do that with spray and pray). Though still a net positive force for creating more startups in the UK. Otherwise this interview was brilliant. Right on the money about AI talent in the UK and our ability to give birth to many more exceptional new companies - if only a few things were tweaked. Matt seems v sharp, comes across very well

[@christopheraaron2412]

18:15 The CCPs obsession with regulation might actually be on inhibiting factor on their ability to get ahead other than of course copying and pasting what we do here in the West.

[@patrickgeyer9239]

1. Buy up all the bitcoin left on exchanges for 1/10th of what the London->Birmingham trainline cost. 2. Back the pound with bitcoin so that every British citizen owns perfectly sound money. 3. Tell the world about why bitcoin is scientific, 21st century money, and watch everyone converge on the schelling point.

[@mrlegoface]

🤡