Going for Growth: Can We Make Britain Wealthy Again? (Matt Clifford)

20:09Amol Rajan:Pensions, how would you go about fixing pensions? So I think we have a lot of the right ideas. So, you know, the Mansion House Accords, which are about trying to increase the share of UK pension assets into unlisted equities, into private companies. These are the right things. It's just going incredibly slowly. And I think this comes to the pathology that to me is behind your very first question about stagnation, which is why is it so slow? Well, I think we have this very, very challenging problem in our country, which is it feels to me like we've made a Faustian pact where we try and protect ourselves from any downside at the cost of any upside.

20:48Are you talking about planning? Well, I actually think it's the pathology that runs across planning, across infrastructure, across energy, almost everything. We've got to a position where, and these are very old ideas in political science. This is not radical, but we see the person that loses out from any change as a real person whose voice must be heard. But we see the people who lose out from non-change, who lose out from stagnation as statistics that can be ignored. That's so interesting. And so whenever anything is a problem, you know, like, obviously, if you want to build a house that overlooks my back garden, that affects me a lot.

21:23I'm going to make a loud noise about it.

21:24Amol Rajan:And you can get on the local TV bulletin doing it. How can you put on the TV the person, the 15-year-old who's going to have to work in a box room instead of having a proper bedroom because there aren't enough houses? But the builders and plumbers and suppliers who don't get to make that house, they're just a statistic. They're just a statistic. And you take that and you can extrapolate that over nearly everything. So, you know, when we talk about pension reform, of course, there are valid objections to a radical upheaval in our pension scheme. But we have the seen versus unseen. The people that want to stop change are very visible.

21:54They're concentrated. You can put them on TV. The people who benefit from change are diffuse. They're not organised. Why would they be? They don't know each other. And so we just we've now got to this position where we're a vitocracy. Like it is very, very easy to say no to things in this country. It's incredibly hard to say yes to things. And that starts, you know, by the way, in number 10. My experience working in number 10 was a place where it's so hard to make things happen because everyone knows someone who will object.

22:21Amol Rajan:Is that the problem? I'm really interested in that because, you know, as you know, I cover this stuff for a living. People sort of say the trouble with number 10 is you've got huge turnover, not necessarily bad people, but there's a level of incompetence. Well, I don't think it starts with this number 10. We're talking about a 20-year problem, right? So I'm very, a bit like your opening observation, I'm loathe to say it's this person, it's this, you know, this personality floor. I think this is structural. I think it's that we've got to the point where, not to blame your colleagues in the media, but anyone who is upset about anything is a hearing.

22:53Yeah, gets a hearing. But I really think this problem of how do we make the voices heard of the unseen people who lose out from stagnation, I think that's a huge one. There is no political constituency for growth. I got told, I remember very memorably by a very senior political figure who I won't name, that growth is deeply unpopular. We mustn't talk about it. People don't think growth is for them. And it's funny how we now want to talk about cost of living all the time. Quite right, we have an affordability crisis in this country. But the affordability crisis isn't actually about prices. It's not actually about inflation.

23:27The inflation that we've had in this country, apart from in 2022, is not historically unusual. The problem is waste stagnation. Of course we have an affordability problem. No one's had a pay rise.

23:35Amol Rajan:I'm not suggesting it was Keir Starmer who you're on WhatsApp's video with who said that to you. But Keir Starmer would say that he, you know, he said growth was his number one mission. Well, if you have a number one mission, then that means you have to make real hard trade offs. And some people are going to lose out. Exactly. And, you know, it is not a comment about this prime minister or any prime minister. But, you know, the truth is your priorities are what you are willing to hold constant while trading off other things. So if growth is our number one priority, then we have to be willing to upset people whose number one priority is not growth.

24:11You cannot actually have growth without people objecting. And one of the reasons I object when people say like, oh, you know, growth is for the rich, which, you know, and there's some people that, you know, as you mentioned, I gave a speech on this last year. And I think I'm a very self-critical person. I'm obsessed about the negative responses to it. And one of us is like, this is just trickle down economics. This is just like Liz Truss with better PR or whatever. It's like, no, actually, some of the people who will object, who would object most to a truly pro-growth program are the currently rich.

24:42They're the people that benefit from, I mean, planning lawyers, people who make their income through, you know, renting out artificially restricted supply of housing. Like this is not rich versus poor. It's about extraction versus creation.

24:57Amol Rajan:you gave the speech you're referring to i think was at by the way went hugely viral and i watched it and i watched the entirety of it i actually searched that and read the speech in which you talked about britain's scientific heritage and you mentioned a huge number of uh truly great inventors pioneers engineers i note that you took pains to make sure you covered all bases so you span centuries you talked about women you talked about men and this was i think for a group called looking for growth that's right which casts itself as an impartial non-party political organisation, which I think is why you, who do the same, were associated with it.

25:30Amol Rajan:But it is also true to say that on some social media platforms, anyone who wants to champion growth or fight decline, I think is quite often associated with a kind of nationalist concern about the collapse of Western civilisation and therefore the fight of decline. Just why don't you, in this safe space for radical ideas, why don't you say why you gave that speech and what your argument was? The reason I gave the speech is that my most radical belief is I do believe this can still be the richest country in the world per capita. I think actually the funny thing is, if you're an alien coming to earth and looking at the world, you'd say the weirdest country in the world economically is the UK.

26:06And the reason it's weird is that if you look at a textbook, where does economic growth come from? You have the basic layer of, got to get the rule of war, right? Okay, fine. But then you get to the intermediate layers of infrastructure, et cetera. The icing on the cake is like frontier industries, like really, you know, difficult knowledge work. The UK has all the hard bits, right? We have the most incredible frontier industries in the world. We're just really messing up the middle layers. We're just not building housing. We're not building infrastructure. We're not building, you know, energy generation.

26:34And so my belief is like, this is an extraordinary country. It's a country that actually built the modern world. I think my more controversial is, I also think it is a country that on net has actually has an extraordinary moral legacy in the world that I really believe in. And I've been very lucky in my life. I built a business that turned out to be way more successful and valuable than I expected. And so sort of at the age of, in my late 30s, suddenly was like, well, I kind of have the opportunity to be able to kind of do public service. I did a couple of stints in number 10 for, first for Rishi Sunak and then for Keir Starmer, and came out of that both kind of energized by what's possible in government, but also pretty depressed by, you know, like where we actually are.

27:18And so I wanted to give that speech to try and tell a story for a group of people who I think are dispositionally inclined to the narrative, but don't hear it. That like, actually, we can't, this is possible.

27:30Amol Rajan:This is true. But to what extent do you think that one of the defining challenges of today is the conflict, and it is a conflict, not just an inconsistency or incongruity, but the conflict between the speed of technological change and innovation on the one hand and the slowness of modern states and bureaucracies on the other? I think that is an issue, but I think it's overblown. And I've been very lucky to be part of two, what I would call government startups. One is ARIA, where I'm the chair, the Advanced Research and Invention Agency, an agency for funding and managing breakthrough research.

28:03And the other was the AI Security Institute, which was set up in 2023 to build state capacity and understanding the scale and power of AI systems. And I've invested in hundreds, maybe now a thousand technology companies. And I would truly say that both AC, the AI Security Institute and ARIA move at least the pace of the best startups that I've invested in. They have extraordinary leaders in public sector leaders who would absolutely smash it in the private sector. They have the same talent density of the best startups I've been involved with. Now, they're doing things that they're doing new things.

28:41It's not like we, you know, in either of those cases went in and like fixed the NHS. I'm not claiming that like all the public sector can move at this pace. But I have seen these, you know, what we would call in my world existence proofs that actually there's no reason why the state has to be slow.

28:54Amol Rajan:But on some of the other areas that you highlight, we've got huge challenges. One of the central arguments of people who, for instance, would be dispositionally inclined to support looking for growth is this is a country that struggles to build stuff and struggles to build stuff because of planning and it struggles to build stuff because of planning because we have this as you say this great interest in all the people that say no who have a kind of veto and we don't have this kind of let's go and do it there's a side issue here which is that you know people say well the chinese state is you know the politburo of the chinese communist party is mostly engineers whereas here we're run by lawyers quite literally in the case of keir starmer but we don't build stuff if you look at size well see if you look at hs2 if you look at the uh lower thames gateway and on and on it goes that's one area where you'd say the state is just not able to get cartsware stuff done so i think there's two different things here that we have to like decouple and address so there's something about state capacity and which is what you're touching on i'm actually pretty optimistic about state capacity i think when we know what we want to do and what trade-offs we're willing to make we can actually go really fast and you know i think we see that in for example in the parts of the state i just mentioned but actually lots of things that I have nothing to do with, you know, I think in national security, I think in our special forces, I think in, you know, certainly in parts of the NHS, we have amazing state capacity.

30:09I think the thing you're actually talking about is how do we resolve trade-offs? And, you know, you talk about SizeWellC, I think this is right. You know, the environmental impact report for SizeWellC was 44 ,000 pages long. We don't have to write 40 ,000 page environmental impact reports. No other country in the world does that. You know, that's, no, you could say that's a capacity issue. I think it's actually like, are we willing to take the hit on this other stuff that isn't building? That's actually a political question. That's not a state capacity question. And I think the reveal preference of successive governments is, no, we're not willing to take the hit.

30:44There is a constituency that wants us to write 44 ,000 page reports, obviously. Planning lawyers. Now, are we willing to take the hit of saying, actually, that doesn't make anyone safer? We've got very into theatre in this country. So I think one of the symptoms that you see a lot is that we play whack-a-mole with bad outcomes. So whenever anything bad happens, we write a very carefully tailored new rule to address the symptom.

31:08Amol Rajan:But sometimes it's necessary. I mean, post-Grenfell, I mean, what was going on with cladding was abhorrent. And some people say, well, if that slows down future building developments, it avoids a Grenfell. Obviously, that's the... But this is a great example because Grenfell was truly horrific. I think it was like one of the most, maybe the most viscerally painful example of failure. in sort of our lifetimes, or maybe it's just recency-ish. That was nearly 10 years ago. I mean, truly horrific and kind of unforgivable. But this is exactly what I mean, is our response to it doesn't actually get to the root causes.

31:44It says, well, here are some of the surface-level phenomena that looked bad, so we'll address those. But what the response to Grenfell has done is added£100 ,000 to the building cost of every house in this country. So the government has a target of 1.5 million new homes, right every year i think the 24 25 is not going well it's 200 000 now i'm not i'm not blaming that all on the on the regulatory response but but that you know when i speak to people who really know housing they are all horrified by what happened at grenfell they all believe that was it was totally appropriate that we had a response and they all think that the response was both failed to address the actual um the root cause problems but at absolutely enormous costs to one of the biggest crises that we have in this country.

32:28Amol Rajan:So other than following all of the recommendations of your AI report, which the government did, how do we stop failing in that sort of way when it comes to artificial intelligence and make this country the home of... I mean, it's not going to be the home of artificial intelligence. That might be on the west coast of California. You did this report. In fairness to Keir Starmer, he accepted all of your recommendations, didn't he? He was great. Yeah, I mean, honestly, I was so encouraged by... His and his team, you know, it's funny having worked for Rishi Sunak as a non-partisan, non-political appointee and then worked for Keir Starmer as a non-partisan, non-political appointee.

33:05They're very different people. But I actually think to their great credit, they both actually saw that this was an issue that even though, especially at the time, was not in the headlines every day, was important that we seize. my sort of self-review on the AI action plan is that all the things that the government could do very directly it's done actually pretty well the things that have been hardest are all the things that touch on the things we've been talking about for the last 10 minutes like what well so like one of the flagship policies of the action plan is AI growth zones which is the idea that we need we need to have domestic compute infrastructure in the UK we need the data centers that will run these AI models.

33:47We'll have as a slight side note, I think it's always very tempting and there's definitely a view in Treasury like we don't actually need domestic compute. We can just buy compute on the open market internationally. And that's wrong because? Well, it really reminds me, and to his credit, he admits this is one of his big mistakes, but there's that famous clip of Nick Clegg in 2011 explaining why we're not going to build nuclear, which is basically like, it won't come online until 2022. Every generation pays the cost of what the previous generation refused to build. And it's always tempting to believe that all the major inputs to your economy are these market goods that you can just buy at international prices on the open market right until you get a Ukraine and energy costs kind of go up.

34:22My view is that AI inference, in other words, using these models is going to become maybe not quite as big as energy, but like one of the most important inputs to our economy over the next 10 years. Can we buy it on the open international market? Yeah, most of the time, right up until you can't. And then you really, really, really wish you had domestic compute. And so I really believe that building this out is very important to the Prime Minister's great credit. You know, he went hard on this. If you go watch his speech announcing the action plan, he talks about this. He's given multiple interventions on it.

34:52Why haven't we moved as fast as I would have liked? Not because of him not backing it. He did back it. All the things that we've been talking about. The other problem we have, as you know very well, is we have the highest electricity prices in Europe. So on electricity alone, if you're an international investor and you have a choice of, do you want to build a frontier AI data center in the UK or in France?

35:13Amol Rajan:I spoke to an AI giant and someone who had just been in to see Whitehall. And he said he just had an hour and a half with the Treasury. And then say, you've got to invest here. You're an Englishman. You've got to invest here. You might work for a company headquartered in California, but you've got to invest here. And he said, guys, I can get electricity in France for about a third of the price. So they took your recommendations. They backed it. Some of these things are going to be very, very difficult to do. What's the big thing where, yes, you've got frustrations. frustrations but what's the big thing where you'd say to the prime minister and say to whoever's in charge it could be a reform government in a few years time could be a Tory government could be weird coalition who knows what's the thing to do now to make sure Britain wins so I do think the single biggest thing is adoption um what I mean by that is I really worry that the UK is in this knife-edge position where if you go read the government's industrial strategy whatever you think about it, it's very clear that the place where our growth comes from, the industries where our growth come from, are industries that will be transformed by AI.

36:18It's life sciences, it's financial services, it's professional services, it's the creative industries. These are all industries where AI, I think everyone agrees, is going to change them completely. And so our entire economy is highly, highly geared on whether those industries adopt AI rapidly or not. We are in a position where if they do, I think we could see enormous economic tailwinds, huge growth from those industries adopting. If they don't, they're also all industries that have international competitors. And so you can imagine a world where the city of London booms if it adopts or shrinks if it doesn't, because there'll be other countries.

36:53If you look at our track record of tech adoption in the UK, unfortunately, at least in recent years in digital tech, it's pretty poor. We're actually slower tech adopters than most of the rest of the OECD. And so my big message is slightly contradicting our earlier thing about concentration of power. If you look at where do the benefits from general purpose technologies accrue historically, yes, there are definitely benefits to being the people that build it. But actually, by far, most of the benefits come to those who adopt it. Yes, it mattered to invent the steam engine. It mattered way more to build out the railroads and connect every town.

37:27The big opportunity for the UK is to be the best AI adopter in the world. And that, I think, is where I would like to see just enormous amounts of energy as we start to see that actually you can now really do things with this.

37:46Amol Rajan:I really hope you're enjoying this conversation with Matt Clifford. What a fascinating man he is. If you are, please do subscribe to this podcast on BBC Sounds. That way you won't miss future episodes. And if you do subscribe, make sure you've got your push notifications turned on. and that way you will get an alert whenever we publish a new episode and you will never miss out. Right, back to Matt Clifford. What does this mean for workers? There's a lot of people who get in touch with us who kind of want a kind of practical takeaway for what they need to do to prepare for this world. And I said that one of the concerns about AI, and by the way, I say this as a technophile and someone who's extraordinarily excited and I use it on a daily basis and think of it as Promethean.

38:28Amol Rajan:I don't like the word miraculous because miracles don't happen. But anyway, it's incredible what you can achieve. One concern about AI is ownership, excessive concentration of ownership. As you say, that's in a sense downstream the fact that we've moved away from being an ownership society. And that's thinking about investment and pensions, all that sort of stuff. What about if AI creates a huge amount of wealth, but it actually eliminates some jobs? Now, I should just preempt what you're about to say by saying two things. One is, obviously, we don't know which jobs. You just can't know, right?

38:54Amol Rajan:Because that would be the future. And if you knew the future, you'd be even richer. But in terms of the jobs that AI is likely to come for, the best guess we've had, based on evidence of what the jobs are participating at the moment, are what you can think of as entry-level white-collar jobs. That's what Dario Amadei, the boss of Anthropic, said on this very podcast. I really want people to hear this clearly. Which jobs do you sense are most vulnerable in the coming years? I think we need to be really, really clear here what we mean. Because I think there's two different things that people mean here.

39:23Sometimes people talk about this, meaning, is your current job in your firm right now at risk? And sometimes we're talking about the macro level. Will this category of jobs exist? I think that there is going to be a lot of disruption. And I think for sure we're going to see individual firms fail, change the shape of their workforce, have to change the shape of their workforce a lot. And that will cause real problems and we need to be ready for that. And I don't think we're quite ready for the level of disruption that's coming on that. So that's point one. I'm actually quite optimistic at the level of will there be jobs for entry level white collar workers?

40:05Almost certainly, you know, to me in any conceivable world. It might mean that our labour market has to be a lot more flexible because, you know, individual firms will rise and fall more quickly. But I don't buy the narrative that there'll be no jobs. I do think things will have to change a lot. I have to say I'm not fully persuaded. So I started my career in professional services. You know, it's a big part of this country's economy. There is this narrative that, oh, you know, like we don't need so many. I wasn't a lawyer, thank goodness. But, you know, we don't need so many junior lawyers to read through everything.

40:37Paralegals. Yeah, yeah, exactly. And then how do we train the next generation? I feel this is a slightly weird, like, failure of imagination. like there's like so much of professional services is actually about things that may always be human certainly will be human for a long time things like trust authenticity legitimacy um these things are not going to go away and the way that if we were trying to design from first principles a way to learn how to convey those things it wouldn't be by reading thousands of pages of legal docs that's just how we do it today yeah i'm very skeptical that that's going

41:07Amol Rajan:to remain but also it's useful to think of this uh in the way that mark andreessen who's a legendary Silicon Valley thinker and investor thinks of this founder of a firm called Andriton Horowitz, where he says that the atomic unit of work is not the job. The atomic unit of work is the task. And the concept of bundles or bundling, which people in the news industry will be familiar with because we've just lived through this grand unbundling. It's quite useful here. If your job is made up 80 % of tasks which are repeatable cognitive endeavours, so processing information or solving some mathematical quandary, those tasks might be done more efficiently by an AI that doesn't take holidays and doesn't have sick leave.

41:49Amol Rajan:But you might, in the remaining 20%, be able to have a very successful job with that remaining 20%, which is more to do with human skills, intuition or dependent on trust or empathy. That kind of can take up the rest of your job. So unbundling jobs into tasks that are repeatable cognitive endeavors and those which aren't is quite useful. But in terms of the jobs that are going to be created, have we seen any evidence of that yet? This is, I think, one of the hardest arguments to have because it is totally understandable why people want the burden of proof to be on people who are less worried about the jobs apocalypse.

42:29But of course, the way I always think about this is I'm the eldest of four. I think all four of us, you know, I think often about my conversations around the dinner table with my parents being like, what are you going to do when we grow up? I think all four of us are doing jobs that either didn't exist or would have been totally inconceivable to us, you know, in the late 90s. I really feel old when I say that. And so it's like, in some ways, yeah, I get the impulse to want to be like, what is the job that is there? But I think the burden of proof is on, you know, if we look at the history of productivity increasing technologies, they've been very, very good for humans.

43:07I would argue that nearly everything that's good that's happened to our species has come from productivity improving technologies.

43:14Amol Rajan:And we've tended to find jobs to go with them in the past. Yeah. And I also think there's lots of things wrong with this analogy, but I think it is worth dwelling on for a moment is to think about what can we learn from the very first jobs that got completely dominated by AI. And it's a sort of niche one, but I think it's an important one, which is chess. So in 1997, you know, as is well known, deep blue. And so, you know, if we were having this debate then with the same sort of set of assumptions, I think it would have been very tempting to say, well, now we know that computers can completely dominate humans in chess.

43:52Chess is going to shrink. There will be no role for humans. Chess today is way bigger than it was in 1997. Way bigger. Chess play. Not well. Sadly, not well enough to play with anyone who can play well. I think my eight-year-old is starting to be a lot better than me. But, you know, what's really interesting as well is that as far as I know, even though the best human chess players, you know, have sponsors today, I'm not sure even, you know, no one has ever paid to watch a computer play a computer in chess, even though the computers are way better than Magnus Carlsen. Now, chess is niche. It's my, you know, I'm not trying to pretend the whole economy is going to look like that.

44:30But I think it does show that it's very hard to build good intuitions about what happens when a job gets fully automated.

44:35Amol Rajan:So one related concern, which I just want to reflect on, then we'll get into some of the extraordinary positives, is people worry about what's going to happen to tax receipts. And the reason they do that is because this is partly gets into technical stuff about how investment income, and if we're all in the future, going to have to just be invested as much younger, which might be a good thing, where investment income is generally taxed at a very different level to wages or labour income, and that might need looking at. But there's another thing, which is if you do lose in the short term through creative destruction, lots and lots of jobs, you lose potentially quite a lot of tax receipts.

45:07Amol Rajan:And just in terms of this sort of worry that people have about a kind of new AI aristocracy, where ownership is concentrated, where work is no longer being created in the same way. If you lose tax receipts, then it's poorer people who tend to rely more on public services who potentially have the most to lose at. So I agree with this. This is why AI sovereignty is important. What do you mean by AI sovereignty? I will explain and I'll try and be really precise. But let me kind of give you the direct answer to your previous question. So for me, part of the reason the sort of modern liberal state exists is to help us manage these transitions, which are really disruptive.

45:43And I think we have collectively decided that we want to provide, we want to dampen the variance for individuals as they go through this. Ideally, we don't dampen the upside at the moment. I think we are dampened into that too much. But we're trying to like prevent the shocks that come through these changes, damaging the lives of individuals too much. And that's why we have a welfare state. And, you know, this is like an eight year project, more than that, like to the start of the 20th century. Right. So that works really well when you have technological change that creates a lot of wealth in the country that then means that we have more tax receipts that we can then choose how to distribute.

46:19Amol Rajan:That was what Wilson's White Heat of Technology was about. That was domestic goods, white goods, gave women the chance to enter the workforce and at the same time created more tax receipts. Exactly. So we grow the pie. And yes, there are people that lose out from that. And our job is not to therefore not grow the pie, but to use some of that pie to cushion the negative impacts from the people that have lost out. I think what's really challenging about AI, and I think there's a very visceral way to bring this to life, which is to look at what was the first speech that J.D. Vance gave abroad after he was elected.

46:52A lot of people think it was the Munich Security Conference where he told Europe off for everything. He was at the Paris AI conference. He was at the Paris AI conference, and he had this very memorable line, which is, the future of AI is American models running on American chips. Well, that's fine, except it does suggest, well, it's fine for the US, what it suggests is that actually what we're about to go through might not have the same political economy characteristics as what I've just described. It could be that the pie grows globally, but far too little of it grows in our country. And so we actually don't have the fiscal headroom to be able to cushion the people that lose out from it.

47:27So for me, this is a big advert for two things. One, for AI adoption. We need to make sure that the benefits are happening here. If we actually have the nightmare scenario where we don't adopt and our growth industries are outcompeted by faster adopting counterparts in other countries, we lose tax receipts at exactly the time when we need more headroom in order to cushion the blow for people that lose out. And it points to AI sovereignty. What is AI sovereignty? AI sovereignty is simply the idea that this is a strategically important industry where we're going to need some domestic companies that are winners.

48:01Amol Rajan:And that we can be masters of our own destiny. It's not the same thing as self-sufficiency. That is impossible, may not even be desirable. I think sometimes people immediately, to me, that's a straw man. Like, obviously, we're not going to be self-sufficient. The US is not self-sufficient in AI. It's dependent on all sorts of companies all over the world, including many in Europe. I think even China probably will not be self-sufficient in AI. So it's not self-sufficiency. But what it is saying is, how do we have enough important AI companies in this country for three things to happen? One, that we have a seat at the table as international governance questions on AI are going on.

48:39Two, that we actually have direct bargaining chips as we negotiate with our friends and allies, including the US, on these questions. I've been on the other side of the table with both administrations and realising how our lack of chips to play, figuratively not, is a real problem. And third, so that we have a safety net of having enough growth, enough value accruing here, that we actually have these options, these political options around how do we distribute the gains.

49:10Amol Rajan:I have two specific questions and I want you to talk to two specific constituencies, please, about them. The first is that 1.1 million needs. A lot of them listen to this podcast. What do they need to do to, and most of them want to get off of welfare, to get into work. And, you know, we're talking about young people here, by the way, 18 to 24 should be in the prime of their lives. And they could be living in places, you know, which are attached to big cities. It could be rural poverty, which is an issue we want to get into in this podcast in a big way. Can you just talk directly to them for a second?

49:40Amol Rajan:What do they need to do to win in this AI economy? me i'm not pretending that it's easy i don't want to be glib i don't want to be one of those tech pros that's like oh just become an ai guy you'll be fine but uh i do think that there is something about this technology despite everything that we've said about concentration of power that is actually really liberating and they they actually do give you superpowers in your pocket you know honestly people like you and me who let's be honest you know we might be you know talented it in some dimensions, but one of the things that has been helpful for you and me, I suspect, is just sort of a verbal fluency, an ability to talk about ideas, memorize things.

50:21AI really democratizes that. Actually, Claude is a lot better than I am at most of that stuff. Anyone can use that. There is a free version of Claude. There is a free version of ChatGPT. Again, I'm not glibly saying, if you do that, you can be a successful BBC journalist. What I am saying is the best way to give yourself a real shot in this era is to be an enthusiastic adopter of ai i think you have to go down the rabbit hole it is shocking to me and to be fair this is mainly sort of in invert commas elites um rather than rather than neats for example who tried it once in 2022 didn't think it was very good and so decided it's it's not for them i think that is

51:03Amol Rajan:not a plausible position to take if i may say in case i get accused of um discriminating against old people i would say that actually particularly i would say that to older people who are worried about technology as well there are lots of free tutorials just go on to youtube or other video search engines you can find the stuff i say this uh talking to my mum who does who's the most amazing woman ever lived and does kind of voluntary classes for people where you know in her community groups in south london she'll sort of show people this technology and they think this is absolutely incredible i can help it book me a holiday i can sort of uh work out where i want to go i can save time i could do my weekly shopping list okay well actually let me just build on that quickly because I think one way in which we've built structural unfairness into this country is actually making the state extraordinarily difficult to navigate for the least well-off people.

51:48It is incredibly hard actually to navigate universal credit. It is incredibly hard to work out what you are actually owed. It's incredibly hard to work out what form you have to fill in order to undo some structural unfairness. And actually, AI here is a great equalizer. This is actually one of the things that, you know, Chagipiti, Claude, Gemini is really good at. And I do think there is this, not to sound too conspiratorial, because I don't think it's a conspiracy, but I think, you know, the country is run by people who are really good at reading forms, writing forms, filling out forms, knowing, you know, like how to play the bureaucratic game.

52:26I think some people call them the lanyard classes. I don't want to get too, but I really think that this is one of the ways in which AI can be a really great equal.

52:35Amol Rajan:So the second constituency, I'd love it if you could talk to directly, because to some of these people, you are a hero, is the people where you grew up alongside in Bradford. Yeah. You're from Bradford. And one of the huge concerns people have about the modern economy is that, to quote a remarkable young man called Joe Seddon, who has sat there just a few weeks ago, runs a tech platform for social mobility called Zero Gravity. And he said that in Britain, geography is destiny. we have this isn't on you but we have a particular an extraordinary for historical reasons problem with regional inequality and there are some people and by the way lots of people in Silicon Valley who have a very very clear message about this people like Chamath Palahapitaya who's an influential VC you probably know and he says the honest advice to young people is go to where the action is at and I say this as a South Londoner I'm very conscious that people might not like me saying it but I've realised as I've grown up and I've taken a very close interest in social mobility and life chances and meritocracy, whatever you want to call it the extraordinary weight of opportunity that comes with living in Britain's capital city for young people living in Bradford is AI this kind of great equaliser because you've got these superpowers in your pocket or is the message bluntly, truthfully, honestly that if you want to maximise your career chances you've got to leave?

53:55Well, I did grow up in Bradford and I went back there in October and actually in some ways marveled at how much it's changed for the better. I do think it's a place where actually economic opportunity is still very limited, as you say. Here's what I would say. I think if I like intellectually, honestly, with the country as it is today, it is really hard to escape the economic logic of what economists would call aggregation. You know, it's really, really hard to avoid the fact that, yes, economic opportunity in this country is extremely concentrated in London. Does that mean you should leave? Maybe.

54:38But to me, the reasons for that are also really addressable. You know, so part of the reason that you would need to leave Bradford rather than just sort of live there and take a national view is because we have really poor transport infrastructure to Bradford. We've not built local transport infrastructure there for a long time. So, I mean, the truth is if all you want to do is increase your income, then yeah, I mean, like, it's just an economic fact that moving to London is probably one of the best ways to do that. But that's not the aspiration we should have for this country. There's a difference between, like, what is the correct advice today and what normatively should we want.

55:15And what we should want is for our, you know, cities that are, you know, between 100 ,000 and a million people to be amazing places to live. And by the way, as you doubtless know, if you try and decompose what are all the reasons why the UK hasn't got the same productivity as Germany, for example, it's actually that our second, third, fourth, fifth, sixth cities are just nowhere near as productive as that. That's the solvable problem. And it's back to all the things we've talked about.

55:39Amol Rajan:Right, so Ed Balls has done some remarkable work. Harvard, where he actually went in the 1980s, he and some co-authors at King's College London have done some amazing work on regional inequality in the UK, which is worse than other countries, has gotten worse over the last four decades, whereas in other countries it's gotten closer. And the two things that he always lights on is, A, Britain doesn't have that second tier of megacities. Obviously, he's got second cities, Birmingham, Leeds, Manchester, Liverpool, and so on, but they just don't compete economically with London. And he puts a huge amount of weight in his analysis, as you do, on transport links.

56:08Amol Rajan:I want to just mention one other huge externality of what we're talking about, a negative consequence. The people who listen to this podcast who want us to get Kate Raworth on, and I hope she will come on soon, might also say that the environmental and ecological cost of AI is A, unfathomably big, and B, is so big that we ought not to not bother with it, but we ought not to try and make ourselves world leaders. And they might marshal for their argument the fact that data centers are not just power hungry, they're kind of cities in themselves. How do you approach that problem, the environmental cost of this technology, which in your analysis is our best hope of being rich and great again?

56:49I think there's a, I want to start by saying, obviously, one of the other great challenges of our time is that for a long, long time now, we've allowed the negative externalities of industrialization to run wild. And that's why we have a big problem in climate change. And I'm definitely not, and the reason I'm stressing this is because of what I'm about to say, I'm definitely not a climate change denier. It's really important. It is one of the great challenges of our time. However, the point of civilization is not to reduce our energy usage. I think degrowth as an ideology is deeply inhumane. I think it misunderstands the purpose of what we as a species are trying to do.

57:36So when people say the ecological impact of AI is too high, I would say, compared to what? Compared to what? It's actually one of the good things about capitalism is that in theory, the price system allows us to allocate resources to their most valuable use. And actually, if AI allows us to stop doing a bunch of other things that are way more energy intensive, that could be exactly the right use. That said, let me also like, I know some people, going back to your question about what should needs do, I know especially young people who are totally paranoid that they shouldn't use AI because of their individual environmental impact from using it.

58:15This is not something you need to worry about. I'll give you an example. If you eat meat, stop eating meat way before you eat it. Stop using Claude or Chachy. So this great stat, which I have double-checked is true. If you have two people, one of whom is vegan and uses AI every minute of every day without ever sleeping, and another who eats the average amount of meat that a UK citizen eats, which has the worst environmental impact, But the second by far, oh, sorry, it never uses AI. The second by far. You just cannot make a dent by stopping using AI in your own environmental footprint relative to just the things that really matter.

58:59If you boil a kettle, that's probably more energy usage than a heavy individual user of AI in a day. Maybe not the most extreme, but a typical heavy user.

59:10Amol Rajan:Really? How do you reconcile that with the warnings we hear about how power hungry these data centers are? Well, but remember, these are where we're training these models. And then we get to spread that cost over absolutely enormous numbers of people. So ChatGBT alone has about a billion weekly active users. And so, yeah, it's enormous. But that's because the world is enormous. And it's one of the great hopes of what you're doing at ARIA, which is part of what you do now. But it's an incredibly exciting thing. For those who don't know, it's based or it's inspired by an American program called the Defense Advanced Research Projects Agency.

59:46Amol Rajan:DARPA, that's it, which was this kind of it was of an era with moonshots and exciting science fiction about the future. It was it had a pioneering spirit which tried to envision an amazing future. And ARIA is trying to do that in the UK. What an inbox you must have. I mean, it must be amazing getting emails on a Monday morning from people saying, I'm just trying to do X. And the X might be, well, I think we're two years from a vaccine for cancer. Yeah. Or we're kind of, you know, quite soon we're going to be able to use AI to massively reduce the environmental impact of tidal energy. What's Matt Clifford most excited about right now?

1:00:25So you're absolutely right. I've had this enormous privilege of first spending 15 years entrepreneurs first, meeting entrepreneurs that have these crazy ideas. And then now the last four years as chair of ARIA, seeing the scientific frontier. And as you say, one of the programs that we've invested real money into is this program on energy efficient computation. So energy efficient AI. It's certainly true. The thing I would absolutely concede is that we have not optimized the way we do computation today to minimize environmental impact. Interestingly, though, and again, this is sort of one of the joys of capitalism, because the energy cost of AI for individual companies is becoming one of their biggest expenses.

1:01:05There are enormous economic incentives to reduce that, which is great. And one of the things that that will come through is new kinds of chips. And we're funding the development of some of them. By the way, another great way that we could have more AI sovereignty in the UK is for some of those chips to be designed and built in the UK. And we've got great examples of that already happening. Actually, though, I think for me in many ways, the stuff that I get most excited about ARIA is the stuff that I know least about. So I do know something about AI. But, you know, ARIA is, as you said, it's broad based breakthrough science and it's pre-commercial.

1:01:38So we're typically not, it's not a VC. It's typically trying to find the very best researchers either in universities, sometimes in companies, in fact, increasingly often in companies to come together. And we like orchestrate these research programs that end up, we hope, having these huge impacts. Now, you mentioned DARPA, which you rightly say ARIA is inspired by. To give listeners a sense of what success would look like, no pressure, DARPA's research directly led to the creation of the internet, GPS and mRNA vaccines. So it's a high bar. And that's what we want at ARIA. We want to find things that will create trillion-pound industries of the future and have comparable impact to that.

1:02:20Amol Rajan:And solve big problems. Huge problems. It's a really interesting thing that, you know, if you follow some of the people in this space and have done for a long time, even before it was kind of fashionable to, it's really amazing watching them talk about science in a way that was almost pre-political. And I want to give you two examples of that. The first is I really strongly recommend everyone listening to this podcast looks online at a speech given by Dominic Cummings in 2014. I can't remember if this was before or after David Cameron called him a career psychopath or where he was on that particular trajectory.

1:02:47Amol Rajan:But he gives this long speech about the civil service and how Whitehall is failing. But he sets out a vision. I'm not saying I necessarily agree with it, but it sets out a vision, which is the optimistic Cummings vision of Britain, which is he says that Britain should be the best place in the world to do science and technology. And he was obviously instrumental in coming up with the idea for ARIA. And I just think that there's a sort of there's a sort of romance there, which sometimes gets lost in the noise around Dominic Cummings. But there's another guy called Elon Musk. I have a slightly strange habit of spending the wee hours when I should be rehearsing for the Today programme looking up videos of things that people in Silicon Valley said 25, 30 years ago.

1:03:22Amol Rajan:And there's a young man called Elon Musk, South African via Canada, who says he wants to solve humanity's biggest problems. And that was where Tesla started. And I'm not saying I endorse everything that Elon Musk has done. But there is this feeling you should take the biggest possible problem, you know, cancer, and fix it. Aging and fix it. And it sort of requires that mad vision and then mad people, perhaps, to try and get there. And the thing is, some of these things are actually possibly happening. And cancer vaccines is a good example. It could be that we're just a couple of years away. And it could be that ARIA plays an instrumental role.

1:03:58Amol Rajan:And that would be a British story, wouldn't it? Yeah. I mean, this is the thing that retains the core of my optimism about this country, is that actually, if you look at, it's such an implausible story that this tiny, cold, wet island in the North Atlantic should have shaped the world in the way that it has. And really the reason that it has, I think dates back a very long time. You know, England had a, before it was Britain, even had a very distinct set of institutions and culture, you know, kind of long before the Industrial Revolution. You mean common law, Magna Carta? Common law, yeah, limited government, but also even just things like property rights, things like the invention of trusts, which were sort of forebears of limited companies.

1:04:45You know, there's all these sort of like institutional innovations that come like way before the Industrial Revolution. And, you know, if you read books about like why did it happen here, a lot of it dates back a very, very long time. And I think at the heart of it is the idea that actually is what I would call like permissionlessness. The idea that actually this country has respected for at least a thousand years the idea that even very ordinary people should be allowed to try stuff. Now, obviously, there's been enormous limitations on that. I'm not trying to pretend that we haven't had enormous amounts of oppression of different individuals and groups.

1:05:15But this idea of permissionlessness, the idea that you can try stuff, actually has very, very deep roots in this country. And I think if you read, I think one of the best periods to read about is like the sort of birth of the scientific revolution. And, you know, one of the most fascinating. Mid 17th century. Yeah, exactly. Exactly. So take maybe the most famous scientist of that era, Isaac Newton. What's so amazing about him is world historical genius, obviously. I mean, everyone knows something about Newton. Was he?

1:05:42Amol Rajan:Just had an apple bop his head. Exactly. Super easy. Was he that cool Newton? But at the same time, he's also really into alchemy and astrology. Yeah. Why do I think this is important? Because I think it's really easy to imagine that, oh, progress in science and technology comes from these really buttoned down, kind of prestigious institutions that kind of turn in the handle and great things. It comes from weirdness. It comes from people who are right at the fringes of both the brilliant and the bizarre. And actually, I think we're a place that can tolerate that really, really well. And at our best, we can really celebrate that.

1:06:19And so, yeah, for me, ARIA, I don't think we have anyone doing astrology right now, but ARIA, I hope, is a place where all over the world people want to come and guide research programs that are almost too weird for for any anywhere else and i think if we do that right let's be clear and i want to make i always say this i say this when i'm in front of the select committee i say this around with the secretary of state we're going to have a lot of failure like most things that are weird are really weird because

1:06:46Amol Rajan:they're wrong um but you know i'll tell you one of the issues might be that a lot of people might trust matt clifford to do it a lot of people say oh he's a cool guy i had him on this amazing He's a podcast called Radical, which everyone else I know listens to. And I kind of, yeah, it's cool that he's in charge of this potential thing. Elon Musk and Sam Altman, not so sure. Jeff Bezos, not so sure. Demis Hassabis, fantastic guy, Nobel Prize winner. But the two concerns people have is, A, an extraordinary amount of power is in very, very, very few hands. A, and B, is Demis Hassabis, founder or co-founder of DeepMind, which was subsequently bought by Google, would say and has said that we have exactly the wrong dynamic for this technology at the moment you know when we had previous extraordinary technologies which by the way ai is different every other technology because it has its own ideas it comes up makes its own decisions and could generate new knowledge but we always said slow down and collaborate whereas now we've got this crazy race dynamic with the wrong people in charge well i i think there is a very important idea here about power and democracy and one of the things i really like about aria to the earlier a conversation is, it's state capacity.

1:07:52It's the UK government paying for this and empowering this. And, you know, I've complained a lot about the failure of the British state in this show, but actually an extraordinary thing that Parliament passed the ARIA Act in a bipartisan way. And the ARIA Act, if you go read it, it's one of the shortest acts of Parliament. It basically says ARIA can do what it likes and we won't interfere. An amazing thing for Parliament to have done. And I think it has real consequences for the distribution of power in these technologies. So I'm not complacent about that. And I do think that AI has implications for power that are really important.

1:08:25But that's why I've spent so much the last two or three years trying to get the state to take AI seriously. Because I don't think the right thing is to try and stop Sam Altman or Demis or Dario or even Elon Musk doing these things. But I think there is a thing about how do we make sure that as a country that can make decisions for us all, we have a seat at the table. And that I think is just really important.

1:08:47Amol Rajan:How did you end up doing this for a life? Working in professional services, having grown up in Bradford, 15 years later, you're the guy working cross-party trying to make Britain the place that solves the cancer vaccine. Do you ever sort of wake up in the morning and think, my life is pretty cool, but I didn't foresee this? Well, definitely didn't foresee it. I am optimistic to a fault. Was the move into politics one you felt at all queasy about? As in government? Government, yeah. No, it wasn't. It happened by accident and it was something that I felt. I think a lot of it did come out actually of the dynamic that you've just talked about, which is this stuff is really important.

1:09:26It's really happening. Do we actually want it to be entirely determined by a very small number of private companies in the United States? I'm actually a big Atlanticist. I spent a bunch of my life in the US. I love the US. I'm not one of those people who has this instinctive anti-Americanism. But I actually think the US is at its best when it has to work with others. And you talked about Demis already. For me, he's one of the great British heroes of our time. Not just because he's this extraordinary scientist and entrepreneur. Actually, I think the most consequential thing he did was saying when DeepMind was bought by Google, I'm staying here.

1:10:01And as a result, we have this incredible ecosystem there. So to me, it was like, well, look, I'm in a fortunate position where I can do this. And I think if we get it right, it will have enormous positive impact. And, you know, I do think that I don't want to oversell the story. I had a very I grew up. My dad, as I already mentioned, was a social worker. My mom was a teaching assistant. You know, I went to the local state comprehensive. But, you know, I had a great childhood. I definitely don't want to oversell the sort of misery of my roots. I had an incredibly happy childhood. But I was always conscious of like, can someone like me from where I am, can I actually break through?

1:10:44Can I play at this level? I think I came to London like twice before I was 18. You know, it was just like this mystery, far away, slightly alien place. I think entrepreneurship is one of the great liberators of ambition. That's why I've spent most of my career trying to help others do that. But as I got closer to some of the limitations of that, much of which we've talked about on this podcast today, you know, like all the ways in which actually we put a ceiling on what people can achieve. Then actually I increasingly saw government as one of the vehicles to change that. And I do worry, you know, it's worth being really honest.

1:11:22Like one of the things I've had a lot of criticism for over the last year or so is Entrepreneurs First, the firm that I co-founded with my co-founder Alice Bentink. One of the things we say to entrepreneurs is you should probably go raise your seed round in Silicon Valley. You should probably go raise your funding there. People say, how do you square being a British patriot with this? And my answer is this is kind of classic British establishment thinking that's the wrong way around. The question isn't, how do we stop our most ambitious people going to the US? It's like, how do we make this the place that they want to stay?

1:11:53How do we unlock those pension assets so that we have the venture capital? Well, I think unless people like me say, well, actually public service in one form or another is one of the unlocks on that, then we'll always just be complaining that like, oh, why do our best companies sell? Why do our best people move?

1:12:07Amol Rajan:Two final brief questions. Number one, what makes an effective entrepreneur? Well, I would say the thing I've probably changed my mind on most over the last 15 years is when we started EF, I thought it was basically like we just need to get the smartest people. And if they're like really, really smart, they'll figure out all the problems along the way and they'll be fine. I think it's certainly true that we don't want to fund non-smart people. But I think I way overrated raw IQ. I think the most important thing by far is just relentlessness. it is extraordinary how the very best founders will be rejected again and again and again they'll be rejected by investors they'll be rejected by customers they'll be rejected by people they want to hire and unless you have this pain tolerance to just take the nose you'll fail and you know the funny thing is i don't want to keep talking about demis but i really recommend if you haven't read the recent biography yeah um the thinking machine uh infinity machine One of the things that's interesting, everyone knows that Demis is a genius.

1:13:10I think many people, unless they've read this book, do not know. He's been an entrepreneur since his teens, and he has taken so much pain. Final question is, what's next for Matt Clifford? I don't know yet. I announced a couple of weeks ago that after 15 years at EF, I'm becoming non-exec chair. I'll still be spending time there, but not full time. I want to do something big. I think I'm too young to retire. I think whatever I do next will be at the intersection of the two big themes we talked about today which is like how do we make this country great and how do we make AI go well for the UK and the world

1:13:43Amol Rajan:I'll tell you what you are doing next is you're answering our listener questions we've had rather a lot thank you so much for sticking around Matt Clifford it's a pleasure and a privilege thank you very much

1:13:59Amol Rajan:well spending an hour or a bit either side of that with matt clifford is like a massive massive jolt of energy and i can see why frankly whatever you think about his opinions his ideas his experience what he thinks of the world i can see why successive prime ministers on both sides of the house of commons wanted him in the room not only because he knows what he's talking about But he is an incredibly energizing force, deeply, deeply, deeply passionate about technology, but also about this country. It was just really interesting. It's a while since I've heard someone, A, so embody this desire to fight decline and to sort of go against the prevailing pessimism or anxiety of our zeitgeist.

1:14:38Amol Rajan:but also just really openly talk about love of country and patriotism and cast himself as a British patriot, not whilst ignoring some of the nastier periods of British history, but saying, as he put it, in the net, this has been a country which has done great things in creating the modern world. There are a couple of specific things, both on tech and the kind of politics and government side of things. And he's seen both that really stood out for me. One was on that sort of manner of government and the nature of politics. when he talked about how the people who, you know, we have to understand that all policies are trade-offs, that there are going to be losers, and that the losers are the people that make a lot of noise.

1:15:16Amol Rajan:We often see them as a real person who's likely to lose out and to suffer in some way. But we don't really grasp how the people who lose out from not building things or not doing something aren't real people. They're just stats. I thought that crystallised something quite powerful about how we, perhaps, you know, including us in the news industry, give too much weight, too much airtime to those who are obviously going to lose out and don't think about those who are going to lose out from not taking particular actions. And then the other thing on AI specifically, I mean, on the stuff where you can't know for certain how things are going to pan out because it's about the future.

1:15:50Amol Rajan:It was just a useful corrective, I think, that he talked about future jobs and the impact of AI on jobs, not as a jobs apocalypse, but more as a kind a form of productivity enhancement. Now, he's obviously thinking about it from the point of being an owner, an entrepreneur, rather than being, if you like, a daily worker who's potentially going to lose their jobs. But that idea that you should reframe it in terms of what you do with the gains from productivity and not in terms of the losses and displacement, though there will be big losses and displacement, was, again, a useful and pretty radical, I think, counterpoint to the prevailing mood of our times.

1:16:26Amol Rajan:God, that was fun, wasn't it?

1:16:31Amol Rajan:And that is it for this episode. To make sure you don't miss future conversations, you can subscribe to Radical with Amal Rajan on BBC Sounds. Please do that. And please feel free, if you happen to listen to us elsewhere, to give us a little five-star review and say something nice about us. Because if you do that, who knows, it might help others find our work as well. We'll be back same time next week. Thank you so much for your time.

1:17:21We'll be right back. style who's got the best beef. We yell across rooms, across blocks, across generations. Yes, nice earrings, Grammy. And now, we're doing it on our phones, too. Video calls from the L train, group chats that never shut up. Neighbors texting you the score before your TV catches up. Get a load of this guy. Ketchup. On a dog. Yeah, we've got thoughts. Because nothing here gets built without conversation. Nothing moves without connection. So, if you think you know AT &T, it might be time to rethink that because their network has improved like a lot which means chicagoans can keep speaking their mind with the top rated fiber internet in the city chicago won't stop talking our improved network backs it up at &t connecting changes everything at &t fiber limited availability in select areas top rated and fastest internet at &t fiber based on analysis by oocalypse pizza intelligence data 2h 2025