52. Do we need to get ready for the age of algorithmic investing?

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

0:09So, Paul, what's coming up on today's The Difference Engine podcast? We have a word from the wise to the London Stock Exchange. We also wade through all and always coming out of London Tech Week. But first, is big money really betting on algorithms to beat humans?

0:30So the co-founder of Revolut, Russian Nick Staronsky, has got a new wheeze, apparently. He built Neobank Revolut, not without controversy, we may add, from zero to one of Europe's most valuable startups, or shall we say scale-ups by now. $45 billion valuation, 50 million users, over$10 billion in profit in 2024. Not bad, eh? Sorry, what's that? But you can't keep a good entrepreneur down. And now, apparently, he's applying his systems-first approach to venture capital. And we hope, new tech category creation with investment vehicle Quantum Light. Now, replacing VCs with AI algorithms with the result that 17 investments have apparently outperformed top investors by 2x.

1:28Now, to be clear, 17 of Quantum Light's investments were identified by algorithms and have outperformed human decisions by 2x. Now, for those of you in the business, traditional VCs rely on an interesting cocktail of research, gut feelings and of course personal networks but the problem is is that most startups fail anyway despite all this expert human judgment um you know the one in 20 blockbuster success rule that's actually kept the vc model working what's that on always paul uh oh yeah um it's my new algorithmic trading strategy what sounds like you're throwing darts into a dartboard well i do call it the go big and luke littler thesis and it hasn't failed yet but maybe that's because i have not had just like a russian friend i've not had to return any capital to actually any actual investors yet all right well good luck with that then um you know back to the main story uh in 2022 steronski launched quantum light with ceo ilia kondraschov kondraschov exited And he's apparently an exited entrepreneur and, of course, a Staronsky confidant.

2:47Now, their radical thesis, remove humans with all their foibles from radical decisions entirely. Now, according to Kontrashov, this lets them almost eliminate human judgment from the investment process. Let's be honest, they're not the first Russians to remove people from pictures. Steady on, steady on. Now, let's go back to the main story, Paul. Here's how it works. Now, Quantum Light analyzes apparently over 10 billion data points, tracks more than 700 ,000 venture-backed companies worldwide. And of course, the AI evaluates startups across dimensions humans can't process. Now, that means no subjective decisions in theory, of course, and in theory, no unconscious bias.

3:32Zero human deal sourcing, zero gut instincts, zero personal networks. Just pure data-driven analysis. Just machines finding patterns humans might miss or are incapable of spotting. Just the good stuff. Again, yes, the good stuff is, just to repeat this, 17 investments at performing humans by 2x. Not bad. But here's the real valuation. They've just raised$250 million for their first fund. Ironically, investments probably not identified by algorithms. And talking about putting your money where your mouth is, Staronsky has put out his money too. Personally contributing 25 % of the funding. Now that in our maths adds up to more than 50 million.

4:21Nice to see him backing his own judgment. But I want to know, did he run an algorithm over it? And the rest came from billionaire tech funders and founders, sorry, and institutional investors. Right. So, yeah, this sounds on the face of it, extremely impressive. But just to take the other side here, hedge funds have been making billions from making millions of automated calculations and have done so for years, right? So, algo trading has existed since the 70s. And that was pioneered oddly by good old school players like HPE now and IBM. And since 2005, algo trading, which is what I'm going to call this, was even regulated by the SEC.

5:01So, nothing new here to see. But what maybe is being missed here is investable young companies, especially unstable tech startups, are run by ambitious human beings. And not all of them have fully developed frontal lobes. So, and you can see this with some of the, you know, some of them also have an odd sense of morality. You remember effective altruism, still going strong in the value I hear, FTX, etc., who just plain fall out with each other and their investors and trouble followers. All of this means it's the value over time, not the initial stock picking, which matters. So no doubt there's lots of variables that can be pumped into an LLM and remove human error from the initial analysis.

5:45And who would not want to cut down on pitch decks? But I still think there's a risk of missing the look them in the eyes, fire in the belly, which can take a so-so idea to a once-in-a-lifetime investment opportunity with people like Zuckerberg at the front of it, or Travis Kalanick of Uber, or over here, we had Mike Lynch or Herman Hauser. These are individuals that are hard to spot by humans, and certainly not going to be picked up on any algorithmic analysis that I'm aware of, because algorithms, of course, are non-empathetic to human beings. And we also know just how opaque these early stage companies are.

6:26It's hard to tell if something's going to fly and be the next Decacorn or not. Public numbers, whether financials are not available, sales prospects you don't know about, and even the increasingly irrelevant, irrelevant because of AI, headcount numbers are just not thrown about. So I do hope good old Nikolai's algorithm has a smart way to proxy these. And maybe he's a smart guy. Maybe he does. So what can we sort of learn from all of this? the fact that this may not be the first rodeo for some form of AI or machine learning. We know in this case that the smart and big money is betting on algorithms to beat humans.

7:09But I think before we do get too excited at the sheer chutzpah of this move, this is not some sort of outlier. It's thought that already, as Paul has just alluded, did, around a third of VCs now find at least half of their deals through AI tools. You know that? Well, that's really bad news for white male MBAs with conservative haircuts and Patagonia gilets. The future of venture capital won't be about who you know or even who you went to school. No. Nobody would have wanted to go to our school. I know. I wasn't talking about us. No, it's just going to be less about what you know and more about what your algorithms can see.

7:48So if we think about what we learned from this discourse, to a greater or lesser extent, frankly, the age of algorithmic investment has begun. And, you know, it's well going now. Categorical category? Category creating nirvana, but with a caveat. Interestingly, in this particular case, once the algorithm has done the picking, Quantum Light has gained a reputation for taking a rather hands-on approach with startups at BACs, looking to instill the operating principles Staronsky used to scale Revolut. This is done via a series of playbooks it gives to its founders. Here we go. So this is the truth.

8:32The initial stock picking may be done, but the playbooks thing rings true, right? This is where the dirty, human, hands-on part of the process starts. And in fact, I think this could be more successful, this blended approach, than a totally hands-off approach. And I've seen this recently firsthand. We ran an AI workshop to look at how the best practices in agentic AI can help with best practices in marketing. And they are taught by feeding your own playbook, because these days it's possible to build your own custom GPTs, your own notebook LM, your own version of an LLM, which teaches the best practices that you know.

9:10And they could include at any stage in the journey helps on HR and hiring and firing. Sales help, which is a big AI area of interest. Marketing, selecting channel partners, et cetera, et cetera. All from a playbook, AI or not, which you create to help these teams once they've been selected. And unfortunately, there's no AI that's been developed or is possible developed that has prior knowledge of events. So you do need that hands-on advice, hand on the shoulder, coaching, empathy, which we're talking about here. Which AI could have predicted COVID, or the fallout from the Trump tariffs, or the current natural disasters, or the current conflicts that we've got in?

9:55Very few. And it's said by Manny that if you want to avoid the mistakes of the past, which Storovsky and all of his portcos will want to do, then you should learn from the past. If you want to be smart and agile enough to guide a struggling startup through a hard pivot. But what you need is humans in the loop. It's not all hands off. Funny you should say that. Because if you are in this particular orbit, if your firm succeeds in being pointed to by the algorithmic fingers of fate, be careful what you wish for. You're still going to be a guinea pig for less than algorithmic ambition and ego because for three years kondrashov has been on the job of providing the firm's ar-led approach to investment the pace is monumental a source told our friends at sifted he's not going to let you relax before lighting a fire under you now let's hope for the investment's sake that's not a bonfire a bonfire of the vanities and perhaps a major source of this legendary by now 2x performance good old human fear and loathing augmented the magic machine perhaps well yeah and my final thought is 250 million in valuation for the fund and 50 million chucked in by a guy who's worth many many more times than that i'm sure he could find 50 million down the back of his revolute branded sofa if he were so sure that only algorithm was perfect wouldn't it be all in on quantum light

11:41you know what this really grinds my gears okay so do tell what is grinding your gears this time well it's london tech listings and particularly uh the move of wise and this is a long overdue wake-up call you'd have hardly noticed this thing had happened if you'd read the general media saturated with the same tired old faces his self-congratulation and dubious hero worship of London Tech Week. But Wise, formerly TransferWise, and one of Europe's tech category leaders has moved its listing from the London-based LSE to the tech-heavy NASDAQ in New York. Boo, say those of us who think, including those of us here at Categorical Towers, that think the UK needs more tech startups going public as opposed to being swallowed up and giving their real profits, the real mature profits to folks from overseas.

12:38Yay, say those shareholders who actually include us too, who have seen the Y share price rise 50 % since it launched and believe there is much more value for this particular stock in the more liquid and widely held US markets. Well, this is just the latest in a series of tech departures and delistings from AIM and the main board on LSE. But this should have been treated as a long overdue wake-up call that it is. Because the only way the UK economy will grow in any sort of significant way and future-proof itself is by scaling and listing successful technology companies and dominating key tech categories.

13:18That's because achieving exactly this has made the fundamental difference between the performance of the US and the UK slash EU economies over the last 15 years following the 2009 financial crisis. It's too easy to blame UK politicians experienced only in public and not-for-profit sectors and their short-sighted policies for our predicament, but we're not alone in that respect. The truth is that America can keep shooting itself in the foot as it's doing spectacularly at the moment thanks to the leadership of the world's greatest dealmaker and stay ahead because it keeps producing the next game changer thanks to a combination of immigration, inspiration, level of investment, exit values all but known in Europe and sheer momentum.

14:02I'm assuming, of course, that ICE isn't successful in deporting all the new and first generation immigrant talent that has created the US tech industry. Google, Microsoft, NVIDIA and Oracle, to name but four, would be looking for new leadership, for starters. And this is not because, as once again reiterates at London Tech Week, that we here do not think big enough. The tall puppy, who is tall poppy poppy gustafson so he did that see led or more accurately school prefected her way through a panel at this show with the breezy title i love this why the uk is the ultimate launch pad to expand your business um not sure who anyone was kidding with this poppy who is now a baroness and if that was not enough overachievement is also the minister for investment in the uk Jolly hockey stick to guests asking them to parrot platitudes about how government grants and tax breaks made Britain the best place for promising tech businesses.

14:57Well, they duly did. But if they'd hung around until mid-afternoon, the very same day on that show, they would have heard the much more familiar criticism of UK and European entrepreneurship from those who really know how it goes. the panel uh starring uh lingi the ceo of tensile europe who was an early investor in monzo and revolute was clear she said we are looking we are not looking to invest for five to ten years we may look at a biotech for an investment of 30 years rightly so yeah and she added the uk has an abundance of ai talent by the way jensen wang said the same thing at the same show that morning.

15:38So the UK does have an abundance of AI talent, but there is a leadership gap. Once UK teams scale up to a few hundred million, do they have the stamina to become a unicorn? I mean, I don't think many UK tech leaders think about getting to the few hundred million. Perhaps we should ask WISE. Another panelist was clear, calling out UK tech entrepreneurs as not showing enough opportunities for the really big opportunities. Well, let's add UK and European-based investors as well, because they're clearly not providing the finance to allow anybody that really does want to scale up to get to the heights.

16:18Back to Wise. Yeah, I mean, let's be clear. Wise was the only world-class B2C tech company on the LSE. You know, before this debacle, you could just about imagine Monzo, the one that, you know, we're a competitor for Revolut, listing there, but because they don't actually have a significant US operation and actually therefore they're not really world class. But, you know, maybe regulatory reform to encourage pension funds to invest in tech might have started to stop the rock, possibly creating a special deal with Revolut as another flag bearer and encourage others to list and stay. So, you know, Is anything happening?

16:59Big question mark. Well, recent reforms to the listing regime, including permission for dual class share structures, have happened. But the problem is institutional capital outflows and liquidity issues continue to blight the market. It wasn't so long ago that we were thinking that Britain had an unfair advantage because we were so good at legislation and regulation. And yet here you have the London Stock Exchange just about accepting that what is completely standard in the States, which is dual class listing, is something they might consider. Yeah, and the reality is, rather than hitting the issue head on, we've got a sticking plaster on an open wound, the Private Intermittent Securities and Capital Exchange System, commonly known as Pisces.

17:47That's a sort of cod. Yep, see what it did there. See what it did there, exchange, based on the reality that faced with a not fit for purpose London, well, with not fit for purpose London exchanges, companies are choosing to stay private for longer. That means there's a demand for investors to trade private company shares easily and efficiently too in some form of organized marketplace. Pisces actually does meet this demand by allowing secondary trading of these shares in a private stock market. Companies can set the floor and ceiling of share prices and have a say over who can buy their shares.

18:26Problem is, of course, it will only be launched later this year, 2025, after the FCA has announced the final rules. Do you think that's going to happen on time? Let's hope it works anyway, but it's a distraction. No tech company, certainly not one that has a potential to be a game changer will now list on LSE without drastic wholesale reform. Now this is urgent because the UK economy will keep stagnating. Indeed it's currently shrinking according to latest GDP figures. The problem is we're an economy addicted to state spending propped up by tax take from centuries-old banks, oil companies and gas giants that are now relics of the past.

19:08Not only that, but they're in an inevitable decline as the structure of the economy changes. And the reality, no matter what any political zealot says, that we have to get to net zero ASAP, otherwise we won't have a civilisation to maintain. The foundation's already shaky. These dinosaurs aren't growing, aren't hiring, aren't paying more tax, yet they dominate our markets. Again, the only thing that will deliver sustainable UK economic prosperity is building, scaling and listing new category leading British technology and technology based businesses. And the same happening across the rest of Europe and the Nordics.

19:50Well, that's funny because if I listen to the tech minister, what we've got to do is regulate more. Yeah, right. And make more policies. Yeah, they're good. Think tanks. Yeah, right. Yeah, sure. Right. Okay, so the problem we've got is young companies can't scale and large ones are shrinking. It really doesn't matter what policy wonks in national government or the EU come up with. Just witness the EU's startup and scale-up strategy. It just shows how disconnected European institutions are from the private sector. Capital markets need fixing. They're the driver of growth and nothing else matters until they are.

20:28Otherwise, we are rearranging deck chairs on the Titanic. Yeah, so, well, it can't be that negative. The one hope we see through STEM, pun intended, STEM, the flow of AI talent from leaving the UK is if we can attract the magnets for talent, such as Hussein Kasai. And once again, I will plug episode 42, give it a listen. But with that, if we don't sort this out quickly, we've got an increased tax on jobs from this government. We've got an exodus of the mega rich who are the people who can invest in these companies. And we've just admitted, as we see, the sales of Arm, the exodus of DeepMind, and now Wise leaving the UK, potentially.

21:12This is not looking great. Right. We need to stop talking, tweaking, and frankly, twatting about and start meaningful, full scale reform of our investment listing and exchange structures. Does that sound scary? It will be if we don't.

21:32Okay, next up, London Tech Week or London Tech Week. So we've been seeing contrasting stories emerging in the aftermath of London Tech Week. That is definitely with two E's, which for those of you not familiar with it, is the yearly bun fight, where prime ministers try to look smarter, Czechies try to be a bit more sympathetic, and big brands try to start at wash their cynical sponsorship and love being spent. Cynical? Us? Right. So once again, the great, the good, the leather jacketed and the besuited talk tech in the birthplace of computing. Wait, no, hold on. No, sorry, sorry. That was Manchester.

22:15That was Manchester. All right. OK, so this year, thank God it was not in the godforsaken Excel, well, but nor the surely passé Shoreditch, or even the cool Britannia nonsense, who can forget that of Tony Blair's days in the Britpop era. This time, it was dual-centered in the classic West London Olympia, newly refurbed, and the edgier but equally historical Tobacco Dock, which hosted the partner event, which was the UK AI Summit. Remember when that was good? Anyway, this year, the supposed highlights were Keir Starmer and lifelong politician with zero business experience, our own tech minister, Peter Kyle.

22:57If they were listening, they would have heard the real star of the show, NVIDIA CEO Jensen Wang, who confirmed that the UK had some of the best AI talent, universities and brains, but like the infrastructure for which read data centers and energy to match the ambitions that it has. Ouch. This year, though, the reviews were, shall we say, mixed. On the one hand, we have the lovely Lara Lewington, who we've worked with and is an amazing tech pro and a fan of tech, who has been amazing for a decade or so on BBC Click. She hosted the AI Summit. And to be fair, we did not go to the AI Summit, but she was, you know, characteristically gushing and she tweeted what a week hosted the main stages at london tech week and the ai summit interviewed brian johnson at founders forum and spoke founders forum longevity buzzing around the uk tech scene and its ai ambitions right now was that was that brian johnson the cricket commentator johnson ac leaders gravelly voiced front man neither he's the he's the nutterist who um is somewhere in his 40s and believes he has a um a biological age in his 20s and measures his own erections he is nutterist all right and wasn't what the funniest thing was that no sooner had tech week begun than an alternative whatsapp group emerged um which sort of blew up to over 600 people chatting about the show um and it's now puzzlingly renamed potato you're right Didn't we manage to get some prime takes out of that?

24:35Oh, absolutely. So as is, you know, the way with entrepreneurs and techs, a WhatsApp group, I think it was originally called the Alternative London Tech Week WhatsApp group, blew up, grew to 600 people. I presume maybe thinking they were infringing some copyright has now puzzlingly renamed itself as Potato. And boy, were there some prime takes. Give us some of those. uh so somebody um who i think we know who it was but who was trying to remain anonymous um what do the rest of you lovely peeps think i know it's sat but any quick fire comments on the gap we can fill together as a community please speak up yeah right so and here it is you know for both barrels ltw's floor was a political clutter of tech virtue signaling felt a bit forced it's odd how detached it feels considering how relevant and jam-packed industry slash tech-specific conferences, meetups, and institutions are in London.

25:31In my opinion, the answer for TWL is to showcase the existing meetups, conferences, and networks, which are genuinely highs of activity. Ouch. Oh, indeed. And there was another one which, you know, took a rather more global scale, which we thought was rather amusing. But maybe part of the problem is our event ecosystem, system and that challenge increased AI dominance making every about the same thing I'm reading this literally yeah so I mean I think the point is there's too many shows and you know every show is trying to be the the AI show you've got but again we're seeing a power shift you know this person goes on to say with LTW, ALS, COGX and more I've no idea of the role of London Tech Week anymore and the floor felt like that hodgepodge of anything uh interestingly makes the comparison I went to leap this and this is I think year in Riyadh which is in Saudi Arabia for those of you not familiar with where Riyadh is all the big boys were there at scale it's the one tech event per year that everyone has to be at but also we're a saturated market so maybe they just don't feel the need to bother here or maybe they have event decision paralysis with too many confusing options like me okay so excuse it but you get the point of that particular thing a pretty fascinating take yeah it's a more whatsapp badinage here and and you can you can tell that the government typically is always one step behind when you've got the prime minister going to shows which are characterized as this here's another take for you plenty of loudmouth vcs swanning around like farts lingering long after their value had evaporated and who holds the most firepower the guys that are pushed to the sidelines after footing the bill when they should have been front and center the goliath organizations with the most customers capital credibility and infrastructure get treated like dinosaurs like yesterday's news when right when in reality they are the ones best place to write the next chapter so that's a spicy one yeah i mean this i love this other bit which is we have we have over romanticized the garage born startups and other shiny things in london as we try to copy Silicon Valley.

27:43People clearly don't know about the history of Silicon Valley. And the only actual garage startup I'm ever aware of is actually HP. Isn't that garage now a national monument? Yeah, it is. So we did sneak a quick plug-in for ourselves, as you do. And we can see that some of the feedback from that was from proper, genuine founders and people that we think are going to do well, Mr. Kasai again from Quench. We've also heard, talking about the show in more general terms now, we've heard from someone who knows more than most that this show, London Tech Week, is paid for or underwritten by Founders Fund.

28:24Now, who is Founders Fund, you may ask yourself. Well, Founders Fund is the vehicle of read, don't start up in London, Hoffman. And he may have forgotten that he said that at COGX. That's Mr. LinkedIn. Mr. LinkedIn, well, now Mr. Microsoft, as he sold it to LinkedIn, and Mr. Not-A-Friend-of-Musk. He's behind the show. So is it any wonder that the government wonks, etc., are there? And it's also run by Publications Goliath Informer, who is owned by the MediaShire UK mini-media mogul, Mike Danson. So it is sort of the show where the guys that have always controlled the strings are still trying to do so.

29:08and they've obviously persuaded their mates in the government to go along. We don't know if London Tech Week is a loss leader or it pays its way in other slightly more shady relationship government lobbying ways. But if there is a bottom line, they might want to consider if they change things up a little bit from next year. And the acid test was the little WhatsApp poll. Was that pay as you want? Yeah. Yeah. So if LTW was a pay as you want event, would you have considered it would you have considered it fair to pay to attend as a full access attendee for those of you who aren't aware a full access attendee would charge you the princely sum of 695 uk pounds pretty much the same sum as it it is to maintain a premium subscription to linkedin over the year yeah and that is just baseline right so there are But the full event passes were knocking at over£1 ,500.

30:13And to see what, as you said, there's one other person I read said, if I want to hear Adobe give an advert for its AI, I'll just go and watch a seminar. I'm not paying for the privilege. So it's a show with some identity issues. So given this poll, Paul, how much was it that people said they'd really be prepared to pay for that ticket given the actual value of the event. The highest answer on this poll, and we've got to say it wasn't an extensive poll, but the highest number of people said they'd like to pay zero pounds. The second highest number was 75 pounds, which was a day pass available to see very little, but go and spend some time with all the sponsors as if that's good.

30:56It looks like there were some crazy spendthrifts who were prepared to pay 395 quid for this. Yeah, God knows who. and I think you described this as an engine. The engine to produce gongs. That seems to be what this is about. Yeah, it does. It seems to have been co-opted into the public sector machine even though it's supposed to have been contracted out. But, you know, Paul, what do you think the future should be? What do we think that London Tech Week should do next year? Honestly, I think it should have a good thing for itself as to what it wants to be. you know is it a showcase for government ministers is it an innovation hub and i think it should probably slim right down and get back to basics thank you for listening if you want to learn more about category design head to becategorical.com if you need help designing and dominating your category then get in touch contact details are in the show notes

32:05Are you still developing separate marketing plans for each of your different channels to market? Can you really sanction messages reaching the same customers from unconnected targets? Listen, it's now the AI age. Things need to change. Positive is a different sort of AI-powered agency. From our very beginning, we've integrated marketing communications from PR to events, social to paid ads. Working smarter means working integrated. Let us show you what a truly joined up strategy could do for your tech business. Contact Positive today with an email to hello at positivemarketing.com.