This Week European Tech: What happens when AI starts deceiving humans?

28:04It is a real open source repo. And the people maintaining that repository didn't know. Yeah, yeah, yeah. I mean so you could say where do you draw the line if it had attempted to hack a bank if it attempted to transfer people's money if it had succeeded like what and and this is sort of what gets some people's you know backsides up a little bit is is sort of saying okay you know how far do you go with this yeah yeah Sean you are you how AI pill are you yeah um I'd say pretty um you know We've got obviously massive exposures through our funds. We're big users of it here internally at Vencap and even personal life.

28:46We were talking off camera. I've just had a new kitchen. Lots of those designs were done by a chat GPT. You're just taking pictures, trying things out. You just fit seamlessly in it. It did an incredible job, actually. The one thing that we see with these cyber attacks is that we're also seeing more and more cyber security companies pop up in our underlying portfolio. Because how do you stop this sort of thing? You've got to fight fire with fire in a way. So, again, it's like every sort of paradigm shift, you get a new attack vectors and then there's opportunities for new companies or new products to sort of come up.

29:16Yeah, yeah, it is going to be a bit more cat and mouse. And I guess it is the only technology that can resolve itself. So we'll have to work out if the intent is malicious or not at what point. Mads, what else is in AI corner this week? Look, I think this was the key thing. And you mentioned Anthropoc have started to roll their own silicon, which is also something we've been talking about for a while. They were leading, of course, heavily in NVIDIA, but also using chips from Amazon and chips from Google. And they're now starting to roll their own silicon. And I guess the question becomes, okay, back to NVIDIA, what's been holding them back from pushing their own open source efforts more.

29:55And, of course, the more their customers are rolling out competing platforms, the bigger incentive they have to just offer open source to the world. So this cat and mouse game certainly is continuing to unfold. It's quite well developed, their open source platform as well, isn't it, NVIDIA? They've been working on that for years. They've been throttling it for sure. I mean, they could easily, it's not a frontier model, but I think they could easily get it there. We've seen some quite unusual moves at Google this week. So Demis Sassabis is stepping aside as CEO of Google DeepMind. I'm not going to get his name correct, so please forgive me.

30:31But Kurei Kovacoglu is now the act, he's the CTO. He's now acting as interim CEO. Hassabis is moving to chair, chief scientist role, focusing on AGI and isomorphic. Some articles have been saying for a while that he's been stepping away for about a year anyway. Same day, Jeff Dean has left. He's been at DeepMind for 27 years with Sanjay Gemawat, Oriel Vinyals, who co-led Gemini, and Kwok Lee. Now, they've all stepped out to found a new business called Discovery Loop, which looks very interesting. is a public benefit corporation trying to automate scientific research by running thousands of experiments in parallel.

31:11Google is investing. So the team is all leaving and then Google is investing, which I guess maybe, you know, keep it in the family somehow. And they're giving them compute for at least a year. Radical and Kostler are in too. Underneath it all, Gemini 3.5 Pro, They've missed their release date time and time again. So there's a little bit of shenanigans going on, perhaps. Is there a low morale? Don't know. Alphabet fell about 4 % this week. Mads, do a better job than I just did about setting up the Google shenanigans and who's gone where, who they are and why. This is a big deal. We've said repeatedly we think, at least I think Google is, a really, really interesting and formidable player in the AI race.

31:55They were obviously early with this 10 years ago. They came up with the attention is all you need paper that your foundational research for the LLM. They then proceeded to build the chatbot internally and not release it. Jeff Dean has been talking about why. He said, we thought it was not a better search experience than search. So we just kept a search and shelved the other thing. And meanwhile, the engineers over in OpenAI told it, march on them. I think to some extent, you could say that what's been happening here is that the researchers, the scientists, Dennis, the chief amongst them, are stepping aside, leaving room for engineers to take over, builders, people that ship product.

32:33But it is a big deal. I mean, Jeff Dean leaving Google is almost a little bit unreal. People that are in the industry may know Jeff Dean. He was employee number 30 at Google. He is somebody who was behind MapReduce, which is how Google indexed the entire web. He was behind Bigtable. He was behind Spanner, which is sitting under the things like search, like Gmail, like YouTube. He co-founded the Google Brain team. He pushed TensorFlow out into the world. Then he spent a decade on the TPU, cutting inference times by 10 to 30 times there. He was one of the most, one of the two most senior engineers at Google.

33:13In fact, there are Jeff Dean facts about him. He's sort of seen internally as the Chuck Norris of coding. And he was absolutely a, you know, almost sort of an unreal superstar in that sense. And he's left with Sanjay Gamawad, who was the other most senior engineer. So the two most senior engineers at Google have now left. And they've been working on, you know, who's who. If you look at who are the most sighted people on the planet in AI and in large systems engineering, the new company they've built, I think, have three or four of the top eight most sighted people in both categories. But they didn't get poached for money.

33:58We've seen a lot of people in the great shuffling of the chairs running for cash, and they haven't. They've stepped out, haven't they, to build their own thing? Dan Jeff Dean, he was employed on the 30th at Google. I don't think he needs money. I don't think he cares about money. For sure. But this is mission-driven. It feels to me almost mission-driven. It isn't, you know, he's no spring chicken. And that's precisely the point. I think there was probably maybe a little too much mission and a little too varied set of missions at DeepMind. We had lots of people that were focusing on research and on publishing and maybe not enough on shipping, which is why we're seeing all these delays we've been seeing to Gemini and why Google have been falling behind over the last nine months.

34:44What does that mean? So Jeff Dean and Sanjay, they've gone off to build a public benefit corporation doing scientific computation for scientific research. Meanwhile, who's taking over? I think you're right. I mean, it's not the easiest to say, I get it to be something like, so it's Corey Kavukcu Ulu. Much better job than I did. I'm going to go with yours. I'm going to dub mine with yours. So he joined DeepMind in 2012 and he was part of the team that built the model that learned how to play Atari games. They could play all 49 Atari games from raw pixels alone which was the seminal paper that led Google to buy DeepMind.

35:28But he's not just an academic. He's been the chief AI architect. He built a WaveNet that is kind of the thing that's making synthetic voices that stop stopping them sounding robotic. He's been working on lots of projects at scale, assisted maps, voice search. I think what Google is really trying to do here is to say, look, DeepMind was a beautiful thing, but now we need to ship. So DeepMind no longer has a CEO. That title dies, the Demis. Corey, he will be a senior vice president reporting directly to Sundae. The idea here is let's get back in the game. Let's build the models the market wants and let's ship.

36:04So they're not bringing, well, I guess we won't know, but they're not bringing a new person in. I would be surprised. I mean, where would they, they could maybe try and go poach somebody from OpenAI, but Google probably has a deep enough bench. I mean, they'd been hiring all the best for decades, right? Yeah, yeah, yeah. No, I'm pretty long on Google. I think they might end up shining after all of the frontier shenanigans. Sean, do you have any thoughts on this? It's a big deal. And sorry, Sean, just coming back to you in a minute, you look at the resurgence of AI labs and companies in Europe, so much of that concentrated in London, so much of that directly related to Google DeepMind spinouts.

36:45So it's been such a powerhouse. And I think that there is a bet now that it will be a different culture going forward. And we're going to have to see what that means for the organization. I mean, hopefully Google can be back in the game and really challenge and compete at the frontier. Yeah, I mean, DeepMind isn't Google and Google isn't DeepMind, and I don't know where the lines are drawn. They are, absolutely. No, no, Google isn't DeepMind. DeepMind is Google, Google is DeepMind. Of course. And DeepMind will no longer have a CEO. The leader of DeepMind will report directly to Sundar going forward.

37:18But it's not like they're going anywhere. I think there's still a lot to do. And I think this is just a slight twist in the story. We'll see. Sean, have you any thoughts on this one? Yeah, I think the way that we sort of see it is that it's talent going from the Google and the public market, which is outside our universe into the private markets, which is. So it's sort of quite exciting to see talent flow that way. I mean, Google was, it was in our portfolio back in the day, but it hasn't been for a long time. So yeah, it is sort of nice to see things going over this way. And then, you know, again, put in my sort of, you know, primary sort of hat on, you sort of look at our managers and saying, look, are they the sort of people that will be able to underwrite this sort of opportunity where it's sort of, you know, the top of tier one level talent, them looking to build big things are they able to you know underwrite it are they able to you know convince someone like jeff dean to work with them and so again these are the sort of things that we sort of think about when you see these sort of headlines you know any sort of assessment on the company is just yeah probably not for us let's get into your world sean let's talk about the the title of this item is the one billion pound uk scale-up fund but i'd like to broaden that to the kind of the uk venture scene but just to just to tee it up a little bit the uk scale up fund this one billion pound fund is a government-led initiative calling in pension providers such as rail pen nest and border the coast it's not live it was announced i think by burnham it was one of his one of his semi-crowd play or part of the crowd plays early doors when he when he joined office.

38:51Some of the participants have stated that they've made a commitment to explore, which I don't know what that means, but this is now in the public domain. Sean, talk about this specifically, this one thing, if you have any thoughts on this, and also how you view the broader market. And I'd love to, you know, I know you've got some performance data, I know you know this world very well, but as an LP, how do you see the UK specific venture scene from the mega funds all the way down the stack? Start wherever you like. Yeah, I think the interesting thing, just like always at the top of our minds at the moment, is that we're getting more and more UK exposure and European exposure as well, more generally, but from our US managers.

39:32So, you know, like we've got, again, people are coming over here because they're seeing the talent. Things like Eleven Labs, things like Synthesia, Lagora, Mistral. These are all people with, you know, American funds on the cap table and not just at the growth stage. So we see that as a big validation of the market when you've got people that are looking for global winners or choosing to spend time in the UK. But then I guess one of the corollaries of that is that there is lots of capital for the best companies. So that's true in any sort of major market. So when you sort of say, look, there's going to be a new entrant, it's going to be coming into a very viciously competitive market as opposed to it being a sort of untapped opportunity.

40:13So again, that doesn't make it any different from any other venture fund, because if there's a sort of part of the venture market where there's no competition and things are easy, I've yet to find it. So again, it's going to be an interesting sort of thing to see develop. Well, a question for you both. Does the UK need a growth fund? Absolutely. We need more capital all the way up and down the stack from post-IPO to pre-IPO to growth to C, B, A, the whole way, because there is capital at the earliest stages. But if those companies can't graduate to raise follow-on rounds, the market, the ecosystem is not going to develop.

40:49And to Sean's point, yes, of course, there's money in the US, but it's always going to be hard from Sand Hill Road to Diligence and PIG companies. And yes, the very, very best companies, the top 5 % will always get funded. But you do need to have a thousand flowers boom. You do need capital to reach more things and more places for the whole staff to work. So I'd say absolutely. Yes. Yeah, I think for me, it's like the more interesting part of this is catalyzing the sort of pension fund or the LP community. Because, again, you do see lots of capital in the UK, but it being US managers, it's coming like that.

41:27Those underlying investors are from the US. That's where the gains are accruing to. and I think like the problem with venture is that it's a very different part of the market just about anything else when we talk to people that haven't spent any time there the first step is basically unlearn everything you've learned about investing in public markets or in buyout funds it's it is it's a very different beast but it's also a very small niche of the market so if you've got you know your private markets portfolio of which venture is 10 % you're probably not spending that much time there and if you're an institution that hasn't done venture before hasn't done it for a long time it's sort of hard to sort of go up the ramp i mean i've been investing in venture funds for over a decade now and bank up as an institution has been doing it for almost 40 years and we still find it really difficult and it's all we do so if you're you know what's unpack that a bit sean what what's the what's the what are the challenges or maybe some success stories and some failure stories without naming names but what unpack that a little bit for us i think at a really high level um you know there's data coming around from carter they do a nice job with some recent their benchmarks and people have been looking at the performance numbers and you know dave our cio posted this on on twitter got quite a few reactions like the numbers aren't that great um to put it mildly and everyone knows like the median venture fund carter's mainly us data though isn't it it's mainly us data and it's mainly smaller funds and you know you can look at cambridge associates or pitch book or prequin or wherever your favorite benchmark provider is.

42:51But it's sort of the performance you see in the benchmarks doesn't match up with what you see in pitch decks. It's quite slightly, maybe for in the top, we're going to do 5x. Everything's 5x, baby. I've seen people promise 10x on the first time fund. Maybe they will. Who knows? Did you sign them up? I'm assuming you were throwing money at them. How could you say no with that sort of projection? Exactly. But it sort of comes to the fact that The venture ecosystem is so skewed. If you're not in those sort of top quartile, top decile funds, it's really hard to make it work. And it's sort of very easy to go in.

43:23And we've met people that have done this as an allocator. You go into venture, you back a load of median funds, because that's what you'd probably expect statistically. And you say, well, this doesn't really work. We've taken a load of risk. We've had massive illiquidity. You know, why are we not just putting this allocation into lower mid-market private equity? So I think that the hard bit just comes through in that performance data. and that's always going to be the case. Yeah and I guess bonds are now really attractive and other vehicles are now really attractive for investors as well so how do you make decisions and what are the key things that you look for the things that maybe aren't so common knowledge when you're looking at firms?

44:00I think like the way we do the venture world is sort of really quite simple and there isn't any sort of like magic diligence question or silver bullet that will sort of get you consistent performance. The things that sort of keep on driving the industry out of an industry level in terms of performance is those top 1 % of exits. You know, it's the top 20, 30 IPOs or M &As of any given year. They tend to be the majority of exit value. There tends to be a pretty small group of VCs that can consistently back those companies. The other interesting sort of component of it is fund returners, where if you've got one company returning at least the entire committed capital of your fund, again, venture is an outlier business.

44:36This would be that exemplified through that. But when we look at our data, about 90 % of our funds that are doing over 3x net have at least one. So again, it's sort of, do you think you can partner with the world's best companies? Or do you think the manager assessing can? Do they have a track record of doing it, which is normally how we like to look at it? And are they able to do it in such a way where it's meaningful for their fund? Because if you're able to put a small check into a very late round of a great company, that's a nice logo, but it's probably not going to translate through to fund performance.

45:06how any any themes or trends that you're looking at maybe china america india europe uk any themes or trends that are kind of driving your strategy at the moment the interesting thing for us is that we sort of view all these trends whether it's sort of sector so lots of investment at the moment into the sort of you know physical ai semiconductors networking equipment all this sort of stuff different geographies but we're sort of passive consumers of that if that's a right way to put it, because what we like to do is back generalist managers, and they'll tend to go where they see the best entrepreneurial talent.

45:41So it means that you get early exposure to, you know, the LLM. So we were in early rounds of open AI and Anthropic, which we're very happy about, but we had no AI thesis at the time, we were underwriting the funds that did that. So it sort of helps keep us on the front foot in a sort of convoluted way, because we're sort of, you know, leaving ourselves open to sort of backing those emerging trends. The same is true for geography. The reason we went to age 20 years ago is because lots of our existing global managers said, look, we see a great opportunity, we see the talent, we're not going to say actually now's a great time to be backing entrepreneurs in Beijing or Bangalore, as opposed to the UK or San Francisco.

46:19So again, we sort of see all these trends. But a lot of what we focus on is, are you able to consistently partner with those best entrepreneurs? No, that makes sense. Any coming back to this one billion pound UK scale up fund, What are your thoughts on this specifically? I think for us, and before this sort of podcast, we sort of cut some of the data we've got because we tend to back established managers. They tend to have larger funds. So if you look at, we looked at the 2010s, we looked at 2011 through to 2020, we had 17 funds that were over a billion dollars. So dollars, pounds, blow your eyes a little bit.

46:53It's a mixture of early stage and growth stage. And obviously, a billion dollar fund investing in 2011 is very, very different in terms of environment to 2020. but just as sort of like a simple cut, you know, what do these funds look like? And the interesting thing is, you know, they're performing pretty well, like as a cohort across the decade, it's sort of 2.8x net. So you're almost at that three mark. We sort of projecting them out to get roughly in line with that sort of three and a half times, which is roughly what we'd expect from this sort of cohort. So again, performing really well. There's, you know, 40 % of those funds are over 3x net.

47:24There's two of them that are over 5x. There's none of them that are losing money. so you sort of look at it and say well look billion dollar funds they work this is great but there's sort of when you dig a little level deeper you sort of see just how hard this sort of stuff is so on a look-through basis there's 759 companies and if you look at the top five they're generating about a third of the total value so if you're not you know power law and power law and power law power law power law and if you look at the top 10 companies that's almost half the value. And these top companies, it's things like maybe a ByteDance coming out of China, a Databricks, or a DoorDash coming out of the US.

48:05There's a big slope of open AI in there. And these are sort of the$100 billion plus companies. So if you take those out of the sample, then suddenly this is not going to look very exciting at all. And the key thing is, are you able to get into those key companies? And then going back to the power law and the the fund returners um you know there's eight individual fund returners in this sample so with those eight fund returners it's not that eight of the 17 funds have one it's that three of the funds have two and there's another couple sort of floating around there so what it means is that you know this cohort has performed exceptionally well or you know pretty well at this point hopefully we'll get to that exceptional level but it means that there's a couple of truly exceptional funds in there and a lot of them that are sort of you know that are doing good but are fine but it's power law at every sort of level.

48:55Again, so when we look at this compared to the billion pound UK scale-up fund, I sort of look at that and say, look, this can absolutely generate and just our returns. The fund size is a bit of a red herring in a way where if you back the right companies, you can absolutely generate venture performance. The problem is coming out of a much smaller market like the UK compared to like a global fund, you might have one shot. you might have two shots and if you miss those one or two companies you might end up being on the wrong end of these sort of benchmarks and if you get those one or two you might be a hero so you know that that is a tall task but you know no one has ever accused venture of being easy no it's not easy and the bank is pulling this all together and we'll see who comes in and who who makes the the real commitments beyond the commitments to explore in inverted commas Mads any any thoughts specifically on the scale-up fund or the UK venture scene more broadly yeah I mean it's just a bit of a reaction to I think Sean and you also Dan you pulled out the point around the industry kind of the main return the median return maybe not being being very being very good I think to some extent it almost seems like there's some schizophrenia going on because the industry we're in is that backing you know massive outliers which takes crazy the ambition, which takes a desire to go for it, which takes, you know, really a will to try and knock it out of the park.

50:24And so if a founder came to us and said, look, I'm going to build a company. I know that most companies fail. I probably won't build a unicorn. I might, maybe I can return the money you gave me or you invested with me, but the stats are saying probably not. It'd be like, I can't back you. I need you to show that you're going to want to knock it out of the park. Yet some people seem to suggest that if a manager who is in exactly the same ecosystem goes out and raises capital and say, look, I'm here to knock it out of the park. I'm here to aim for 3, 4, 5, 6x. I'm here to create a big fund. It's like, no, no, but that's not what the base rate says.

51:00And you're like, guys, this is the game we're in. We are here to try and move the needle. And we all know that most funds won't be 10x returners, but that doesn't mean we shouldn't try and go for it. Obviously, you can't promise it. everybody who's on either side of the table knows that it's not a promise. It's an aspiration. The founders that come to us are not promising us we'll get our money 100 times back. When we raise money, we're not promising LPs we'll give them 10 times their money back. But it's the aspiration to create something big and meaningful and ambitious, because that's what fuels our industry.

51:34And so I just get a bit confused when some people, they look back to the stats and say, but what about the stats? You can't possibly promise that. Well, the whole idea is, of course, you can't promise it but if you're not trying to go for it then what are you even doing and this thing just confuses me a little bit i don't know i've dan or sean whether you guys have other thoughts well i'm gonna hand that out to sean because you see this from from a very different perspective don't you sean yeah i think a lot of it comes on the the credibility of doing it because again it's very easy to sort of like say look like we're going to go swing for the fences we're going to try and get a 5x 10x fund but then sort of recognizing how rarely that happens you've got to handicap that quite significantly.

52:11If someone's got a track record of delivering that sort of performance of working with the founders that can help generate that sort of performance, and that's sort of much more credible. If someone can explain it quite clearly to the portfolio construction, you know, how you would get your fund north of 3x, again, it's sort of something to sort of dig into. But then what it often comes out of is sort of naivety on the fund manager sort of side. If someone comes to us and they're sort of promising a 10x, and they sort of don't really or realize how hard that is. Again, you are in the business of outliers, but the distribution of fund returns tends to be a lot more on the normal sort of side compared to what you sort of get at the company level.

52:47Fair. I think there is a, you kind of need to be, what's the phrase? Get it the end of the way, then. Yeah, that one, but smart enough to give it a go, dumb enough to try. And I think that you are, we all know that PowerLore exists, as we've just spoken, is up and down the full stack. So you've got to go for it. anything else sean anything else on the specific uk fund or anything from your perspective that i haven't asked you yeah i think like the key thing for this as we sort of watch it is you know who's going to manage this because i think that's sort of a very sort of critical point whereas you know obviously the managers in our portfolio we entrust them with lots of money they raise billion dollar funds if it was one of them that were to manage it you would say look this this is looking in pretty good shape um you know almost by definition if it was someone that doesn't have the experience of running you know a scale-up fund or investing at this sort of stage um you could get quite worried um so i think like choosing who is going to do it will be quite key and again the big thing for me is that this can encourage more institutions in the uk to look at venture properly for the first time um that's gonna be a great thing no it can only be a good thing i think maybe coming back to the question I asked before, it's like any capital at any level in the venture ecosystem is a good thing, even though the US is obviously dwarfing us.

54:06I mean, we look at just one company in the US, multi-billions compared to this one billion scale-up fund in the UK. And I just like to see more coming in across the board. I've got one prediction for you. Please shoot me in the face. I reckon Bending Spoons is showing us the way and by the end of next year, there will be a a European firm dedicated to AI roll-ups with a€250 million or more as its first target, sectoring industrial and physical AI. Sean, would you invest in a VC firm looking like that? I think it would be pretty off strategy for us, unfortunately. That's a very diplomatic answer.

54:45I do love you. A few deals of the week. Mads, what's your deal of the week? and olex which is one you raised back in february where they had they were raising just over a billion six months on they're raising another 312 million dollars at a 3.3 billion dollar valuation it's james there's your three x sean there's your three x there we go absolutely within six months he's a school dropout from yorkshire he's a feel fellow he still runs another company called cool mind. It's a brain company he founded when he was 17. So he's raised more than$350 million across two companies, and he's not even 26.

55:25So well done, James. I think it's maybe interesting to look at what they are working on, which is a chip for the slow half of running AI. Now, when we're running AI, you're writing the answer word by word. And the chip they're developing here is one that has no HBM, no kind of no high bandwidth memory, which is this scarce memory we're all trying to get our hands on. Now, on the other hand, it's too small in itself to hold one model. So you have to spread the model across hundreds of chips or even a thousand. So they don't sell a chip, they sell a rack full of chips. But it's interesting, it's made for coding agents and other agentic type of workloads.

56:05And because these type of workloads, they generate lots of tokens and that's the slow half of the whole in friends game and so all lakes they've created a chip that can do that a lot faster now it's important to say they haven't shipped anything yet they haven't seen any benchmarks yet so this is still all it's all in the works it's all in the labs but it's great to see an interesting company getting money to go out and take a big swing at something really important that's great uh mine is uh there's an amsterdam startup or any i've got a bit of fixation with batteries sean you probably won't know this but i'm massively fascinated by battery tech.

56:40So mine is a battery tech company. They're all energy. Now they are building iron air batteries that store power for days. They've just raised 43 million. I wish you every success. Right, gentlemen, I think we are done. Mads, before we head off, what's happening this week that we should keep an eye out for? We're going to get US consumer price inflation for July. It's landing on Wednesday. That's an important inflation prank. Up or down? What's a bit? I reckon up. Yeah, yeah. Don't forecast that. But yes, seems likely. CoreWeave is announcing on Tuesday, kind of a big data center provider. They've got a big backlog.

57:23Can they make it grow? They're still running at a quarterly loss. Can they trim that somehow? Does AI in for demand turn into cash is the big question. Those are the two main ones. But we're still very much in the summer period. Yeah, we've got the rest of August to go. Blimey, it's only a time of pod. It's only the 7th of August, Friday the 7th. Sean, what's on your agenda this week or next week, should I say? Anything fun, exciting personally or work-wise? I think you're sort of seeing if the quiet summer period ever arrives, to be honest. It just definitely, it hasn't felt like anything. Do you get a lull?

57:54We always think we do. I think the last couple of summers, we really haven't. But between closing our own funds, investing in new funds, all the information coming out, there's been no sign of the summer lull yet this year so maybe next week who knows no no rest for you at vencap well um good luck with the new kitchen i hope chat gpt did you proud and no no doors fall off and the dishwasher isn't sellotaped to the ceiling so enjoy enjoy your new kitchen i'm hanging out with the kids this week i've got i built a little app i was telling mads off camera i built i used claude to build an app to scour the local area and find all of the cool things to do with my kids who are eight and five and I've published it it's at why don't we.uk if you want to have a look at it I built it in a few hours Claude did all the all the hard work and it's bloody incredible okay it errors and it's a bit buggy but it's bloody good and it finds things in my local area that I didn't even know existed listeners watchers thank you so much and we'll catch you all next week

59:01I'm sorry!