This Week in European Tech: Can the UK restore growth and stay competitive in AI?

25:20And it's out there. I mean, it's already out, right? It's out there. It's out there. So I think you can't do that. What you can do is you can hit the procurement line. You could say, just like, you know, we've had GPL, right? We've had folks that are unfamiliar with GPL and copy left. So open source, kind of the whole idea is that you have software that has an open source code that people can download the source code and modify it and embed it in their own technology, et cetera. And when you have open source source code, it's distributed under a licensing machine. And there are different licensing regimes.

25:50Some of them are very permissive, meaning the open source can be used for whatever you want. Some of them are more stringent. It may be the licensing regime says you can use it for your own private purposes, but if you use it commercially, you have to pay. And still others are saying, if you use this open source code for any commercial purposes, we own you. And that was sort of the idea of GPL and the whole copyright, but copyright movement that says, look, if people are being mean capitalists and using this free-spirited open source software for nefarious commercial purposes, we'll get them. And that's a problem, of course, for businesses.

26:26And so as an industry, we've been all over this to try to make sure that nobody uses the wrong open source code in their products. Well, could the US government do the same thing? Can they say, if you use a Kimi K free anywhere in your supply chain, can we then say that that means you will not be eligible to sell to the US government, for example? So we're not banning free speech. We're not banning the use of models. You just can't sell to us. We're going to cut off a huge part of the market. And of course, it's complicated because even if you as a business only sell to other private companies, well, they may be selling to government.

26:59So government has got a huge reach in that sense. And that is probably the way one would go if one wanted to do this. And could it happen? And if you open AI and if you're anthropic, that is, of course, what you're hoping for. because you're looking at you charting Fable at 50 bucks a million tokens and saying somebody could download some open source off the web running on their own servers and do it with an inference at a fraction. Interestingly, there's a point here because we're talking about 50 bucks versus 15 bucks, I think was the number. It's not as big as I thought.

27:33Lomax Ward:I thought I thought it was going to be a 10X. It's not as big as I thought. Well, so that is the cost to run it off moonshot servers. It'll be cheaper. where they'll have a different price in different places. However, a token is not a token because different models have different reasoning efficiency. Now, it used to be I asked my model a question, it'll come back, it'll do a single pass and then come back with an answer. But the powerful thing today is the reasoning where the model reasons with itself and thinks through different problems. And what we have seen is that Fable is actually quite an efficient model of reasoning and can get to answers with fewer tokens than a KMEK3 can.

28:10So a token is not just a token. And so expect lots more interesting evolution here. Now, I think the Chinese will eventually figure this out and become much more efficient because they build for efficiency and they'll keep hammering at it until they work it out. But right now, the gap is not as wide as one would be led to think. The Kim K3 weights are being released on the 27th of July, right? So Monday.

28:36Lomax Ward:Yes, yeah, Monday. Yeah. But Mads, we talked about this before, is that the Chinese have been thinking about basically stopping that, right? Stopping the companies from releasing the weights. It's been discussed. So if that happens, then you have a problem, right? If you're sitting in Europe. Potentially. I mean, I think in that situation, I think NVIDIA would clean up because as we talk about, they've already got their Nemo Tron. They've already got the open source model. They've been holding back from perfecting it because their customers are all the Frontier Labs. But they're all rolling their own silicon today.

29:15Google is rolling their own TPUs. Anthropic is using TPUs, right? Everybody is doing their own stuff. So what's holding NVIDIA back and saying, well, great, we will just up our efforts on the Neymar Tron and release an awesome open source model. What is the primary complement to a GPU? It's the model, right? if we all have free models what do we all need to run them gpus so i i i think this is what i could see happening if the chinese pull back so i actually don't think there's a whole lot for

29:47Lomax Ward:them to win here but we'll see any any notes on the hugging face hack mads i think it's super interesting i mean it's pr uh no no it's hard to say where the line goes i do think it was great Dan, as you said on episode 87, you said, this is going to happen. It happened more quickly than some people had thought. And what was interesting in all of this is hugging face. What do they do when you're hacked? Well, you turn to your trusted AI to try and work out what's going on. And what is your trusted AI? Well, that's the good one you have from Anthropic, right? But that's Godrails. So as they were being hacked, as they turned to Fable to try and work out what's going on and defend themselves, oops, the classifier came out and said, well, you know, you're using it for some cyber defense stuff, you're not authorized to do that.

30:37So they had to go to the Chinese models to defend themselves from the open AI attack.

30:43Lomax Ward:Oh my goodness, how the web weaves. Oh my goodness. And also who knows what's going on behind the scenes as well. Anything on the circular financing, the AMD, Anthropic deal, anything in there that caught your eye that was worth talking about? It doesn't excite me that much. I think, you know, everybody's so excited talking about Lucent and Nortel. I think this is completely different. It's not like they are putting money towards buying kit that isn't being used and sits in a dark, dusty room somewhere. It's been around for years. I mean, the auto industry has been at this for years. It's not like a novel way of doing business.

31:18It doesn't excite me that much. I think we should watch out for does the revenue keep growing? Can the firms keep monetizing? And if they can, this is fine. If they can't, that's an issue. But then the issue is not the vendor financing. The issue is the monetization at the front end.

31:32Lomax Ward:yeah yeah yeah so everyone's going to keep dancing while the music is playing baby uh anything else in there i mean the the mccrory piece it feels like a nothing burger but is there anything from either of you on on the mccrory ai jobs thing you know dario who is mccrory's boss has been saying that the ai is coming for your job yeah he's been saying the opposite for years yeah and and now mccrory's like yeah well it's the numbers that aren't actually supporting that And I think that's what we're seeing in general. So there is a huge amount of steermongering. And is it plausible that there will be displacement?

32:05Absolutely. But we're just not seeing it yet.

32:08Lomax Ward:No, I think the interesting thing to watch will be the entry-level jobs piece. That's the thing that kind of worries me the most personally. But it's just not sizable enough in the metrics yet. So we'll see how that one pans out. Right, the big meat and potatoes, gentlemen. man, I want to know if you woke up tomorrow as the UK's eighth prime minister in 10 years, what would you do? So within the context of investing, startups, innovation, VC, blah, blah, blah, all the good stuff that we know and do at our place in the world. So just to containerize it a bit before we kind of go off into random directions, what would you do?

32:49Lomax Ward:I've got my top five, but I want to hear from you guys. But first off, I want to just have a quick review on what Burnham has done so far and a quick look at who John Healy is, who is the new UK Chancellor. So, so far, we've seen from Andy Burnham, he's scrapped digital ID, he's removed VAT on domestic and electricity bills, which I think people may not know that electricity bills is split into three. And on the domestic side, we're not horrific. But on the industrial side, the cost of industrial electricity is, I think, the highest possibly in all developed economies. He's also reinstated a two pound bus fee cap.

33:28Lomax Ward:Lovely man. Business rates cuts for pubs and clubs. Pledged$340 million to end rough sleeping. Killed the Department of Science, Innovation and Technology and installed a new minister for AI, Kanishka Narayan. And I feel like I've missed something. What else have I missed that he's done on these crowd pleaser policy launches? What have I missed? What have I missed, guys? Anything? I mean, I hope you're not counting Killing Decent as a crowd pleaser. No, no, no. I'm sorry. I'll retract. Although, to be fair, every new government tweaks acronyms or changes. Some of the actual services stay the same.

34:11Lomax Ward:they just get absorbed elsewhere or an acronym gets changed. So I'm not convinced this is such a big deal. And maybe one of you will have some notes on this. But yeah, I mean, I'm not putting that in the crowd, please, Apotno. No, no, no. I mean, so what they've done is they've asked one Department of Technology and AI we had, they put the remit under the Department of Business and then back into central government, into the cabinet office, split it up, anatomized it. And I think Kanishka was already AI minister. Now he's going to be allowed to attend the cabinet meetings, but not as a cabinet minister.

34:44So he can sort of sit in the corner and maybe be asked questions if others need to know what AI is and how it works. It feels like a step backwards to me, I must say. I thought this was a pretty good idea.

34:56Lomax Ward:Before we get into some more of the detail on this, I just want to have a quick check-in on who John Healy is, our new chancellor. So it wasn't Mahmood or Miliband, as many people had suggested. So who is he? Well, he's politically safe. He's a steady hand. He's the ex-Defense Minister, who I'm sure many will remember for resigning under the Starmer government for not being allocated enough capital to do what he needed as Defence Minister. Born in Wakefield, a very soft-spoken gentleman. He's 66, married for 33 years. I couldn't find much on his views on the economy, but I saw a lot of reports saying, oh, he's, you know, he's unknown.

35:36Lomax Ward:But actually, he was the PPS to Chancellor Gordon Brown, his private secretary from 1999 to 2001, then economic secretary to the Treasury 2002 to five, then financial secretary to the Treasury 2005 to seven. So he's not he's not a nothing burger here. So he's involved in fiscal policy, regulation, economic planning under both Blair and Brown. Is it a defence hawk? Will he be a fiscal one? What's going to happen with guilt? What will his first moves be? I don't know where we start about. Maybe, Mads, you can set a bit more of a scene and then we can get into some of the good stuff. Well, so I think the problem the government is inheriting is, and is a big problem, is we are all relatively poor.

36:17We're all£10 ,000 a year poorer than we would have been if the economy had continued to grow the way it did from the late 70s to 2008. because if we continue to grow on that trend, GDP per head would be 10K higher than it is today. And that's massive. You know, you think about 10 ,000 pounds for the average family. That's just huge. And that's genuinely agreed, you know, across the board, right? That's both the TUC and economists on the right. So everybody's looking at this issue and that impacts living standard, impacts tax take, it impacts everything. If you compare the UK to the US, you've got Mississippi, one of the poorest, least functioning states in the US.

36:59And our output per head is just above Mississippi. And if you take out London, it's below. So the UK is now poorer than the worst performing American state. Make of that what you will. The typical polls, Poland, is going to match UK GDP per head within a decade on current trend. that's really not good. And you look at some of the implications, you look at what it's done to debt. I think Lomax, you've been digging in a little bit into that. It's a debt interest today. It's£110 billion. It's the third biggest budget line. It's more than all of education. We spent just servicing debt. Welfare is killing us.

37:39£330 billion last year, above£400 billion by 2031. And the issue is that today, welfare, the spend we have on welfare, exceed the entire income tax take. It's mental. It's absolutely not. It is mental, right? And so it is clear that this is unsustainable. And what I fear is that the government will put their head in the sand and say, well, we can't possibly take money from everybody. They're poor, they're weak, we have to support them. But the challenge we have is unless we make some changes, we will all become a whole lot poorer. And the current trajectory is just not good. Now, how big is the hole we have to pluck?

38:23Well, sometimes people talk about the 34 billion pound current deficit that happens every year. That doesn't sound like a too frightening number given the size of our GDP. But there's also an 82 billion pound investment line. And it is sometimes said that you can exclude investments because they'll go towards funding stuff that'll be good and yield good things in the future. But it challenges some of the stuff you count as investments are things like HS2 that will never yield a positive NPV on the current track. There's a ton of stuff we're stuffing in there. We're buying new stuff for the armed forces.

38:59We call that investment. Guys, it's expenditure, right? It is important. We should do it. But it's not going to yield a positive ROI. It is not what you would classify as an investment for any right-thinking investor. GDP is currently what? So debt to GDP is currently 95%. By, I think, 2060, it's forecast to be 300 % at the current rate of deficit and investments that you just alluded to, Matt. That is crazy. It is. They're crazy numbers. and let's unpack some of the ways we're thinking about investment and why we are where we are. If you look at HS2, because this is sort of this NPV, I used NPV, net present value.

39:40It's the, you know, if any of you listeners are not sort of, you know, investors, investment professionals, when you make an investment, you have to put some money up front. Let's say we invest$100 million in something and then you're going to get some money back over time from the revenue, the profit you're going to get from that asset. And what you try and work out is, of course, if I'm investing 100 million today and we're going to get 100 million back in 10 years, well, today's money is worth more than the money will be in 10 years. So that's a bad investment. And what I'm trying to work out is the money that I'm getting in the future, if I discount that, if I value that at today's value, is that going to be worth more than the investment I'm making?

40:20And if an investment is worth more than what I'm putting in, then it's a good investment. So if I'm investing 100 million a day, but the thing is worth 110 million in today's money, that's a good thing. We should do it. We should just keep doing those sorts of things. Now, the challenge we have is HS2 today, it looks like it could exceed 600 million pounds per mile of track. Is that saving the newts? What is that? It is the newts. It's the bats. It's the fish. and fish. I mean, it's all that stuff. And so people are saying, oh, it's just because we wanted to make it really fast and people were getting too ambitious.

41:02Guys, come on. In Spain, they built a track that was slightly slower, not at 600 million per mile, but at 25 million per mile. Okay, France managed 46 million per mile. Japan, 50 million per mile. And that is even with half of the new expansion they did is tunneled. Okay? So lots of expenses. We can fill mountains in the UK. We don't have that excuse. Like we just, you run out of excuses. And why is this happening? It's because of all the rules, all the regulations, all the things we are putting in place to make it difficult to do stuff. Where does it come from? It's this notion of the risk-free society we've been trying to create over the last 30 years.

41:48Brick by brick, we have been putting in place barriers to business and to entrepreneurship. And it's all for reasonable reasons. In 1991, Robert Maxwell, he died at sea. And it was found that 450 million pounds were missing from the mirror pension scheme. So in fact, he'd been robbing his pension members. That was a bad thing. So of course, we acted. We put in place the Pension Act, minimum funding, FRS 17. Lots of things were done to then protect pension savers, including developing new ways to look at how much money should go into pensions and how they should be valued. One of the key things there is if I'm running a pension fund and I'm investing, I have to value that as if it was invested in GILTS.

42:34Even if I buy tech stocks, even if I invest in venture capital, I still have to project the value of that in the future as if it was invested in GILTS. So lo and behold, everybody invests in guilt. People don't invest in inequities and venture capital and the stuff that could actually yield good returns. That's really bad for us. That's why pension funds used to have more than 50 % in equities in the UK and now it's 4%, right? You've got Grenfell, 2017. It was terrible. 72 people died. So of course we acted. We put in place the Safety Act, the Gateway Approvals, the second staircases. Now the bill has arrived now.

43:11In 2015, in London, 33 new homes were built. Last year, 5 ,000. Okay, an 84 % collapse because we made it so difficult, almost impossible to build that nothing gets done. Another example, 2008, the banks nearly took down the economy. So of course we acted. Core equity was increased from 2 % to 7%. Retail was ring-fenced. The bill has come back. SME lending, lending to small and medium businesses, has decreased from 10 % of GDP to 6 % to 7%. All right, so less money for the young growth companies, the small companies that need to create jobs and grow. And on and on and on. Brick by brick, we have been building this risk-free society.

44:02This is what's killing growth. And unless we reverse it, we can't fix these issues we have around debt and our public finances.

44:10Lomax Ward:You are on one, my man. I'm seeing you in my mind's eye in public office Standing next to Andy Giving it like it needs to be given Good, lovely points Let's talk about what we would do So Lomax, what would I've got top five things I would do If I was, I guess, Chancellor Not necessarily Andy Burnham But what would you do, Lomax? You need to move away from risk-free society As Mad said Which cuts across Andy Burnham's happy husband home and every happy heart and every postcode, right? We just simply can't underwrite everything at the state level. So the state needs to shrink. We also need to maintain fiscal credibility at the same time, right?

44:53So we can't necessarily go straight into a, you know, cut, cut all taxes, kind of end up in a kind of potential Liz trust territory. So there's that trade-off. I would though, in no particular order, the welfare state needs to be addressed, right? We need to, we need to bring the cost down. We need to encourage more investment into technology in our society, particularly in opening up the ability for pension capital to flow in, to growth capital. We need to make an attractive place for foreign people to invest. So there does need to be a look at things like corporation tax, some tax incentives for people who are actually investing significant capital.

45:34Clearly, we need to have a cutting of the a red tape, a deregulation drive. I mean, look at this data center build out, right? We're trapped by two things. We're trapped by the energy problem. So we need to bring that down. We need to deregulate nuclear, for example, and make it easier and quicker and cheaper to build nuclear plants and to cheapen electricity. Prices, particularly industrial ones, at the same time, we need to make it easier to build data centers. So that also is a deregulation play. You know, look at what happened since the Americans came over last year and with Donald Trump, and we had all these big announcements.

46:13Basically, all of that's been unwound because everyone found it's too difficult to do that in the UK. You know, the Stargate Open AI Project in the UK got canceled. So it's a combination of, and you know, the problem is, it's very easy for me to sit here and be like, well, we should, you know, we should run austerity and we should cut welfare. And at the same time, we should increase spending in these critical areas. It's very easy to do that theoretically, but this is a theoretical exercise. So sure, let's go for it.

46:39Lomax Ward:Yeah, no, it's tough. I think the one thing I would do above all else is to bring back the feeling of being proud to be British and energetically instilling, I don't know, at schools, through national service, through clubs, through a multitude of ways, just feeling a sense of national pride. I think there's an energy piece that we're missing in the UK where we don't feel together. We don't feel like we need to get up and get out and do and build and be innovative and learn. I'm being a bit unfair here to a degree, but I think there's a, I've always felt a sense of a lacking sense of pride. Dan, I wonder whether that's caught the full horse.

47:30Okay, so I actually think you can find sort of roughly 30 billion that you need to balance the current account on the fiscal position. That immediately will start to bring interest rates down. That will stimulate lending. It will stimulate house building. You start to unwind some of the regulatory issues we've created that will further boost growth. The economy starts to move again. And I think that leads to all the things you're talking about. I think we've got plenty to be proud of culturally in the UK. I was proud when you saw the England player on Saturday, right? What a great match, what a great run they had at this World Cup.

48:09Yeah, we got to the semis.

48:10Lomax Ward:Yeah, we should be proud of that. I just don't think that was, I don't think we were. You don't think so? I don't. I think we are lacking a confidence, an energy, a drive, an ambition. And I think we are stuck in a very southeast bubble. I've not for a while now, but I've spent a lot of time in the Midlands, up north, in the northeast especially. And it's just a different country. And there is there is it's hard. It's hard work. It's hard trodden. There's a there's a there's a there's a difficulty that we just don't see down in London and the southeast. And I think I understand why Burnham is now doing the political piece and, you know, not working from London number 10, but working from the northern number 10.

48:55Lomax Ward:But I think we just I think if energetically we start from that, from that, I'm going to get up and do. I feel confident. I feel positive. I feel energized. And we're going to go. I think that if you if to me, that's the starting point. If you look at it a bit like a startup or any business, you know, you have this like trade off or two things you need to get right. You need to get the narrative right, and then you need to get the fundamentals right. And by the fundamentals, Mads is totally right. Everything really does flow from the growth rate, from the economy. So everyone clearly will be feeling great about that if they're making money, they're feeling richer, if there's a lot of opportunity for them.

49:34This is my point.

49:35Lomax Ward:What starts that? At the same time, you need narrative, like a startup, a pre-seed or seed stage. right or you need i mean you don't have the data in a way narrative like you say we know starmer failed on narrative right and and so actually tony blair was amazing at narrative and he wasn't so bad at fundamentals either fundamentals of course is there's a little bit out of your hands as a politician right because the uk is so it's so like you know look at you know what happened to gordon brown you know suddenly the gfc happens on your watch you know you can argue it's his fault but also probably the fault of what happened a lot in the us right so there is an element on the fundamental side that's a bit out of your control but the narrative is way more on your in your control but i do think mads is basically right like for this for this all to work like you need you need the economy to be people to be feeling richer and for their people to feel like there's opportunity right but but while you don't have that or in the gap before you start getting it you need the narrative which is what we have not you'd be such a lovely moderator like max you should come and join our meetings you'd be such a lovely monitor if you think about the narrative also is like if you've got seven you know i was looking at a table of you know in the uk we often laughed at laughed at the changes in um in prime ministers in european countries now the uk's like had the probably highest or second highest turnover apart from bulgaria or romania in the last 10 years so you know that is a lack of narrative you had this trust going let's go this way and then you know she did the lettuce thing and then you have um you know what i mean like no one's had enough time and then the one dude who came in with a 320 whatever majority seats he had starmer who had the opportunity to bring in some narrative failed because they didn't have it yeah not a strong suit he started to put some reasonable fundamentals in place but the sad thing is now even like for us in technology and ai actually in the ai opportunities plan like there's actually some quite good things in there but now burnham's probably gonna just rip them up or ignore them lots of lots of tears no riz as they say i've got i'm going to quickly run through my five and then mads to see if anything that you would add or poke in the eye so i would rule out a wealth tax day one declare war on industry industrial energy costs yeah fix planning get reform with teeth restructure taxes rather than raise them and force pension capital into uk assets i would force it personally what would you do mads or what would you challenge out of what we've spoken about so far yeah not surprisingly agree with you on most of these bits i wouldn't force capital i think you can't confiscate people's pension savings and tell them what to do you can put incentives in place i think that's far more powerful yeah you've always been big on that they're massive incentives around pension savings today right it's costing us more than 50 billion pounds a year we're talking about fiscal you know gaps i mean that that is a huge amount of money we're spending Yeah, you nailed it in one.

52:31Yeah, direct some of that towards what you invest in. Say, okay, great. You want to save in your pension to invest in max seven stocks. They probably can do fine. They probably don't need a subsidy. Fine. Maybe you get less of a tax incentive there. You're going to invest in UK assets, you have company, startups, whatever. Great. You're going to get the full tilt. I think that's pretty simple. And I think that in itself could solve a huge thing. Then there's also the regulatory piece, right? How are pension funds measured? What constitutes a good fiduciary, a good administrator, a good investor in a pension fund?

53:06Is it somebody who just minimizes fees and minimizes volatility? Or is it somebody who maximizes long-term returns? But clearly, it must be the latter. So there's some of those things that have to change. Like, what the hell is Burnham going to do vis-a-vis AI and technology? That's the kind of thing that we care about in this podcast.

53:26Lomax Ward:he gets it he's not anti i mean all the reports that i've read he's not anti i i just worry as i've i've mentioned here before i do think ai is a bit of a vote loser for for many many but he's a political maneuverer i'm sure he is cognizant of the reality and and will know how to politically maneuver well yeah and if you well yeah no no he could do you know he could do the kind of cheap things that keep people happy, like ending rough sleeping costs 40 million. Look over here, look over here, look over here. Whoosh, over there. Yes. Well, I mean, you know, he obviously, I think he opened up licensing in the North Sea, right?

54:05So he's clearly doing things that... Oh, that was one of the things that I missed.

54:08Lomax Ward:Oh, God, that was one of the most important things. Don't! Yes, he's pledged to open up gas and oil in the North Sea. Right, let's move on. We are running out of time and I'm conscious that we might be starting to sound like grumpy middle-aged men. So let's move on. Mads, deal of the week, go. So two deals of the week. The last week, we were talking about the Palantir diaspora. This time, this week, it's Act Two. We were discussing Valarion last week, which is sort of the anti-Palantir. Palantir, NEA's first European defence check. The founder, X12 years at Palantir. This week, it's the same story running again, twice more.

54:45The alumni are fanning out. The American capital keeps writing the checks. So we have first up Cusp AI, which is a super interesting company that specializes in looking at using AI to develop new materials. The whole question is, you know, how do we find the new materials? We need an industry. And so the kind of the thing they're claiming is they're evaluating 300 trillion candidate structures with the new AI model. And that's led them to raising a$450 million Series B at$2.6 billion, led by Kleiner Perkins and NEA. So NEA that we remember from last round, co-founded by Chad Edwards, who was from Continuum, and Max Welling, who was the co-inventor of the variational autoencoder, which is the maths that sits underneath generative AI.

55:35So the UK government was in cusp, wasn't it? So they're in the round. Yep, they're in the round. It's lit by Americans, but it's very much in there. So I think that's interesting. The other one is so cool. Arrakis, for all Dune fans know that that's the place where spice flows. The name is really in the pitch. In June, the spice is kind of the most invaluable thing. In our world, it's data. And what they're about is to swap the spice for the industrial data. So it's procurement data, supply chain, project delivery, kind of all that stuff that's embedded in in a large enterprise it's the 30 million series a laid by blossoms well done blossom capital yeah they're building what seems like a really cool

56:20Lomax Ward:company so uh maxis do you have a deal of the week sir well i mean other than delivery hero the delivery heroes um story is ending with uber actually acquiring it delivery hero is a listed listed company scaled to 50 markets predominantly done i think has probably hit a ceiling yeah fair enough fair enough that was something that's the german right so that's 14 billion euros yeah let's see what comes recycling because right gents we have run way over thank you so much for your time lovely listeners and watchers we will catch you next week bye see you in the next one

57:07Thank you.