53. Is the UK keeping up with Quantum?

0:00Welcome to the Difference Engine, the show for tech founders, investors and innovators.

0:09what's coming up today paul we're asking if all you need is love for ball how is the swedish startup faring in the ai category competition i will also be donning our managerial boxing gloves to find out why when the game gets tough the tough thing different but first the uk government is splashing the cash on quantum are they finally taking it seriously Here in the UK, we've seen almost simultaneously the government preening itself at allotting 550 million for quantum development. That's 550 million of our British pounds. But simultaneously, Amazon has just announced that it's going to invest 40 billion in the UK economy, which makes us wonder, are we taking tech seriously or selling ourselves cheap?

1:07Right. So let's look at the face of it here. So 550 million, it's a big number. You don't find it around the back of many people's sofas, but 40 billion is a much bigger number. And the problem we've got is that most people have difficulty conceiving a million, nevermind a billion. So if we look at those two numbers, 550 million for quantum, half a billion, 40 billion from Amazon, 80 times the number. And let's even look at the 550 million that we're talking about here from the government. This is over the next four years. And the total will be 670 billion as part of its digital and technologies sector plan.

1:44Let's bear in mind, it may or may not be in power at the end of that time. Am I missing something? Or is this just a really massive defunding dressed up like a government reshuffle? Somehow is good news. Previously, we had 2.5 billion under the previous administration allotted, OK, we don't know how, for quantum. This new number of 550 billion breaks down as 125 million a year for the rest of the duration of this regime. That's just 5 % of the previously announced total. And then it's dripped out at 3 % of the total over the course of the next decade, if they're still in power. Somebody is taking the media for a massive ride.

2:24As a former journalists i feel we need to remind our friends in politics many of us can do elementary maths right so what are we to learn from this well this is nowhere near good enough right and if history is any guide most of this government money will go to quangos think tanks dare i say consultants and not to the engineers who build businesses now moving on to amazon it says and i have to believe it's going to invest 40 billion over the next three years including building four new warehouses and creating thousands of jobs god knows they need them in hull northampton and the east midlands the plan also includes uh expanding their london offices which i'm sure are very lush yet to be invited um upgrading delivery infrastructure and redeveloping brave film studios hammerhouse of horrors i think was that right like we don't have enough film studios being massively developed in the uk at the moment you ever tried to draw through shepperton at the moment i haven't doesn't this also cover the 8 billion they already announced.

3:22Right. So they're re-announcing, so 40 billion, which has been proudly touted by the government, 8 billion is already previously announced for data centers, which they need anyway to sell us stuff. And who are we to decry such a significant investment by the everything store? Yeah. I mean, this man's got weddings to pay for, right? Albeit in our continent, Jeff Bezos and Lauren, I guess Bezos, recently got married in Venice. And that, you know, Jonathan, that stuff costs money. Is there some, to use a sort of current phrase, a Zempicization, if that is indeed a word? Was it Retail Gazette claimed in September that 2024, there was 17 million on sales of more than 1 ,000 times?

4:072017 million of taxes on sales of 1 ,000 times that at 27 billion. Now, that's some Zempic-style tax slimming. Yes. Is it avoidance? Is that the word we use here? No, no, no, no. I would avoid saying that. I just think it's smart tax planning. Tax planning, that's the word I was looking for. According to Retail Gazette, that meant their effective rate of direct taxes was 3.45%, and who wouldn't like to be 3.45 % of their current size? Amazing. Okay, so this amounts to them choosing to put money into the UK rather than being compelled to put money into the UK. Is that what we're talking about here?

4:42Yeah, that's right. So what you avoid in tax, you claim as an investment. uh and uh anyone paying only 3.45 in direct taxes that could pay for a lot of wedding confetti but only if you bought it from a chinese e-commerce store you don't want to buy that from from anywhere else but you know is this is this a top down or a bottom-up approach you know what's actually going on here is it pushing us in any particular direction well it's clearly much more than the government itself is prepared to to spend um and we would be saying that you know as 8 billion of this 40 billion has already been spent.

5:16It's sort of double counting. That, though, would be to suggest that somehow low taxation rates and job promises were linked. That's a very cynical thing to say. Well, doesn't Amazon employ about 75 ,000 people in the UK or thereabouts? It's actually one of our biggest employers. And it would be churlish to say that the 550 million, jumping back to the quantum piece that the government's talking about uh that that is wasted because that race is very much on uh and recently the chinese uh announced that they had using a d-wave quantum rig they potentially earth-shatteringly or at least bankbustingly made a claim that they had a beaten encryption right right what was was this was this um claim debunked or is it still it was debunked and just to make the leak what's that got to do with quantum if you use a quantum computer the theory is that you can beat encryption which is of course what the whole internet and all banking etc is based on but the truth behind this claim it wasn't the full bifter not the rsa algorithm it was a 22-bit encryption algorithm the real rsa algorithm, I believe has something between 2048 and 4096 bits of encryption.

6:34So it's a whole other order of magnitude bigger, two orders of magnitude, in fact. But it proves that the quantum race between governments and private companies is raging. Which makes you think, is 550 million going to touch the sides? Yeah. So what do we do? Is this about bottom-up or does it have to be top-down? Yeah. It's the age-old questions. Experts in the field, Jason Sirocco is from Sectigo. He claims that we've got until 2030 to sort this out. Just around the corner. Which will take us to the end of this 550 million thing. So, you know, that scare is similar, he claims, to Y2K and Snigger, if you will.

7:15But that scare woke everybody up. So now's the time, I think, to get a little bit more serious than a meager 125 million pounds a year from the UK government. Yeah. So what can we really take away from all of this, in your opinion? I think they bring to the fore the critical importance of tech. So the dismal drip, drip, dripping of UK government funding into tech and this much vaunted job creation of big tech, which the same government sings for the rafters. You've got to take that with a pinch of salt. Tech is right now about to become once again a very high-stakes game and quantum computing is at the bleeding edge of tech.

7:54Yeah, so we're talking strategic national importance. Maybe, though, to your point about top-down or bottom-up, perhaps the best investors in tech, history might say, are private ones.

8:08Leave that. Why even bother wasting a meagre£550 million worth of taxpayers' money on foundational research? Just leave it to the big tech guys. Oh, wait, what the government's got left to do, what happens then? The big tech guys that exploit the foundational research, I think? It'll be the big tech guys who don't like paying tax, that's for sure. But maybe cutting these deals with big tech, which is a bit of a deal with the devil and taking reduced tax take for decades, will give you not just the tactical headlines, but the inward investment you need. It certainly seems the price that some regimes like the UK are willing to pay.

8:49I guess what we take away from all this is that some moonshots, because this is what ultimately it is. Like tech dominance in quantum computing, nuclear fusion energy, to another example, or as we've seen in the last 24 months, drones are far too important for a nation or society to lose. But abandoning leadership in tech is a choice. And if we choose to duck that choice, then we can have little say about how quantum computing or big tech chooses what it chooses to do, or more importantly, how it doles out its favours for the next five years and beyond.

9:46Think there are too many lookalike tech companies chasing the same markets? We do. We think it's because so few tech startups are able to build and lead their own categories. What matters most is to be different. Stop following and start building your own unassailable leadership position today. Working with the category design gurus at Categorical brings decades of differentiation expertise to your team. Book a consultation from our website today and we will send you our one pager detailing how to start designing your own highly differentiated category.

10:25you've got to learn to earn okay what are we learning today well i thought we'd actually think a little bit about some of the realities of building tech firms and becoming category contenders and even category leaders and uh you know we all know that the one thing you clearly need to do when leading a business is keep going but there are moments when you have to make a decision so i thought that we'd think about you know when the going gets tough how to think a little bit different about what you might do. So, you know, we do know the road to being a category contender, let alone a category leader, is never a smooth one because it's ultimately based on constant risk-taking.

11:09You know, the market may take longer than you expect it to get it. You know, your wonderful idea may not be apparent to other people. Development costs may be higher than you think. you may do some wrong hires that's not beyond the bounds of fantasy and of course as is so often the case for European firms cracking the US market may take a lot longer and be much more difficult than you expected so you can quickly find yourself in a situation where you're burning cash and rapidly running out of runway because of that your funders could be signaling that they're not going to do a follow-up round worse than that they could be possibly plotting to take the company from under your feet or some other dastardly move including exiting you from the firm that you've founded i refer you back to episode 43 let's say no more all right yeah i think you're right mining a winding road on the way to success lots of pivots these days etc and timing is absolutely everything for a wannabe category contender.

12:16I listened to Marc Andreessen of A16Z recently on his excellent A16Z podcast. And he talked about timing being everything and those who were slightly too late, as they thought, to market. Actually, in his experiencing, that was a good move, as opposed to those who had just thought they got early market dominance. In his view, Those who arrived at the market too late or thought it was too late are actually the ones that succeeded, which is an interesting turn, not what is traditionally thought. And generally, every great investor knows, and he knows because he knows this from Netscape, his own experience, that a silver medal in tech normally amounts to first loser.

13:04However, this news thesis seems to blow out or at least reframe the doctrine of finding early market fit as being a signifier of eventual category dominance. You need to fit, in his view, just as the market explodes and, interestingly enough, when the early leaders have already been declared. Now, suppose you don't get that exactly right, which, you know, maths would tell us is going to be the majority of people. If you're stuck in Europe and your financier has zero experience of running an actual tech firm, coddled in the comforting rows and columns of endless spreadsheets, any disquieting numbers may cause a sudden panic.

13:47Now, this panic often manifests itself as cut overhead now, now, now, this week, not this month or this quarter. By the way, a note to our US listeners, neither the timescale or overhead reduction with impact immediately can be actually possible thanks to European employment laws. So frankly, this extortion may be rather pointless. But anyway, the point is you're under pressure from your financiers to cut the burn. and whatever you cut it's likely to involve people so you know beware of any grand gestures after all the time effort and expense of existing people from a business is going to distract you from growing it so the idea of doing it this way could be a pyrrhic victory in itself absolutely and you need to be sure I noticed the there was some rumbles at dating site Bumble when the CEO in announcing big layoffs asked everybody to calm down which is just not going to happen right it wasn't calm down it wasn't it wasn't one of those it was um yeah which whatever it was it had the exact opposite effect as you might as you might think you know you spend every waking hour building a team you know and you've invested in that team heavily um which if you just cut them they're going to be gone forever um and their absence is going to impede your attempt to get the business back on track you know that is the very definition of a pyrrhic victory um you know i've been there and uh at that moment you know your world potentially comes crashing in it is time to think creatively um but there is an approach that can give you more breathing space you know keep the team together, actually expose those who aren't totally on the journey, which there's always a few, the Zs, and get you back on the growth track and on ultimately to a successful exit.

15:52And it's not actually about cutting. It's about concentrating, concentrating the remuneration, concentrating on surviving and thriving, and ultimately concentrating shareholding. um but you know but you've got to realize that if you're running a business in europe um your team could be making a choice between a tax-free payoff in the event that they let go it's tax yeah yeah sure but they might they might quite like that might fancy a little holiday um but of course they will have the challenge of finding a new job and of course go through the mental disruption that comes with such a such a move um and of course a severely reduced personal cash flow for some time.

16:33That's a big risk. So what we did was that we proposed that all employees receive 50 % of cash in their remuneration and 50 % of equity for the next quarter. And of course, the situation would be reviewed at the end of that quarter. You got a lead from the front. Senior team had deeper pockets, took deeper cuts, no cash at all during that period. I mean, And that absolutely shows commitment. It was a strong signal that the leadership team were committed to pulling through. Of course, what did they know? The spreadsheets, the numbers became nicer, instantly reducing the burn that so bothered the board.

17:18But the intent was there. It was keep the band together so that nobody felt forced out. you know and on the basis of doing that it was proposed that the board sign off on bridge funding to match that equity pool that had been created um ultimately now that is intent matched by action and you know it is risk but there's plenty of upside if it works so i would say that all of that sounds like a really good case study as to you know what was possible perhaps in the pre-ai age I think now we know that it's clear that AI is threatening a lot of jobs, a lot of tech jobs. You just look at the Microsoft layouts, et cetera.

18:02So we have to add in those effects. And we've seen how tools on the engineering side, like Replit, Cursor, Lovable, who we're going to talk about elsewhere, and others are decimating these coding jobs. That potentially means a lot of engineering talent being scared, but some of them leaving the tech space forever. However, more likely, knowing people as we do spending, you and I spending our lives in this sector, will go and start something up. And I have a number of conversations going on with people who have done exactly that. The other effect, though, per employment is the AI effect means smaller team sizes.

18:40And you see that in the tiny teams that are being funded, certainly in the US and beginning to in Europe too. And these small teams can win faster by moving faster, obviously, back to your point about concentrating. And it also, though, may mean subscale startups. If you don't scale and if your ambition isn't to scale, you are subscale by definition, which oddly, and this is a weird twist, is where Europe has done really well. Because one man's subscale is another man's perfect acquisition. so perhaps a hidden benefit from all of this chaos and worry and a need to pivot is that we get all those wannabe subscale AIs we have here in Europe finally break through.

19:25Yeah but let's not forget that you know AI is just another manifestation of technology and even with the AI you need a certain amount of people to run that firm they are the overheads and there's always going to be pressure to if you you know if your runway is shortening there's always going to be pressure to reduce costs and that usually means people so um you know this is all about getting the strategy right and and some of these basic lessons do do pervade over time and i think it is true um you know running a business is a tough thing to do um and you know as they always say the darkest hour is just before dawn you know that particular experience did work it gave gave the space for break-even and even profitability because it allowed us to pull in those customers in the pipeline who would almost certainly have departed had their contact been exited from the company and because of that their faith in our offer would have evaporated but you know for the Europeans it avoided all the costs disruption and distraction of making people redundant you know from from you huge bureaucracy from consultation periods to severance payments, which can take up as much time as actually running the business.

20:44But what we did do, and this is an exercise in positive thinking, which is really, really important in tech, we turned the problem into an opportunity. It basically created more staff members with equity and strongly reinforced that team culture, which we had been working very, very hard on for years. And the firm then continued to grow until it was time for exit though. There is one word of warning and I say this with my mergers and acquisitions hat on. When it comes to exit a lot of small equity owners can be a problem for potential acquisitors. So it makes sense to plan a process of open concentration back to that word concentration of shareholding before actually starting any exit processing.

21:32Just clean up your shareholder sheet. So de-risk them, buy it back and concentrate. Yeah, clean up the cap table. Brilliant. Great advice. Thank you.

21:55So lovable, could this be Europe's new category leader? So if you've been reading the Financial Times in recent days, they were reporting that Sweden's lovable AI startup is set to raise $150 million in funding led by VC Accel, West Coast based but possibly in the London office leading on this with participation from Nordic VC, Creandum and London's 20 VC. So what you might think? Well, in a few months there have been awash with bad news. It looks like Sweden's Lovable is set to become Europe's newest unicorn at a$1.8 billion valuation. Okay, I mean, we've said before that we actually think that unicorn valuations are bollocks.

22:42But let's not allow that to spoil a good story. And the valuation itself might actually have increased during the process of fundraising because of, and I quote, overwhelming investor interest. Now, this is quite interesting because the company hasn't been around very long. Only a few months ago, back in February, it raised$15 million in a pre-Series A round led by Crandom, who clearly like Lovable. So, what's the big deal? What's to love about Lovable? What's to love about Lovable? Well, it's about using generative AI to create software and not just words, which frankly most of us use generative AI to do now.

23:28And this has been one of the use cases with the largest potential from commentators since the technology appeared a couple of years ago. But platforms like Cursor and Copilot are mostly confined to a world inhabited by trained engineers and other inhabitants of darkened rooms sporting snow-white screen tans. The same being more true of potential lovable competitors like Bolt that have also delivered some mind-boggling AIR growth numbers recently. All right, so let's step back a little bit from the hype and let's put all this into a category context. That's what we're here for, right? So the alphabet soup is on the hop and the jargoneers, including, I have to say, myself and Jonathan here are on the march.

24:16And it's time for the curious to take note. And as with this podcast, why don't you actually take some notes Here goes. Let's break this down. So first of all, we've got LLMs and AGI's, large language models. These are the dictionaries of AI, the brains, if you like. And then we have artificial general intelligence, a sort of promised land, which some believe is Nirvana. And others, as in some of us in real life, believe it's a heaven which doesn't exist. Let's dive a little bit deeper into the mechanics of AI and machine learning, let's think about the concepts of tokens and context windows. Tokens, a way to mete out compute power measurably, which is important for those of us like Jonathan, who cares deeply about rising energy and water consumption, especially when it comes to AI.

25:03And context window, that's the number of tokens, essentially the limit of prompts that the AI can take. It's a bit like IQ. The bigger the number, the brainier you are, allegedly. And for simplicity bigger is better. Unless, like our guest on our landmark 50th episode, Amber Vodigal, CEO of 28X, you want to build AI for smaller devices like smartphones for usability, or in her case, very valid privacy reasons. So we now have a rampage of new subcategories based on this alphabet suit, and even potentially groundbreaking innovations recently from old school names like Cloudflare. Cloudflare recently announced its paper crawl attempt to make people a little bit more responsible for all of this rampant tokenization and context window expansion.

25:51This paper crawl attempt is trying to make people like the BBC and Sir Elton John feel a bit happier because previously they felt shoplifted as the LLM companies have used their proprietary content and scraped it all to train their models. So in this context, we have our friends lovable. Yeah, so I think Sir Elton was not backward in coming forward about this issue in recent weeks. Almost cross enough to announce another farewell tour, but not quite. But things are getting pretty lively, aren't they? Oh, yeah. This is one of the liveliest times in country recreation that Jonathan and I have seen.

26:27And we've seen a lot, right? We've seen an awful lot. PC, internet, mobiles, apps, and a whole lot of stuff that didn't make it, which we don't want to talk about and have probably forgotten anyway. Nah, come on. Virtual reality and NFTs. Oh, yeah. And so, back to Lovable. Such was the shot waves of this impertinent development. How on earth could normal people develop software? People who like sunlight and games and things. They actually got onto the front page of very coderish publications like Product Hunt and Hacker News. Just last year, after they announced that they really were allowing anyone, yes, anyone, to create production-ready software just by using prompting.

27:15No coding knowledge needed. Naughty devils. Absolutely none, right? So in addition, apparently, to building prototypes and websites, it's GPT Engineer, not the most original title, but, you know, there we go. You know where to come if you want to change that. They were saying that it can ship fully functional web apps, Lovable now claims to have half a million users who are building over 25 ,000 new products daily and says its financials are growing too. According to Lovable themselves, it's now reached$17 million in annual AIR, that's annual recurring revenues, after scaling to 30 ,000 paying customers in about two days.

28:04That's 48 hours. Money, money, money. Yep, that great Swedish band. Joking apart, these numbers could make it one of the fastest growing startups in Europe. Possibly ever. I can't remember anybody claiming to have gone out this quickly. OpenAI did okay. DeepSeek did okay. But as if that wasn't impressive enough, they actually, this is lovable, claimed to achieve this milestone with only$2 million in capital. From$6.8 million pre-seed funding round, it closed only last October, not even a year ago, from Hummingbird Ventures and by Founders, which is actually an angel syndicate. So they've gone from an angel syndicate to full-scale, world-class VC investment in way less than a year.

28:58You know where that is? Why do you think, Paul? Because the winner takes it all. Oh, see what you did there. Anyway, talking about the winner taking it all, they are getting very punchy. Their founder is a guy called Anton Ossica, told the venerable, but much reduced in recent words. Of this parish, once of this parish. Once of this parish, TechCrunch. The platform is different to competing AI-driven code-building platforms, right? I.e., their direct competitors. and this is what I thought was incredibly punchy. It's the best way to get something that actually works. All our competitors are a bit shit.

29:40I think handbag waving Swedish style, I think. So careful there, Anton, careful. Apparently the platform is using a combination of other people's stuff. So OpenAI, Google Gemini and Anthropic distilled into a platform that can generate the software. You know, sort of like Jack Daniels is distilled from pretty much the same stuff as everybody else to generate more rock and roll hangovers, perhaps. Anyway, they're claiming these commercial apps are being built on a platform at this prodigious rate. And we said it earlier, 25 ,000 apps a day. I mean, makes you wonder if the market can absorb that.

30:21But, you know, while we're talking numbers, this pales into insignificance when you look at their claimed ARR growth. We talked about the top line numbers anyway. Four weeks after they started trading, they were claiming$4 million in AIR. Three months after beginning trading, it was$17 million. Then,$50 million AIR after six months, and a month later, claiming to have done$75 million. So just bear in mind, this company was founded in 2023, and its product was only released late last year. Yeah, right. But to caveat all of this loveable, adorable, loveable Swedish growth, let's remember the new winners.