Loading summary
Transcript167 lines
- [00:00]
A
Listeners, we here on the Anglo Futurism podcast love self sufficiency and sovereignty. However, we've been doing an audit. We found that our list of dependencies is growing longer and longer.
- [00:11]
B
Not only does it include war mail and beef sticks, it also, as of recently, includes this very studio, which Harriet Green from Bases Capital has done up for us on the King Charles III space station. Now we want one more dependency which might come from you, the deep pocketed future sponsor of the show. So get in touch and enjoy the show.
- [00:32]
A
Who now has anything to say about
- [00:34]
C
the deindustrialization of this country?
- [01:03]
B
Georgian townhouses on the moon, the highest
- [01:06]
A
GDP per capita in the Milky Way,
- [01:08]
B
small modular reactors under every village.
- [01:10]
A
Green, this is Anglo Futur.
- [01:32]
B
Welcome back to the Anglo Futurism Podcast. I'm Tom O.
- [01:35]
A
And I'm Callum Drysdale.
- [01:37]
B
And this is James Wise, our guest, the chair of Sovereign AI. James, welcome to the show.
- [01:41]
C
Thanks for having me.
- [01:42]
B
Well, we're going to talk today about how James became the chair of Sovereign AI, what else James has done in his career, what the UK government is doing to try and make sense of this incredible moment in scientific history and a few other things.
- [01:55]
A
Well, let's get started. James, you have taken up this sort of tour of duty. You've gone into the government to lead the Sovereign AI unit, the unit that is going to try and establish a kind of place for Britain in this evolving and rapidly changing AI world. How scared are people internally? Like, are people freaking out sufficiently, do you think, inside the government?
- [02:21]
C
I think the good news is most people are freaking out less than me. And you know, what you want when you go into an organization is to see that sort of British calm approach to what is, as you say, like this transformative moment in scientific history, in our national security, in our national economy, in our culture. And so I day to day work in the venture capital world, which can be known sometimes for its bullishness and exuberance. Working with the civil service has been a much more tame affair in that regard. So that's the good news. I think there's areas where the government are taking things really seriously, and rightly so. We could talk about those. And there's plenty of areas where we need to be doing a lot more. And I've been so involved and enthralled with what's been happening on the venture side and in the British technology industry as a whole over the last decade. I felt it was really important for more people, not just me, to try and help the government navigate some of this Challenge and opportunity as well.
- [03:24]
A
And okay, but you know, this I think is we would say as a is it is going to change everything, right? I mean, you know, we on this podcast are fairly bullish on AI progress and we think that, you know, the line will keep going up where, you know, what is the attitude internally has that sort of, that level of, you know, the sufficient level of real, like, Christ, everything is about to change verbatim
- [03:54]
B
quotes from senior ministers would be very well,
- [03:58]
C
you know, I think internally they've set up this sovereign AI unit which I now chair as a nod to. To how important is us becoming at a national security level and like for a national domestic economic level. They had a big opportunities plan set up last year and announced Matt Clifford, which signified in lots of areas that we're going to invest in. But at the same time they're balancing that with a bunch of other pressures on the uk And I think one of the challenges is syncing up the part of the government that's saying AI is really important to. It's important for national security. So the agencies are working on stuff, it's important for our economy, so Soviet is working on stuff and it's important for our general talent base, everything else. And so we're doing smart stuff. They're building stuff like AC for capability. But syncing that up with, as you say, everything is going to change in health, everything is going to change in education, everything's going to change potentially in areas like transport. Syncing those two things up is a much bigger task.
- [05:00]
A
Right.
- [05:00]
C
It's not just one silo of government like DCIT gets it, you need everyone to get it. And I think just as technology takes time to diffuse across an economy, these ideas take time to diffuse across a very large organization like Whitehall. But it is happening.
- [05:14]
B
When you talk about syncing up between different parts of governments and I wonder, how much guidance are you getting from different departments? Are the security services saying, we've got to keep a bit of that company, for instance.
- [05:25]
C
So there's very many different people who provide input into some AI strategy. Oh, I bet some of those people work in those rooms. But we have, you know, the Prime Minister's advisor, Jade Leung, she sits on our board. We've got the civil servant who's also responsible for nsif, which is the national, which is the investment fund that invests in defense technologies. You know, obviously they'll be overlapped. She sits on our board as well. And so we get input and, you know, get input, as I've already mentioned, for People like AC on cybersecurity. So that's happening. It's not yet. I mean, we're two or three months in, but it's clearly the thing that we need to do, not just with agencies, but we need to make sure that when we're looking at a company that could be of absolutely critical importance to health care, for example, that we get senior people from the health department involved. That piece, getting joined up, government around, that is still something we're working on. But actually everyone I've spoken to knows this is the right direction of travel. So look, we have to get this right across government and then there's what levers can you pull internally and how quickly, even if you pull the lever, can you get people to move? And that's one of the big differences between working in the private sector and working in the public sector or working for small entrepreneurial organizations, which what I do in my day job, like startups and working for big organizations like government is how fast can you make those things change? Direction of travel is good, but it's not just direction, it's a vector. We probably need to increase the speed.
- [06:54]
B
I think we should talk about the concept of sovereignty more broadly. So of course yours is the sovereign AI unit and the way I see it is that you're a bit of a VC fund and the aim is to keep chunks of these companies in this country in DeepMind, example, perhaps are the one that got away and you have various levers available to you that VC firms don't. So you can fast track visas. You can probably help companies embed with British government work more quickly. Is that a kind of complete picture of what's going on?
- [07:25]
C
Yeah, sovereignty is a very broad term. Right. And I think what we're trying to do is expand sovereign capabilities when it comes to AI across all the levels of sovereignty. Some of that is just resilience. So, you know, it's important that we don't entirely rely on one company or one country for that matter, for providing all the different parts of the AI stack. There's lots of different parts that I stack. There's chips, there's energy, there's data centers, there's architecture, there's data itself, there's the models and there's applications. Right. You've got all those pieces and you can't just have one country or company providing you all of that stuff. And so part of our job is to make sure we build resilience, that we have lots of different providers at each part of the stack. That's the lowest level of sovereignty, but it does matter. People kind of overlook this. Like, we can't have someone turn off our chips overnight. And that's really important. Doing that doesn't mean you can only have British chips, but it does mean you have to have three or four providers to make sure you have resilience. The next level is economic sovereignty. And actually this is where I push back a little bit on people talking about DeepMind. Because economic sovereignty means that you benefit, you as a citizen and Britain as a whole benefits from the success of these technologies. And you benefit in different ways, right? You benefit from having this talent here because they can go on and build other companies, but you benefit from the income taxes they pay. We benefit from the capital gains taxes that investors in them pay. We can benefit from all the other services they buy locally. There's all these other economic benefits. And actually, in DeepMind's case, yes, I think it's probably a shame that we ended up seeing the greatest scientific mind possibly of our generation in sadness. And all the people he recruited ultimately contribute to, to Alphabet's top line. However, he managed to keep a vast majority of the team here. He managed to use billions of dollars of Alphabet's money to train and build that team, which was just completely unavailable to him in the UK at that time. And so had he not made that decision, perhaps DeepMind would have been nowhere and another country. Maybe it would have been a team out of Paris or a team out of New York or a team out of Singapore that would have ended up doing all of that. So actually, the economic sovereignty we got from DeepMind making that decision was quite strong. Britain has benefited a lot from not the full way because, you know, the profits from DeepMind will go to Alphabet and very few investors manage to get into DeepMind to benefit from that. But a lot of income taxes are great. And then there's a final level of sovereignty, which is your British full stack. Put the flag on top of the building, you know, legal, regulatory, alignment with the British state. And we need more companies in that bucket. That really matters as well. So sovereign AI is about expanding the number of companies we have in each of the levels of that stack, right? Resilience, economic and legal sovereignty. We are set up and run like a venture fund, right? So we have pot of money, we're investing it. We're not a grant organization. Most of the government's really good actually at grant and forget, right? Give someone some money, walk away, we invest, the taxpayer gets upside. If it works, we lose money. If it doesn't work. But that means we're involved, we have information rights. Sometimes we'll sit on the board, but we're there for the long run. And hopefully I think more than just the capital return, which obviously it'd be great to make money for the taxpayer, but in the grand scale of things, there won't be a lot of money more than the capital return. By being investors and not grant makers, you're really forced to think differently as a civil servant or as the kind of experts I'm bringing in for the fund who are venture capitalists. You're forced to think about the long term commercial viability of this company because why would you invest in a company that's going to die in a year's time or two? And I think that mindset shift is different. It's not just about investing some money and hoping to get a return. It's adding a sort of commercial prism to the work that government has done historically around supporting industries as a whole. Because I think AI needs that kind of effort. So we've got a bit of money, but it's not really about the money. It's about everything else we're doing to make the UK a good place to build AI companies and also sharpening our mind when it comes to the kind of companies we want to support with ultimately taxpayers money.
- [11:28]
B
We should also ask James what happens if we don't manage this? You know, I think maybe perhaps 10 years, in the 10 years from now, maybe lots of companies are, lots of countries are reliance on firms like OpenAI for government decision making. Maybe at some point Donald Trump Jr. Is in the White House and he says to some African country, I want your cobalt mine and we're going to turn off the compute unless you give it to us. Is it that kind of scenario that we're actually afraid of where we have no leverage?
- [11:57]
C
So I think you're right that increasingly trade of goods and also services and so technology services sometimes fall into both those buckets in terms of both chips and, and software are being used at international scale. Have been for a long time for various political ends.
- [12:17]
B
Right.
- [12:18]
C
And there's been great times of collaboration and sort of the British and American intelligence services are great examples of both technical and expert integration, which is, you know, much to the benefit of the uk, but there's probably been times of massive overreliance as well. And if you look at one of the challenges I presented earlier on DeepMind, you know, for every pound that, every pound that Dietman
- [12:46]
B
the plumbing in our space station is Changing a bit.
- [12:49]
C
I think that's just emptying the oxygen tanks.
- [12:53]
B
That's a replenishment. Yeah.
- [12:55]
C
Donald Trump, change our gas cylinder, Pull the thing for the future.
- [13:00]
A
Go and put another 30 quid in the Elekki.
- [13:04]
C
30 quid? Wow. It's cheap.
- [13:08]
A
We already need you for another hour.
- [13:10]
B
We use a bit of solar up here.
- [13:11]
C
Yeah. Where did I.
- [13:15]
A
You were saying about how there have been periods of collaboration.
- [13:19]
C
But, yeah, so there have been huge periods of collaboration, I think many great examples of this. And the American and British intelligence services and the co integration between the expertise there and technology there have been to much of our credit and hugely beneficial in particular to Britain for defense.
- [13:36]
A
Well, hold on, surely it has been, you know, America has often benefited here and as in so many cases, when we talk about sovereignty, the Americans have retained capabilities and we have given them up in order to use American capabilities. Right.
- [13:52]
C
Like, yeah, political decisions were made where we have decided not to spend as much in defence over the last 20, 30 years and put that money to work elsewhere. And those were political decisions that were democratic decisions in many cases. Very few political parties ran on the message we should invest more in defense and less in the nhs. In fact, almost everyone campaigned, rightly or wrongly, from your perspective, on the other way around. But that was only possible because we were very aligned in collaborating with NATO and America.
- [14:20]
A
So this, I think, is the interesting point, right, is that sovereignty here is expensive and it means spending money, less money on the other things. Do you, you know, what level of sovereignty do you think? Because you talked about the economic, you talked about the. The legal, you talked about the sort of, you know, all the layers of the stack. What level is kind of appropriate now in the world that we are in? Because I think the American example is a court is. Should almost be treated, surely, as a cautionary tale rather than like a, you know, it was from a different age.
- [14:57]
C
Well, no, I think international collaboration will continue and I think it's really. There is a naive version of sovereignty where you say, once again, everything from the chips through to the application layer, everything has to be made between Somerset, Swindon and Glasgow, or it's not British. And actually, already software AI was criticized for supporting a company that dared open an office in San Francisco, as if you can build a global AI company without having some international offices. Right? And so I think you have to be realistic about what sovereignty means for anyone. You know, there's some paper I read, it's maybe six months old, but I'm sure it's still true that a Single token. So, you know, a sentence or so generated by GPT models require 35 countries in terms of the resources from, you know, the rare minerals through to the wafer design, through to the etching through back to America, let alone, sorry to produce that token, let alone looking at this international talent base they have as well. When I was in OpenAI's offices four or five years ago, so this is before the GPT model series, you'd hear polish as much as you'd hear American. So these are international efforts. So America, the wealthiest country in the world, cannot sit there and say every single part of this stack is under American control. And I think J.D. vance did shake things up when he, I think it was at the Munich security conference two years ago where he said the future of AI is American models on American chips and certainly having huge advantages like anthropic be a US company is really valuable. But America, China, Britain, none of us can build full stack alone. And so the sovereignty question, much like the defence question, is where are the parts of the stack you have to have a baseline of capability and where are the parts of the stack where you have a ability to be world leaders and if not the world leader and create for yourself a negotiating position where you say it's great that you're brilliant at X, it's great, we're brilliant at Y together we're worth more than the sum of our parts. Separate. Both of us feel the pain.
- [17:03]
A
Sure.
- [17:04]
C
And I think it's fair to say in defence there are certain areas historically, in particular with America, although not just America, where we probably haven't had that level of negotiation across the stack. But actually I think in the British intelligence services we do have that level of negotiation as it happens. And you know, there's high level political discussions people have and you know, statements from politicians saying, oh, you know, we're going to break out of this agreement or we're not going to go through that. You talk to American soldiers or American officials, talk to French officials, talk to British officials, they know the value of what Britain does. In particular on the global intelligence level, there will be certain areas of the AI stack where we can be indispensable at that level. I would argue in arm we already have a company like that. Right. And so the question sov AI is set up to answer is where else do we need to have that level of global relevance and how do we accelerate that?
- [17:56]
A
Well, let's talk about that then. So, you know, you've what you said, I mean, going from the bottom up Energy chips infrastructure in terms of data centers, data itself, architecture models and talent. Is that sort of reasonable? That covers a lot of it, that captures a lot of it. You know, roughly where, where are you seeing? Because you know, you know every day you'll, you know, you'll read, oh, there's this one Japanese company and they're producing the, you know, the foil that goes between the GPUs in the center and that means that they will be, you know, the bottleneck. Are you looking for those kind of things? Do you think we can be building these? Is that what you're looking for?
- [18:40]
C
Yeah. So there's two aspects of it. So one is, you're right, there's parts of the production. By the way, it's not just the AI infrastructure stack Today we have to have some vision as to where it may go. And so that's the beauty of having a venture approach that you take. It's not all eggs in one basket. I think that would be the very wrong thing to do at this stage. Right. And so you invest in many portfolio companies, some of which have competing visions of the future. And you say Britain is a leader in photonics. Undeniable. You look at the Photonics act globally. The there are three or four companies in Britain which you know, both legacy, legacy companies. A bit unfair but older companies and newer inserts.
- [19:14]
A
The old geriatric.
- [19:15]
C
Yeah, those old photons still knocking around but you know there are some, some companies with more of a legacy, some new insurgents where we have world leading talent. Photonics will be a really important part going forward of network switches and data centers as a whole moving data around at light speed. That is a really important infrastructure piece that we should probably be supporting. And if you look at what sort of AI has done, we put out a call to support novel chips and some of that may go towards the photonics industry. But then there's other things as well like Isomorphic in life sciences. Isomorphic is led by Sodomis CEO. The president there, Max Jedeberg is brilliant scientist himself who was at DeepMind for a long time. They're on a mission to cure all disease. It's a grand title but I love big ambitions and they've got some approaches which I think are really novel. They need data, they need some really bright people and they're getting both of those in the uk. Do they need compute based off Isambard outside of Bristol? A bit. But they could also use compute out of Portugal, they could use compute out of Norway, they could use compute out of America. They could use compute anywhere. They feel that they get a properly regulated environment. And so you could build a globally relevant company of the scale of the big pharma companies of the past, and I would argue many magnitudes larger because they're going after all disease, not a handful of therapeutics, and have a company that's so integral to world health that people wouldn't mess around with where they get their compute, with where they run their models, because they need this level. So I agree with some of the people who are like, this is a hard infrastructure problem. And it is to some degree, but it's really, I think once again, you can't overplay the idea that unless the electron that comes in is generated under British, you know, on British soil, all the way through to the chip, all the way through to the data, that you can't build really relevant companies.
- [21:18]
B
What if we tried to do the whole stack in this country? Maybe this is a naive question. I wonder whether the results, if we turned all our effort resources to it, would be creating a sort of Soviet Lada car, whereas the alternative is some kind of, you know, the Porsche, that the modern AI industry might be like, what is actually separating us from being able to do that.
- [21:44]
C
So there's some, probably a very long list, really long list of reasons why we can't do that. Let me give you one. Asml, it's a Dutch company, you may be aware of the lithography machines. Maybe there's 40 or something of these in the world that are used to create the wafers on which the silicon
- [22:03]
B
down to the nearest atom price.
- [22:06]
C
Exactly. It's incredibly precise. You know, they've got, I'm sure China, I know America have got their brightest minds with unlimited capital behind them trying to replicate these machines. It is really, really hot.
- [22:19]
B
Yeah.
- [22:19]
C
And so you can say Britain should all the way down to like, let's create, you know, well, silica. We can, we can create like the very type, the very niche type of silica required to create these kind of wafers and sort of the different mineral mixes that go into creating the silica. Let's, let's go and invade a small island and take that island which has this kind of raw material.
- [22:40]
A
I thought it all came, doesn't it all come from like one quartz mine?
- [22:43]
C
Exactly.
- [22:44]
A
In North Carolina.
- [22:45]
C
Exactly. So, you know, you can take the extreme view of how would you get there? And it would be military intervention. It would be, you know, mass mobilization of the entire country. It would be an effort beyond a Manhattan Project, beyond, you know, rescuing Our fishermen off the beaches of Normandy. It would be a much bigger effort to do the whole stack. And that's why I think it is. And the result will be worse. Well, maybe, maybe I'm, you know, ambitious about Britain's capabilities, but, you know, this would make Bletchley park look like, you know, a weekend crossword. This is like really seriously difficult technology for anyone in the world to replicate. And the companies that have been experts in this area have been doing this stuff for decades and decades and decades.
- [23:27]
A
Right.
- [23:27]
C
ARM is now the UK's most valuable company, full stop, I think, certainly most valuable public one. And you know, they have been around since 70s, late 70s, I think, and they're the best in the world at what they do. Right. There's probably 10 arm chips at least in this room right now. But ARM alone can't build anything. A single company, they couldn't produce a single chip. Right. They rely on dozens of other companies, if not more around the world. So what sovereignty means is not end to end, it is having enough power to mean no one can turn you off. And I think the closest example to this is sort of, you know, is in defense. It is like, you know, yes, it's really important that you have parts of the defense industry that you are solely involved in. But if you look at the parts that we need to build a nuclear submarine, it can't all be done in Britain. And to end, we have to have strategic relationships with other players in the world. But the good news is those are the players rely on us for certain parts of their defense, for example, the intelligence industries. And so we're integral part of that network. Sovereign AI to me means much more economic benefit for Britain, much more of the companies which are relevant and hyper relevant, like seriously of the top five companies in the world in this space, based here and aligned here with British values. But it'd be really naive to say shut everyone else off. Right, we're going to do it ourselves.
- [24:58]
B
Sure, sure.
- [24:59]
A
But I think, and I appreciate, I think this happened, started before you joined, but I think there is a sort of concern among some that some in the government and some in this kind of sphere might not be sufficiently hard nosed about like what it actually takes to be. So I mean, I'm thinking here about kind of open bind, you know, where we're going to open up these data sets and we're going to provide this great benefit to the world which, you know, I understand their arguments. One actually I'd quite be keen to hear you make that case. But to me, this, like a lot of this criticism comes off a sense that we have been kind of insufficiently hard nosed and somewhat naive. And I mean Tom, Tom is a, Tom is a big believer. I'm a big believer. I don't know, throwing you this.
- [25:55]
B
What are you about to throw me under the bus?
- [25:56]
A
Yeah, exactly. We'll go down together. That kind of, you know, we are going back into a period of far, far greater horse trading where frankly the sort of internationalist scientist, you know, let's advance global science kumbaya approach is now not really going to hack it.
- [26:17]
C
Yeah, and I think that's fair to some degree. And if you look at all the advances in computer science over the last 20, 30 years, when most people other than sadness and others were saying AI is going nowhere, machine learning is kind of interesting around the edges. What you really want is a HR SaaS app. That's what flows money, not AI research. Computer scientists were giving out their research. Right. We were publishing it publicly. All the Oxford University, Cambridge University, Imperial Edinburgh, all the best universities were publishing it. And that was happening in the U.S. it's happening in China. And then the private labs start to realize actually there's something really valuable here. And then they start paying the researchers millions of dollars to work for them and then they keep a lot of the research in house.
- [26:57]
A
Well, I mean, you talked about the, you know, Polish being spoken at OpenAI. Right. Poland hasn't necessarily benefited a huge amount as a result. Sorry, God, God.
- [27:05]
C
But yeah, and you know, I'm fairly certain Open Air does have a, have a Warsaw or Krakow office, as it happens. But you're right, they haven't benefited in the same way. But you know, why is that? It's not because the American government said, you know, to work on AI, you must come here and we're going to block you. And you know, it's because we're in an international collaborative economy where talent ended up aggregating in the best places for them. So there's lots of individual decisions. It wasn't like an American attempt to like take over Polish talent. So I think that the, and you know, everyone's benefited to some degree from software getting better. Everyone's benefited to some degree of the Internet. And I think these are all globally collaborative efforts. However, as I said, I agree with you that the private labs, in particular, some of this research and some of the techniques which are now improving the capabilities of these models are so valuable, so much money involved that you're right, that era of global cooperation in some areas is becoming more narrow now. I wasn't involved in openbind. We are working on something called sovimmune, which will come out, which is. Which is working on data sets to help toxicity and people basically for therapeutics to work better for people. And, you know, that is something we're doing privately. We're working with a handful of British companies to do that, to make sure that drugs developed with that data are, you know, the most effective they can be and genuinely save people's lives. Now, at some point, of course, we should open source the stuff which. Or share openly the stuff which benefits everyone. I don't think that you benefit from taking a completely closed approach, but we can be more strategic and the things that we want to stay in Britain and the things we want Britain to benefit more of than we have been over the last 30, 40 years, I think that's fair.
- [28:56]
B
What do other people in VC or the tech world make of Soviet?
- [29:03]
C
Yeah, well, the nice thing is everyone's annoyed so you're doing something right. I think there's some fair criticism from some people and by the way, in some areas of the economy, I'm in their camp, that say this isn't a role for government, this is a role for the private sector. British venture capital and European venture capital has been on a tear over the last 10 years. We're just starting to get going. There are some really great businesses being built in AI in the uk. What role does the government have? You know, stay out, don't crowd out private capital.
- [29:38]
A
And then forgetting that the British Business bank is a.
- [29:41]
C
Well, indeed. It's very easy to. Yeah, exactly. To a lot of these new funds, I agree. But they would say it's great to be a fund of funds. It's great for the British Business bank to invest in my fund, but it's not great for them to invest in another fund and it's terrible if they invest directly.
- [29:57]
A
My middleman fee, My middleman fee, you
- [30:00]
C
get paid for your expertise. So I think there's that angle which is saying you shouldn't do any of this. And there's the other angle which is saying you're way too small, you can't do anything with. Have more money. If you're going to do this, do it properly. The Europeans have just announced a 5 billion euro growth fund which is a combination of EU money and pension money and other sort of large institutional money from across Europe.
- [30:24]
B
Fast track, three year application process. It's so cheap.
- [30:28]
A
It's so cheap.
- [30:28]
B
I shouldn't do it,
- [30:31]
C
you know, I actually think they've done it in an interesting way. We could talk about that more if you want, because it is the first time. I thought, actually that's quite a smart way of doing it. But, you know, so what good is your £500 million? And I think both of those criticisms are fair in certain ways. On the you shouldn't be doing this. In lots of areas of the economy, the government shouldn't be as involved as it is, for sure. And it certainly shouldn't be subsidizing companies that, you know, they think are the right things to do. The market pricing mechanism is the best way to do that. Right. Entrepreneurs are some of the best ways. Private money going to entrepreneurs are some of the best ways to do that. That's proven, we all know that. But there are certain times, and there are certain parts of the economy where actually the shift is so big, you do need government involvement. We've done that historically in defense. And I think AI is one of those technologies where we do need to be involved. Then on the scale side, you know, we're trying to do something new. We're introducing this form of venture craft statecraft into government work. Like in central government, we've brought in external VCs to do some of that. People have been adventure their whole lives. You know, that is a big thing for the government to do. And so we can all say do more, do it faster. But we kind of have to earn our right to do that. Right. We have to prove that we can do it. And actually if we do a good job at sovereign AI, there's hundreds of millions of pounds through British Business Bank, National Wealth Fund, many other global institutions that could come to bear. And as it happens, you know, we've made four investments in the last two months. Equity investments, deliver computer, many more. And three of those investments, we've had global funds knock on our door already being like, how the hell did you get to that company? We'd like to be in that company. And that means we can crowd it. Well, you know, we've hired some rich people.
- [32:15]
B
Which ones are you thinking of? I mean, there's Colosum. That's pretty cool.
- [32:18]
C
Colosum, Ineffable and Isomorphic Labs. Right. So Callosum, you know, Oxus spin out, backed by Aria, Oxford, Cambridge, I think backed by Aria. They are, if it works, going to allow people to use heterogeneous compute, so lots of different chips to do this stuff. Which solves one of our issues, which is, are we overly reliant on one chip Company, if it works completely upends the power of certain parts of infrastructure. Stack ineffable intelligence in David Silver's company, one of the most important AI researchers in the world, spent over a Decade Building at DeepMind has now thinks he can take a completely different architectural approach to AI which would completely upend what anthropic and OpenAI have if it works. And Isomorphic Labs which we've discussed already which are on a mission to cure all disease led by Britain's greatest living scientist. You know those three companies alone, if one of them works, we have done our job at sovereign AI. Now they may not, but we've managed to do that in a couple of months and I hope we can do a lot more in the, in the years to come.
- [33:20]
A
It's very exciting and I think a lot of people would actually say that Britain's venture scene, well if you listen to the VCs is this sort of great success. The thing that is less discussed is the need then because of the kind of faltering, I think in the nicest possible way of putting it, state of the London Stock Exchange means that the kind of path of going from small company funded by VC up to being sort of what small cap and then you have to go, you know, and you have to IPO abroad. And the fact that our pension funds aren't necessarily investing in equities means that, you know, your job here is effectively feeding promising companies to the US like brilliant, brilliant, brilliant. You know, okay, they're actually making some money, ship them off. Like how do you think about this? I mean there's other points here about like the difficulty of getting loans if you're a small family company. You know, VC backed companies are not the only form. Like I'd be interested in your thoughts on this in the sort of business in general.
- [34:33]
C
Yeah. And also I separate out those two things. So I think it's a really important discussion about SMBs, family businesses, people who just want to, you know, this in no way saying this in a negative way, saying it's really positive way. People want to feed their family. Building a business, right. Make a living building a business. That's an incredibly important, often overlooked part of the economy. And then there's a slice that I operate in which is venture and venture backs companies which do have the potential to operate at a global scale and do that via exponential growth.
- [35:03]
B
Right.
- [35:03]
C
Which means they have a lot of impact very quickly. And that's a very different type of business. Britain should benefit more from the very now sizable number of companies that are in that slice. And it is, there's loads of problems in Britain. One of the things we're actually really good at is building these companies to a certain scale. And as you've mentioned, there's some limits on going beyond that scale. Now, when I started Inventure, which is like 12, 13 years ago, a everyone said, don't go into venture in the uk. You're mad. No one's ever been successful in doing that. So go to San Francisco, which is helpful to hear. But luckily I made the naive choice to not do that. And everyone said, well, there are no technology companies. And we had some technology companies, we had a bit of venture. And everyone said, well yeah, you've got some, but there's no billion pound technology companies. And 10 years later we've got a lot of billion pound technology companies. Now everyone's saying, yeah, but where's your 10 billion pound one? Why aren't they listing on where's your 100 billion pounds? And by the way, now we've got two, I think 100 billion pound companies, private and public, at least in the technology space, and we're going to have more. So we're on a improving curve. Where does the curve really benefit Britain is something I am concerned about because as you say, over time companies need different levels of capital. Anthropic Yesterday announced a $65 billion funding round, not valuation funding round. And there is no institution in the UK which could provide that level of capital. And you would struggle to find institutions across Europe that could provide that level of capital. And so you need to think about how do you build this ecosystem? And we have. You're right to say that the domestic stock exchange in the UK has its challenges. You know, there have been technology companies listed there. You know, ARM historically I think was listed there. But you know, more recently darktrace, which we were investors in, were listed there and you know, technology companies like Deliveroo and others, lots of interesting small cap sort of stuff in photonics, one area, as it happens. But it's limited. And it's limited partly just because the size of the country, but also because of listing rules which are trying to reform. That is a problem. It wouldn't be such a big problem if at the same time it was the case that loads of Brits had invested in these companies privately. Right. So obviously public investment is great because you can get more money in that way, obviously much more pools of capital, but also more people can benefit because you can own a single stock or you can own a million Stocks, you know, it can go up and down, but you could as an individual benefit. In private companies, it's harder to buy a single stock and it's riskier. And so generally speaking, you do it through an institution and generally speaking globally, those institutions are pension funds because they have the largest pools of capital. And in the UK for the last 15 years, basically since the start of the movement towards DC, we've completely kneecapped our ability for average people through their largest asset, which is liquid, which is the pension, to get exposure to these winners. And at the same time, so it's bad for Britons, bad for average Brits. At the same time, it's not bad for British companies because they're so good they can attract international capital, but it means they have to make difficult decisions. For example, do you move wholesale to the us? Do you set up in the US from day one knowing this is going to be a problem in future, as some people, I think wrongly, but still choose to do? Or do you sell up early? Because you say the only way for me to achieve my mission of building that AI company or building that E commerce company, whatever it is, is to join ebay or to join Google. And we have to change that. And just as a small sort of anecdote around this, the fund I joined when I first joined Balderton, became a partner of Balderton. That fund was near the end of DB schemes allocating capital because they've had to start winding down, right? And that was over 10 years ago. And in that fund we had a coal miners pension fund, we had a major supermarkets pension fund, a few others, and it was like their last hurrah intervention before they had to start winding down. That fund has been enormously successful. And so maybe it's a one off, right? But that fund alone will mean that those people who worked for 30, 40 years in that supermarket or worked in that coal mine will have thousands of pounds and possibly tens of thousands of pounds more in that pension pot because they allocated a tiny amount, 1 or 2% to venture, because they had exposure through us to Revolut, to Gocardless, to Nutmeg, to Depop and a bunch of other companies that are in that fund. Since then, since that has happened, we've had barely any UK exposure at all. In fact, British Business bank, which is small investments of our funds, is probably the best way. Brits benefit from it. Canadians have benefited, Americans have benefited, Danes, Swedes, Germans, some Australians. Very few Brits have. And it's really sad that that's happened, but it was entirely a decision made by the government at the time to make a change to force them into bonds. I don't think that was explicit the reason why they changed it and there was lots of reasons, but the reasons that impacted the ability to invest in venture was the fee cap. So government came along and for lots of reasons wanted to change the DB and DC pension funds. But they said it is outrageous that you can use the Internet to buy a single share of BT. Now it costs nothing, 0.001 cents and people are charging 2% fees to manage your money. That is outrageous. Putting a fee cap in. And they treated people who like us spend a decade investing in companies built on the board of a company, working for the company or all that time the same way as a hedge fund that presses a button to buy or sell a share. Now if you press a button to buy or sell a share you can drop your fees pretty low if you are building a 10 year relationship with a company. The thing that they said they wanted patient capital value add. You know, like all the things that if you're Labour or Tory, if you believe in capitalism at all, even on the left you want, they basically effectively made impossible to do through a UK pension fund. You know, lots of private institutions in the UK still decided to invest in us because they could. They're not bound by these things and they saw that what matters isn't the fee you pay up front, it's the long term return. It's like saying, well, you know, if I get charged 1% or 2% for buying, you know, this, this, you know, whatever it is, rare vintage watch or whatever I like. It's irrelevant if you're going to sell the differences between selling it for £100 or £500. Right. It's not about the upfront fee, it's about the outcome. And unfortunately through well meaning legislation we focused the whole of the pension industry on the upfront costs rather than the long term results. And we're now in the long run, we're now 15 years later and the pinch is being felt.
- [41:55]
A
I've spoken a lot. Do you have any questions for our guest?
- [41:58]
B
Well, we were talking about how to. We've defined sovereign AI in perhaps a broadsheet way. How far have you got on this question?
- [42:07]
C
Yeah, so it's really hard because it's so easy. So you see political opportunities doing this already. We've seen it in the States and now, you know, the political parties who just follow whatever happens in the States and try and do it in Britain are Doing it here to say AI is bad, right? It's change, it's going to be disruptive. Even the people building these models are, I think, particularly bad at explaining a positive future. They seem of sense obsessed with the idea of explaining sort of this dystopian future sometimes. And you know, personally, I find it, I think it's complete rubbish. I'm a Schumpeterian economist. If you look at the history of all technologies, you know, it took us out the fields, it took us out the mines, right? It's amazing what's happened over the last hundred years thanks to the progression of technology, making our lives better. But the challenge is you have people who can explicitly say something's changing today. And the other side of it, which is saying, but we're not sure what the future is going to look like. We just think it's going to be better because of these economic laws around productivity surpluses. It's not a very attractive pitch. And so I've been thinking a lot because I've lived this as an investor, at least through technological change before, because I am Schumpeterian. I've been thinking through how do you land these questions? And Schumpeter doesn't help much because he says is creative destruction. And the problem with creative destruction is they're not necessarily co located. Where the value is destroyed and replaced creatively with even more value doesn't mean it has to be the same town or city. And so we have to find ways to like, explain how we're going to go through this transition, why it's going to be a good thing and why it matters to individuals.
- [43:47]
B
I think there's some scales all very well, but there is, I think, a significant chance that we're entering into an era where it's not kind of Schumpeter's sort of like gentle puff, but in fact, like Schumpeter's tornado. What if kind of all cognitive work and all manual labor is automated, that probably gives you the strongest case for sovereign AI. A lot is going to change in the world. We're going to try and ensure that the country stays solvent and that, you know, the individual man or woman can, can have a slice of it. Isn't that the.
- [44:17]
C
No, I think it's an extremist argument because you look at, you know, what people were saying around the turn of the century with flight and cars coming online saying the same thing, right? What does the average working man in the field do when a tractor can do the work of 50 men? Right? There are going to be millions and millions of people made destitute. And some countries that navigated that badly did struggle, and lots of people did struggle, but the countries that navigated it well saw an economic and health benefit that's never been seen in the history of humanity before. And then we did it again in the mines when we bought huge manufacturing and automation to the mines, and we've done it again with mass manufacturing and factories. And so actually, I don't believe this shift into, like, oh, well, this time it's cognitive work. It is certain areas of cognitive work. But let me tell you, a calculator is better than you at maths and the spreadsheet is better you at statistics. And almost everything we've done in healthcare over the last 20, 30 years is better than all our top researchers. Right. Because they've had to automate and use statistical techniques which no human could do themselves. So we've had this level of change in certain areas of the economy before. It's coming faster and it may be broader, but I don't think it's sort of existential. I do think it's really important, though, that Britain benefits from it. So overall, the global economy is going to benefit a lot from AI. Overall, I think humanity is going to benefit a lot from these new tools in many other ways than just economically. But Britain's slice of that is what really matters to me. And more broadly, Europe's slice and the pie grain for everyone. And that's what we're trying to secure at Solve AI. And so I think the really important thing is we do make that more tangible to people. I personally think the most tangible way we're going to see it in the UK is in healthcare and life sciences. And so, you know, when people say, well, what are you doing the chip layer? What are you doing around photonics? What, around these rare minerals? That's all important and part of the plan. Data science is part of the plan. But I do think life science is probably one of the areas we're going to see the biggest impact.
- [46:18]
B
Are there any particular fields or areas within life sciences where you're. You're most interested in hearing from new companies?
- [46:25]
C
Yeah. So personalised therapeutics. And I think that the.
- [46:29]
B
Can you deal with snoring?
- [46:32]
C
Well, you know, the red top.
- [46:34]
A
Turns out, Tom, there's already a cure for that. It's a pillow.
- [46:40]
B
So I just want to be really clear. We don't share a bed, we live in the same flat. And I can hear it through the
- [46:45]
A
wall,
- [46:48]
C
you know, you know, maybe GLP1 secretly cures snoring as well. You know, it seems to cure.
- [46:52]
A
There's not that much more.
- [46:57]
C
No, I think there's like this offer out there around AI, which once again political opportunists and it drives me mad already because it's so sad saying, oh no, AI in healthcare, this is a terrible thing. And you think, well, so what happens if you've got a loved one who goes in with a rare cancer where you can take a blunderbuss approach? And let's be clear, most healthcare over the course of human history has basically been guessing and we've been getting better at guessing. We've taken a probabilistic approach and saying we think this approach of chemo is going to work or we think this particular treatment is going to work. And we've used the X rays now and we've got slightly better view and we can do a blood test now we've got a slightly better view and now we have the world's greatest predictive model of human health at our fingertips that's going to help design a drug perfectly targeted to the cancer that is killing your loved one. And you say AI in the nhs? No thanks. Right, well let's go back to moon cycles for predicting this stuff. And so I think the way that I think AI is going to have the most immediate impact in the UK is probably going to be through health care. The way that people feel it on a day to day level, beyond getting slightly better in their job or beyond maybe some more entertaining use cases in gaming, whatever it is, whether or not people recognize it's AI, I don't know. But what really worries me is already people popping up online and these aren't just Twitter trolls, these are like serious people in political parties polling at double digits saying we don't want AI in the nhs. Well, throw out the calculators as well, get rid of the spreadsheets, get rid of the blood tests. It's a really naive way of approaching it. And what I'm hoping is that we're going to have some breakthroughs as we already have from an isomorphic or something someone else that show. Actually, you know, this is transformative stuff and it's going to really help a lot of people.
- [48:46]
A
It's an interesting point because actually I think a lot of Britain's success over kind of the 19th century can be explained by, you know, we were just much more liberal and that we were willing to bring on technology and we were a free trading nation and you know, that the workers saw Free trade and technology as a thing that potentially, you know, would, would benefit them. There are extremely, you know, I mean you've sort of skirt around it but there are extremely opposed people. You know, these myths about water usage, the myths about, you know, it's all, it's all theft, it's all, you know, still a stochastic parrot. You know, I actually don't think many people are well of that ilk will be listening to the podcast. But what is your, how would you talk to them? What is the, the way that you actually taking it even more, Trying to compress it down even more. Yeah, like how, how can we talk about this as a positive force?
- [49:51]
C
So I think, look, first of all you always can have dissenting voices, right? Political opportunism is always going to be a thing. There's always going to be people who see votes and fear and ostoke fear to get those votes. You know, you talk about historical, you know, our historic and rich past as sort of scientific engineer merchantiles who combine these different mindsets and traditions and you know, Florence Nightingale went out to the Crimea, introduced the use effectively of data visualization to healthcare and created the friggin hospital. Her and Isaiah King de Brunel and you know, that kind of incredible legacy, you know, wouldn't have been possible in most countries of its time. But she had dissenters, she had loads of people being like this is, you know, what are you doing? This is the wrong way to approach it. You're not a trained physician because you couldn't be because you're a woman. You know, there's always going to be political dissenters. But I think most people look back now and say it was probably a good thing Florence Nightingale came along and applied scientific endeavor to this real world use case. And what we need is more people taking up that kind of mantle, Britain's historical mantle of being these scientific entrepreneurs of combining sort of rigor and a search for truth and a willingness to have an open debate with this idea that actually we have a role to play on the global stage.
- [51:10]
A
But isn't that you, aren't you the one that actually made that argument?
- [51:13]
C
No, I'm making, we need more people to make those arguments. Can't be a sole voice. I'm making the argument, you know, chair of sovereign AI for a reason. Right. Because I care about this stuff and I think it's really important I do venture capital across Europe but you know, because the European values matter as well. But specifically in the UK because you know, not only do I see this as like a big economic opportunity because those two things are combined often, you know, social impact and successful companies, but also because I care about it. So yeah, I made that case and I think you can say, well, where is this golden land that AI offers us? There's a lot happening today already, a lot that's happened already. You're a kid in Manchester, where I grew up. Twenty years ago, you couldn't afford to build a frigging website, you couldn't afford to launch one. I did web agency work when I was a kid. It was nigh on impossible. You needed a server, you need a designer. Some people had to rip off Adobe Photoshop licenses illegally to access that kind of stuff because it was friggin expensive. Right now you can launch an online store like that, right? You can go on Depop, you can launch one, Etsy, whatever. And that's a tiny part of the economy, right? Online shopping, a bigger one now, but admittedly small one overall today. That kid in Manchester, who's a teenager thanks to AI, has the abilities of a software developer who just a year or two ago would have cost 30, 40 grand for a project which they could never afford. A marketing expert which once again would have cost 30, 40 grand. You know, advice tailored to their needs on how to launch an application or a business, all thanks to a model which, you know, has a free tier, but it's probably £20amonth at its premium tier. That is a fundamental change and people are seeing benefits around this already in the economy. And so those benefits and those wins need to be shared more wholly. I was at dinner a couple of weeks ago with my wife's auntie and she's a physiotherapist and she's not an early adopter of new technologies. She doesn't use Apple Pay or anything like that. But she said for the first time in her very long career as a physiotherapist, there's been a technology chain that's really helped her and that's being able to dictate fluently as she talks to a patient. And she's tried all the technologies the last 20 years, none of them work well. And so as a physiotherapist, you spend half your time tap, tap, tap, you know, stretch, stretch, tap, tap, tap, stretch, stretch, stretch. And now she can fully dedicate her time to the patient and look at them and talk to them about their issues, knowing that what she's saying, the correct words are being picked up on a secure device. It's an Australian company that does this, but they have got a license with the NHS Secure device, which turns it all into free flowing text, which you can choose to put into your electronic health record system with human oversight. That is astonishing change in care and that's available today. So these things are happening. I think the big thing I'd love to see over time in the UK and we talk about lots of areas of sovereignty that we need, but where's our comparative advantage? I think that is in healthcare. And you're sitting on your Anglo futuristic space station up in space, which is great, but other countries may beat us to space, as it happens. Britain's got a very long history of providing great health care. We invest a huge amount of our gdp, you may say, sometimes erroneously or inefficiently rather into health care. But we also have some of the world's leading laboratories and private companies developing pharmaceuticals and drugs for treating people. And I think we get this right, we get AI right in Britain, we get our health service right, we could be the envy of the world again in providing health care to people. And when someone goes in and they get treated better with a personalized drug or faster because an automated system, or they feel it's more personable because the physician can look them in the eye and talk to them rather than spend all their time on bloody paperwork, they will say, you know what, it's great. This AI stuff is great. And I think that articulating that is one thing, but delivering it is going to be harder. And I hope that, you know, while you guys are looking down at us in our space station, we've got the world's, you know, you know, Brian Johnson, who thinks he's going to live forever, desperately trying to get his British visa to come over here and get treated in, you know, Nightingale Hospital Cambridge or whatever, because it's got the best stuff in the world and we could do that. That's realistic. It's hard, it's going to cost a lot of money and take a lot of time, but it's not beyond our capabilities. And so when people say, well, we don't have an Nvidia, we don't have an OpenAI, sure, okay, and we need some of those things. We need partnerships and resilience, infrastructure and da, da, da, da, da. But what happens if we had AstraZeneca 2.0, or what happens if we had, you know, next generation? Nightingale might be far more impactful in some ways. And I can promise you, Jensen Huang, if he's getting treated in there for his cough is not going to turn the chips off. So there's these areas and you know, I don't be flippant about application layer stuff or whatever, but I think life science is one of those areas where we really could lead the world.
- [56:21]
B
And what is our advantage? Like the. Apart from all the data, like is your case that we have a lot of talents in this area. I mean, I probably agree with you,
- [56:31]
C
but we've got A extreme talent in this area. We've got B. We have got big British pharmaceutical companies still. But you know, if we get it right, the most important bit, the largest employer in the UK is the National Health Service, one of the largest employers in the world. We have more people working in one organization focused on one goal, which is supposed to be the health of the British people, than any other country in the world. There is no country in the world that has a higher number of its people working on that single goal under one organization with one clear leadership. Now, can you make it work really, really hard? Have we made it work? Not always, but it's a hell of an opportunity. And you know, once again, maybe Schumpeter's going out of favour at the moment on this podcast. I also believe comparative advantage seems to be the other economic term going out of favour at the moment. But I'm still a believer in comparative advantage and that nations should lean into their strengths. And I'd say this is still one of our strengths.
- [57:33]
A
Yeah, I suppose again the. And I think we should stress we are not at all opposed to technological innovation or comparative advantage.
- [57:42]
C
Am I out neoliberaling the Anglo Futurist podcast? Because that was not the essential.
- [57:47]
A
I think it's more probably a doorman's dorman's fallacy. Right. That motivates our questioning, right. That some things are not being measured and that leads to long term misallocation. Despite everything looking like rosy on the surface and in the kind of spreadsheet. I mean, I think a classic example in this sovereignty stack would be, for example, you know, the ability to make things, you know, also our sort of lack of steel production and you know, our lack of PCB fabs or chips or whatever. So it's a sort of defense point. So I think, I think it's. Yes. And I think there are massive benefits that we could reap if we were the one polity that remained really friendly to AI as everyone, you know, imagine, imagine. I mean imagine if anthropic came and came to, to Britain. It would be, it would be astonishing. Yeah, but I, but that I think is where the, the, the worry comes from. Is that, is that fair? Would you agree with that?
- [58:56]
B
Our worry is, is that, is the,
- [58:58]
A
the sort of the Dorman's fallacy that, you know, that you can, you can arrange everything and you can say it looks good on paper, but you follow that logic too, and it produces a sort of.
- [59:13]
B
There are so many other problems as well. Right. It must be a great frustration to you that we have some of the most expensive energy in the world. That is not helping what you're trying to do whatsoever.
- [59:23]
C
No, I mean, electrons are really an integral part of training these models and to a lesser extent, running these models. And we need more electrons. But as I said, there's, you know, there's comparative advantage still. You can work with other countries and, you know, I mentioned, you know, okay, OpenAI isn't building a Stargate in the UK anymore, actually not building one in Norway anyway, but Microsoft's going to build one and we can work with them instead. So, you know, there's a challenge around bringing energy costs down, but it's not,
- [59:50]
B
it's not just energy. I say that to kind of grope at the vast numbers of obstacles to entrepreneurship in this country. So it's also incredible housing costs. Right. Or where could I even. Like, where do we even start? I mean, the high taxes.
- [60:09]
C
Yeah. Cost of doing business. There's a bunch of things where you could say, so, you know, very explicitly. You know, my day job at Boulderson where I work with entrepreneurs building software businesses mostly, although what software business is, is much broader now, still software businesses. And my day job, sorry, my. The other day job I have as chair of sovereign AI, which has felt like two jobs a lot recently of working with AI companies. You know, despite everyone being critical of like, oh, where are we versus America? Well, software business and AI. Two areas where Britain's really strong actually compared to our European colleagues, any other peers.
- [60:45]
B
Right.
- [60:45]
C
We're the third largest venture economy in the world. Maybe we're a magnitude too small, but we're ahead of everyone else. There isn't a country in the world other than American China that wouldn't want Britain's position in AI and venture right now. Housing, energy, broader costs. Right. Actually, those don't really affect our ability to compete in AI and software that much. They really hamper other parts of the economy, which are obviously really, really important. So, you know, well, how do you make sure you diffuse AI quickly? We want lots of companies buying it and testing this stuff. The companies don't have high margins to Spend on R and D because of having to invest elsewhere. Well that slows down the diffusion of it. If people can't feel like they're getting on the housing ladder or can't afford like you know, to build big hardware projects which could benefit from AI because the energy is so expensive, that's obviously going to impact the long run benefits to the UK of AI. So there's a bunch of other things and there's, you know, I'm not the first person to say this. There's this weird trope in Britain where we're really good at really hard stuff. Life sciences, AI, you know, academic research in, you know, plentifold attracting world leading talent and we're not great at some of the basics and I think that's. That is a broader issue. My role at Sovereign, I'm very focused, you know, I don't have any responsibilities beyond that within the government. So there's other things we need to fix.
- [62:09]
B
So you're not, I mean you're not that far from Ed Binniban's office. You're not going to knock on the
- [62:13]
C
door across the road.
- [62:13]
B
Yeah, please give a license to Jackdaw.
- [62:18]
C
I'm focused on from the electron through to the model, not where the electrons come from. But you know, look, I think the energy landscape needs to improve for lots of reasons and I think, you know, you make a good point about people just spreading misinformation around the amount of water that there's like one study based on a data center in Texas on old technology about how much water a data center uses which is being applied to data centers being built in ways calculated.
- [62:48]
A
It's only by three orders of magnitude.
- [62:49]
C
Yeah.
- [62:49]
A
So there was then a retraction.
- [62:51]
C
Yeah, forever. I mean it's just like. Exactly. It's just ludicrous.
- [62:57]
A
But that's who you're actually talking, that's who you need to be talking to. Right.
- [63:01]
C
Well there's always going to be people spreading misinformation. That bit doesn't with people on Twitter saying crazy stuff stuff or wherever that's always going to exist. I don't know if that is the case. I think what matters is people feel there's benefits from this technology and they're going to feel it in their wallets, they're going to feel it in their health, they're going to feel it in day to day bureaucracy. There's a national pride that we have some of these winners. We need to make that happen. I don't care about what people spending lies in the Short run does, unless it really changes the political calculus and then there's maybe a broader problem. But to your point, there's a bunch of other stuff to fix. In Britain. My role in AI is a really important one or at least what sovereign AI is doing I think is really important because it's such a significant technology. But actually on a day to day basis there's a bunch of other things which need to improve as well.
- [63:49]
B
You invoke national pride. And I think it's interesting to observe what seems to be a bit of a culture of doing a tour of duty from people in B.C. like Ian Hogarth, who's, who's been running AC and Tour of duty is this term that people are now using to describe going into government and like earning peanuts to do something they think is important for the country. Is it just like a small subset of people who are talking about it or is it reflective of a broader shift in attitude or what? How can we understand this?
- [64:18]
C
Yeah, and to clear I'm not paid, so I even. I pay for my.
- [64:21]
B
Have you had to like give up. Ian Hogarth gave up his stake in Anthropic?
- [64:26]
C
Yeah, he sold. Yeah.
- [64:27]
B
You have to do something like that.
- [64:30]
C
No, but there's certain things I'm not investing in which I would invest in to avoid any kind of perceived.
- [64:34]
B
So you're probably taking quite a big hit.
- [64:38]
C
Look, we're doing fine at Balderson. So you know, I think that the sacrifice is time and effort and personal reputation. By the way, there's a lot of people now who just really don't like me. Only for the fact that I've said AI might be an important thing for Britain to be involved in. And then you get, you know, government shouldn't do anything or government shouldn't do I people. But it's not. Anyway, it's not. I mean there's people who I've done the least because there's people who, you know, you know, two of the people that we bought into the Ventures team have come out of like a decade long career as a seed investor, someone who was at Google nyc. They've taken big pay cuts. They're doing this full time now since, since we set it all up. So there are more people doing this and it's actually to the credit of the civil service and the previous government and the current one that they've recognized there's a skills gap because Britain hasn't done venture from the central state. We haven't done anything around AI. Obviously historically we haven't done enough probably outsourced it to some institutions which are slightly more academic than practical. And so three or four years ago, whether it was Mac Clifford coming in under Rishi for Bletchley park, whereas Ian coming in for ac, Alex to pledge, coming in the treasury to work on some entrepreneurs offers there, or what I'm doing at Sovereignty and others, there is a. It's great that government has said, you know, we need to bring some people in. It's a short term fix. What I hope is that you don't keep bringing people in constantly from the external world for short stints, that you build capability internally and that, you know, the vast resources of some service and put to work to fix these things and that we're short term stopgaps to do it. But I think there's a broader issue which is a sadder one really, which I've recognized and I think that I hope that it's not what little I can do, but hope is changing, which is a combination of the state probably getting less productive for lots of reasons, Brexit really splitting the country and the Leviathan, the state being really targeted in that regard and a bunch of other perceived failures over the last, let's say, decade have meant that a lot of the public really, you know, there's a negative view of the civil service compared to the esteem it was held in maybe in the 60s, 70s, 80s. It's part of a broader trend of Britain becoming less enamored with its institutions. Right. It's a broader trend of all of us, not necessarily showing the respect for nostalgia that maybe we historically have. And in some ways that's really good because it means more accountability. But it's quite sad because there's some brilliant people working in civil service and tour of duty from people who've done really well in the private sector for 10 years and spend a few years out is one thing. But what about the people who spend their entire career trying to do hard work, good work in the civil service? I think those people don't get enough recognition for the efforts they've made. And partly it's mistakes the governments have made because maybe they hired too many people to do X or they managed it badly. But we can't have a situation where the only people who get recognized for making a sacrifice are Ian and Matt and me, who disappear for a year and then go back into the private sector. There's people who do this full time and they're good people. And so what I hope is that sort of the status of working on these projects Internally increases because it can't just be one offs, it has to be consistent career opportunities.
- [68:12]
A
But what explains then that loss of. Because if you were saying they're actually great and people outside are saying actually no, we find this incredibly frustrating. What is your explanation then? Or is this just people making stuff up?
- [68:32]
C
No, no, no. Look, I think there's lots of institutions which where, you know, historical levels of respect have fallen. Right. This is a cultural thing and that's often a good thing. Right. I think you can hold institutions in too high regard, which means they become sclerotic and don't change and need to. But I think that it's completely fair to say there's large swathes of the state which haven't worked well for people because either they haven't caught up with different types of demand or shift in demographics or just haven't been well run. Right. That's a completely fair criticism of a system. But not everyone in the system is the perpetrator in that regard. Generally speaking, people follow the incentives you give them. And if you give a very large organization like the civil service a bunch of bad incentives, then you're going to get a bunch of bad outcomes. One of the ways you fix that, if I fix in the incentive structure, civil service reform, sure. The other way you fix that is by getting great talent in to build from the ground up. I'm a big believer. My job is to find exceptional people and give them money to build great businesses. That's what I do and I've been doing for 12 years. I want more exceptional people in the civil service and I hope that some of that. I'm not sure respect's the right word, but certainly admiration, aspiration and ambition around what the civil service can do return so we can get some great people working there again, because there are great people working there, they don't get enough credit. And you therefore get this negative flywheel where you get people saying, well, Jesus, I'm working really hard, trying my best, doing what I'm told to do, and everyone tells me I'm useless and I don't get paid well for being told that I'm useless. And so you know what, that private sector job looks pretty appealing. And you start losing really, really good people. And I've worked with loads of great people in DCIT and been really reassured by that fact. Loads of institutional issues. But, you know, the Torah service thing sounds great. And honestly, you know what Ian and Alex and others have done fantastic. But I don't think it deserves anywhere near the amount of praise as the unsung civil servants who are working harder than any of us did.
- [70:49]
B
So if your work and the work of those unsung civil servants goes very well, what's the kind of Britain that we could perhaps imagine 50 years from now?
- [71:01]
C
So in sort of, I don't know whether it's 10 years or 50 years time, I don't know if the change will come from the state, the positive change that Britain needs to see. The state's important, but it's not everything, right. I think most things, you know, once you get some level of basics, right, and we do have the A level of basics, right, Well, a bit too negative. It's a load to do better. Most things are downstream of culture, I believe. And what I would love to see and what I think would benefit most Britons and the world as a whole is more of an entrepreneurial culture embedded within British culture. And actually, historically, we've been incredibly entrepreneurial. Right. I think there's a combination of British values around fairness, around rigor, around hard work and, you know, scientific endeavor and merchantism, you know, bordering on buccaneering to some degree. But all of those values which really defined a lot of, you know, the great growth we saw through the industrial revolution, the Victorian era, the early 1920s, 30s, even the post war era, to some degree, those values naturally lend themselves to people who will use tools to push scientific frontiers and find ways to create value out of it. We've been really good at that. And then over the last 30, 40, 50 years, I think more and more people's careers, and therefore your culture gets captured in this as well, has been caught up in this pincer movement. And the pincer movement was part of the post war settlement where he said, okay, you're going to have a steady career and we're going to create some big companies to give you steady careers and we're going to have a big state. And so here's your career options. Go work for the big state, go and work for big corporate. And that was great because we went through this massive period of disruption and suddenly you have this whole system which is job for life, learn a skill, job for life, move up the managerial ranks. Your mortgage is based on a very predictable salary. So that's great. You know, personally, you get more dignity over time because you move from, you know, being the underling to being the big boss.
- [73:07]
A
From the corporate through the range of like Volvo cars.
- [73:10]
C
Yeah, yeah, yeah, exactly. Badge on the back changes. I was more Volkswagen. It was like Polo Golf.
- [73:17]
A
The seats Recline all the way back.
- [73:19]
C
And you know, that idea of progress, and it happened for lots of reasons why, but that idea of progress, it was very appealing. But actually that corporatism has been short lived, I think. And going forward we're gonna go through this massive period of disruption. Part of it is technological for sure, but part of it is just, you know, a lot of the big companies that employed us are going to employ fewer people because of automation and the big state that employs so many isn't going to be able, like the maths doesn't matter anymore on having a massive state. Technology will make it easier to do stuff better. But also you can't employ more than 50%, you can't pay more than 50% of the people with fewer and fewer people's taxes. It just doesn't work, especially as more people don't work at all. So you've got people who don't work, we need to support them through pensions. You've got people who work in the public service, which is really important, but get paid through taxes. And so you get a smaller portion of your people generating those taxes. Eventually this system breaks, right? Right now we're propping that system up through lending and paying the interest on that lending is 10% of our spend, right? So you've got a system where the public sector needs to find a way to deliver better services with at least the same resources because you just can't continue to grow those resources. And you've got a shrinking private sector. The way out of that is entrepreneurship. The problem is that we've built a system, whether it's university debt, whether it's bank lending only to big companies, whether it's literally getting a mortgage, try and get a mortgage. An entrepreneur versus someone with a salary, career. All of these things have been built to entice you into a corporate career. While at the same time as you are living as these entrepreneurial journalists, podcasters, futurists, the tools at your disposal to start something new, to try something new, to build a business, to experiment, to side hustle are getting better and better and better. But it doesn't matter what happens with those tools if the system itself doesn't incentivize you or support you to take risks. And on top of that, we have a culture which is still saying the high status thing to do is have the fanciest Volvo in the cul de sac, right? That's the high status thing to do. Or the high status thing to do is go and work your way up to your grade six or seven in the civil service over 20 years, which doesn't reward you if you're better or worse than anyone else, just rewards. Tony. Right, I'm building straw men but broadly, historically has been the case. So what I want For Britain in 30, 40 years time is a far more dynamic economy where more people are willing to try different jobs, to try different ideas, to build, to ship, to fail, to rebuild, start again, where risk capital is available to anyone, to have a go because there's enough winners in the system to pay off the experiments that don't work. And hopefully two things happen. One is we get back to sort of thinking of ourselves, Britain as not like, you know, a big corporatist state where you have a massive state and you also have like five big brands, employers, you know, the Tesco's and the Barclays, you know, that is not an economy I think is going to be sustainable. But also you get back this, you know, idea of us as, you know, scientific merchants, as people who are going to think globally because to be a successful entrepreneur you do have to come and think globally and to be engaged in the scientific frontier. And I want people to reward that. And there's, you know, it's a bit of a trope but you go and ask most people today like who's the most famous entrepreneur, you know, and British entrepreneur and the number one, this was maybe 10 years ago now, number one was Del Boy because literally people couldn't think of a real entrepreneur.
- [77:03]
A
Not Lord Sugar.
- [77:04]
C
Well maybe, maybe just after the Apprentice, Lord Sugar got his way in, you know, even so. But like it's not a long list.
- [77:11]
A
That is a buccaneering man though.
- [77:14]
C
Well yeah, got to be careful of the definition of buccaneer actually. I think probably mercantile is probably a fair one but there has to be a cultural change and then a system change to get there. So my vision of Britain in 30, 40 years time is still the great institutions that we have providing the infrastructure of risk taking. Come to universities, do your research where you get funded in some way, can be a great health care system, which means you don't have to worry about like in America, if I lose my job and try something else, I don't have any health care. Well, we've got, we could have once again the world's greatest health care system that's serving everyone, including us. But you've got your safety net, you've got a good welfare state which is highly efficient, providing that safety net so you can take risks and you fix pensions, public capital, whatever else you need to provide risk capital. You've got your platform ready for this new world and then we get this, you know, cultural nudge to say, you know, maybe you don't need to go to university, maybe you can drop out halfway through. Maybe you can get a job here for a year or two to get your apprenticeship in. Or with this vision, you're going to go off and start something yourself.
- [78:25]
A
How do you, how do you think though about the fact that, I mean, what, the Greens are the most popular party among the youth? There seems to be actually. We are in the process of cultural pro cultural reaction against exactly that attitude.
- [78:42]
C
Yeah. I find it really disheartening that so many young people don't see opportunity in their future such that they are therefore rationally saying, well, the only way I'll have an opportunity is if we take stuff from everyone else. Because it's not a long term solution and it's not a very British one. I will say that. So there has to be a strong counter narrative and not just narrative. Right. Narrative is a starting point, counter offer to people. But I am heartened by things like LFG movement seems to do does genuinely cut through to people and say that sounds like a good thing. I'm heartened by the number of people, by the way, even if they say they're going to vote Green, ask them, did you have a depop store when you were 14? Were you selling sneakers and flipping them at school? Are you playing around with side hustles? Do you one day want to launch your own thing? That part of the culture is actually getting better. And I think if you talk to young people now, entrepreneurship is seen as more of a path. There are more people willing to experiment and you know, the generation above and lots of my peers, not kids, I'm an adult now, will say things like, you know, my daughter, she wants to be a YouTuber when she grows up. Well, great. What the hell? That's amazing that she wants to like take on a subject, do something herself, build something herself. In the same way that, you know, I think having a Saturday job 30 years ago, 20 years ago was a really great thing as well. Like bloody try something great. Fantastic. Use the tools. And so we have to have a bit of a culture shift, we have to have a bit of an offer shift and then we have to have some radical policies. I haven't fully costed it, but one of the things I wrote a book
- [80:25]
A
about, we're actually against costing.
- [80:27]
C
All right, okay, good. The OBR mandate of the state needs to probably Be a little more lenient. I'm all about risk capital, very hard to price that. But I wrote a book about called Startup Century a couple of years ago, which is about this rising wave of entrepreneurship. And some states were going to squash it and some states were going to embrace it and how that might happen. But one of the simple ideas in there was to say, you know, you can go to university and you can borrow 50 grand, let's say roughly, you know, some short courses longer than others from the state, and pay it back if you make money over time. Rather than giving everyone that offer or give everyone that offer, rather than saying you only get it if you go to university, why don't you also say, well, instead of going to university, you can use some of that start business. We're going to give you a no recourse loan, which is venture capital, by the way. If it works, you pay it back. If you don't, if it doesn't work, you've had a good shot, right? And say to people, look, you're going to get this business funding. You used to be able to get it from your local bank, but local banks don't do that anymore. You used to be able to get it from, you know, the angels in your community or whoever, some wealthy individuals, well, you know, now half of them moving to Dubai or they're worried about taxes. You know, there's sort of, you know, there may be some issue with getting that kind of funding. So why shouldn't the state say, well, you know what, we only get paid back half of the money we give to people through student loans. Like 50% of the loans won't ever be paid back in full because people won't earn enough over their lifetime to pay back in full. So we know it's high risk capital anyway and we know it's going to some people who fundamentally don't benefit from it because we can see that in the earnings data already. So let's open it up and say, come to us and say, why you want to do an English degree? That's a really good idea. Here's the money going to your English degree. Or come to us and say, actually, I really want to explore this new use of radar for cricket. Here's 50k to go and do that. It's a small thing, but it might start to change some of both the offer to people in terms of the fundamental resources they can get, but also just signal actually have a go and get that right. I don't know if the maths works out, but it might make small change to that more entrepreneurial future that we'll see in 30 years time.
- [82:38]
B
Well, personally, I can speak to the buccaneering potential of the English literature graduate.
- [82:42]
C
Yes, There we go. Yeah. Hey, look
- [82:47]
B
where I could have been.
- [82:49]
A
Tom's return on going and drinking a lot and then memorizing enough lines of poetry to kind of woo various other Fae undergraduates, I think, if only.
- [83:01]
B
Well, we've reached top of the hour, James, so we should let you go. But thank you very much for coming on the show. We wish you the best of luck at SOV AI.
- [83:07]
C
Thanks for having me.
- [83:08]
A
Where, James, if they want your book,
- [83:10]
C
Starry Star Century, it's on Amazon, Waterstones, Dorn Books, all the other places. And you can, if you want, download the audiobook. Although sadly, because Bloomsbury, my publisher, they had an American read it. So if you want the full Anglo futuristic version of my book, you have to just listen to this.
- [83:27]
A
Get your. Get your friend read it.
- [83:29]
C
Yeah, yeah, exactly.
- [83:30]
A
And anything else that you want people
- [83:32]
C
to read, no, go on the Sovereign AI website. Read what we're doing.
- [83:35]
A
Very nice.
- [83:36]
B
And if you're creating a world leading
- [83:37]
C
company, come pictures at Balderson Capital, follow me on Twitter or X rather. I'm Ameswise, because I got in there very early. And yeah, if you're building a world leading company, we're here to provide that risk capital. Get you going.
- [83:52]
B
First off, James, thank you for coming on the show.
- [83:54]
A
Thank you, listeners and goodbye.