E467 | Dave Haynes, FOV Ventures: How AR Glasses, Multimodal AI, and Teleoperated Robots Will Reshape How We Work

0:00What if the next revolution in computing isn't about faster phones or bigger screens, but reimagining how humans and technology interact completely. I kind of somehow wonder if there will be a point in 10 years time where actually people with perfect vision will just be opting to glasses on because it helps them so much. That's Dave Haynes, co-founder of FOV Ventures, a specialized VC firm betting on Europe's founders who are redefining our computing future. Dave sees limitations in how we currently engage with technology. At the moment, most of the work we do, the work is kind of disintermediated between these 2D screens.

0:32So it's very much like text in, text out. But the boundary between digital and physical worlds is dissolving, creating unprecedented investment opportunities. We're now moving towards this like chat GPT moment for robotics where robots have trained or kind of generalized models in order to take action in the real world. Yet for years, European founders with breakthrough ideas faced a critical problem. Europe really didn't have like a go-to, like those founders were getting on a plane and going to kind of San Francisco, or they were just having to do a lot to kind of convince that like, hey, I'm slightly different.

1:03That's where FOV Ventures stepped in, with a thesis built on identifying the right ideas at precisely the right moment. Three or four years ago, there was a company called Wave Optics that got started there. Amazing, technically talented team out of Cambridge, making components for hair glasses, right? And they had investment from Optipus Ventures. That's now sold for$500 million to Snap. The opportunity spans across the continent. It's Europe-wide, but I would say those are the places that we've mostly invested in. is kind of, I haven't talked about UK yet. You have some amazing talent in UK, obviously.

1:35But yeah, UK, Finland, Germany, France, and the Nordic countries as well, China. Join us for a deep dive with Dave Haynes of FOV Ventures as we explore how the merger of AI, spatial computing, and robotics is creating a new era where digital and physical worlds converge and why Europe is poised to lead it.

2:00Here's a few words from our beloved sponsor.

2:47This show is not investment advice and the hosts of this episode may be invested in the funds and companies featured. Welcome back, everyone, to the European VC podcast. Today, I have Dave Haines with me. And Dave was so kind to say, I'm a bit nervous, man. And then I made that all go away by fucking everything up on the tech side. So now we're finally underway. Dave, let's talk about FOV Ventures. Tell us all about yourself and the firm. Yeah, well, first of all, yeah, thanks. It's great to be here. I feel like I've listened to a lot of EUVC podcasts. I know the standard is high. Yeah, hopefully this will be an intriguing conversation.

3:25Yeah, so FOV Ventures, we've been going, we had our third birthday a couple of weeks ago. We invest in the next era of computing. So it's myself and my partner. I'm based in London and my partner Petri is based in Helsinki. And what do we mean by investing in the next era of computing? Really, that's any kind of field that founders building at the intersection of AI, spatial computing, 3D, gaming tech and kind of related areas, including robotics. We love investing in founders like in Europe from day zero. And we kind of specialize in this. So we're happy to get conviction around great ambitions founders from the early stage.

4:07And we're going to talk a bunch about investing day zero, also the geo of Europe and how you think about doing that. The risks associated to your specific strategy and much, much more. Of course, also specialist versus generalist. we're going to pitch into that debate because that's obviously with such a special thesis as you have, something that's important to you. But let's start with your thesis. Expand it a little bit because yes, computing, robotics, but what does it really mean for you? Yeah. So I think for us, if you think about it, we've had all these different eras of computing, right?

4:45And this is, well, this won't be new to anyone, but we had the era of mainframes that meant a certain thing for compute then the hardware changed we then had pcs we moved to laptops we moved to mobile we moved to web 2 and there's been these distinct phases a phase and it's usually like maybe 10 to 15 years where a few things fundamentally shift so the hardware fundamentally shifts the connectivity fundamentally shifts but also the kind of software and applications and use cases that are built then also fundamentally shift and i think where we are now we're kind of just coming out of the end of the kind of mobile era, also the kind of cloud era.

5:23And we're moving into this very different time. Like we're only seeing glimpses of it now, but I think the way that humans interact with computers is just going to vastly change beyond recognition. And how do we think that's going to happen? Well, first of all, the hardware is changing. Mobile phones aren't going away anytime soon, but we do have different types of devices. So we have AR glasses, we have VR headsets, we have kind of intelligent wearables, we have things like robotics. So there is a kind of a new surface area of devices on which compute can be done. Second of all, those devices are just getting a lot more intelligent.

6:02At the moment, most of the work we do, whether it's kind of, you know, work or play or, you know, anything we do in industrial setting, the work is kind of disintermediated between these like 2D screens. So it's very much like text in, text out. We're surfing the internet, making documents with kind of words and pictures and sometimes video. But I think the content that we're going to be producing is just much more intelligent, way more interactive. And actually, if you think about it, becoming more 3D and putting all of those applications into the real world around us as well. You've got to unpack it further.

6:39because you know tell me how do you think how to what extent do you think about it from a technology standpoint versus a how we interface last use it's a bit different i think whether you're you're thinking about computing is going to change in terms of we we've gone from from floppy disks to more digital we've gone from now we now we have a next level uh microsoft chip as well, that's, I don't understand shit, but they've come up with a completely new, not element, but you know what I mean. It's not liquid state, neither is it gas, neither is it... Quantum state. Yes. Tell me, so what drives you here?

7:23Is it this UI shift or is it the shift in the application layer or is it the shift in the underlying technology? How do you think about it? Yeah, I mean, I guess we can take those one by one, right? So we are definitely, there is a number of technologies that have been emerging over the past five, 10 years, or maybe we're even on the kind of second or third cycle. So, you know, the first of which, you know, it's fairly obvious, but artificial intelligence. So AI has just, you know, really kind of in the last few years turned the corner where it's getting much more powerful, especially with things like LLMs, right?

8:00So now AI can pass a certain test or it can do very well. If I ask it a question or put a prompt in, it can give me text out. We've slowly increased that capability. So most of those models are now multimodal. So I could put an image in and get text out or get a text out and vice versa. We've also got audio models. We have a company called AI Acoustics that has an audio model for cleaning up. Audio, you also have things like 11 Labs. So we're also now kind of interacting with our computers much more via voice. We're also getting video models, which is super interesting because there's a lot of stuff that you can do around that.

8:39But actually, if you think about one thing we're really excited about is actually if you go on to ChatGPT and you say like, hey, describe how, you know, if I've got a piece of paper in front of me, like if I ask ChatGPT, how is this piece of paper going to fall? it would try and describe it but it wouldn't just know how to do that because it's just it doesn't have understanding of the world around us and the kind of physics that are underlying that so we're now moving into this world where we have these kind of world models and then physical ai which you know goes on to areas like robotics and we can talk about that in in much more detail later on so that's that's one area just like everything that we're doing is just getting way more intelligent and that goes beyond just kind of like 2d uh text and images the second thing that we then have is the kind of category that we would call spatial computing so again like if you think about most of the software that you use you're using excel you're using google you're using the you know 10th sas product again that is all just we're working like in the abstract we're working with like 2d screens boxes and buttons but actually the real world around us like that's in 3d like i'm i'm sat here i've got physical objects if i want to shop for something then i could see a 2D photo of it, but wouldn't it be great if I could just kind of like visualize it in front of me, like in 3D, or if I was at a construction site, wouldn't it be better if I could review the plans, you know, as they were kind of as a digital twin rather than just like 2D drawings, right?

10:08So that's the kind of the next capability, just having much, much richer kind of 3D applications that can kind of simulate the world around us. What's really interesting is I could do that on phone, I could do that 3D work on a 2D screen, but we're also now with kind of advancements in hardware, we're now seeing very accessible VR headsets and we're also kind of just on the tip of also having very capable AR headsets. Maybe that's going to be five years time, but there's certainly a roadmap for that within many of the big tech. And tell me just Dave, because when we first got to know you, this was very much a big piece in your thesis, the spatial computing, meta meta versus meta world.

10:50I think this is super important to bring out in this podcast, right? How does a VC firm like yours go from one place where you're very focused to saying, no, there's actually a bigger, wider space where we think we should be playing? It's super interesting. When we went out, we're currently out talking about Fund 2, but when we were raising Fund 1, we talked a lot about spatial computing. And I think that the challenges with this industry, especially when you're kind of working with these more frontier Frontier Technologies both the founder community and the VC community and the analysts like we tend to get stuck into the buzzwords du jour right it's like if you think about when we all talked about the internet back in the 90s people were talking about the information superhighway and lots of people came and said oh that means we'll get TV direct to here other people said oh this will change the way we research science and you know everyone knew that this information superhighway was coming but it was it kind of many people sort of like took took out of it out of it what they wanted now we now know the information superhighway is just the internet and it's been a huge like you know value driver and it's disrupted many many different industries so when we in fund one when we were talking about something like the metaverse i feel like it was kind of that it's like a it was a a term for its age but for us it really represented this kind of this next paradigm shift i would say when you think about metaverse for most people i think the way the media perceived it and because companies like meta came out and like owned the term i think that became much more about virtual worlds whereas i think what we get really excited about is how the next era of computing is really where the physical blends with the digital world so yes i could fully simulate that kind of construction yard and, you know, do some planning or do some, you know, fear of heights training or something like that.

12:39But actually, if I'm on that construction yard, and I've got digital information coexisting with the actual physical building, or I'm able to understand it instead of robot to kind of automate building one of the walls, then that becomes really powerful. So I think maybe what people missed about the metaverse was all of the same technologies that that were kind of, you know, this shift in the way that we were kind of experiencing applications and building virtual worlds. Actually, all of that extends into real world applications where kind of the world of like bits starts to smash with the world of atoms.

13:15I think, you know, it's famous, Marc Andreessen is famous for saying kind of software is eating the world. And we kind of think about like the world is like blending with software, like the digital is really like kind of starting to merge with the physical. And Mark, of course, since has been gotten known for saying many other things as well, comes to the source of great one-liners. I want to ask you, so you're saying, you know, we have the AI layer, and then we have the spatial computing, or AI layer, the AI being a driver of this next level computing. Then we have spatial computing. What other elements are we missing here?

13:51Then we have computing in the form of robotics, which I guess is one in its own. Yeah, exactly. Yeah, one other area to point out is that, you know, all of these kind of new applications, whether it's AI or spatial computing, it is requiring more hardware. It's requiring more compute power. Kind of importantly, it's also requiring just more like networking and connectivity, right? So certainly as we kind of like we're still early in this kind of 20-year cycle. So actually as FOV, it's been as much about investing into the kind of the tools and infrastructure layer, which is literally like we may have invested in a optics company or we have a company called Flow Computing that is accelerating CPUs to be able to do kind of, you know, kind of greater workloads.

14:39So we've invested in that kind of infrastructure layer. Then you think about it, you also then have the kind of picks and shovels. So if you're going to build all of these rich applications, what are the tools? If you're going to build a 3D virtual world, that's much harder than building like a standard kind of B2B SaaS application. So what are the sort of the tools of tomorrow? People are building 3D worlds on top of game engines today like Unity and Unreal. And that's powering like huge amount of software beyond games. But in order to get these like full simulations with which we can really create that value at scale, creators and developers are going to need more tools.

15:15So, so yeah, we, we've really in these early innings, we've also looked at those areas. And I think one thing that maybe some people miss is like, okay, it's not about just, okay, we have these new technologies. It's really about the convergence of all these technologies. Right. So if you think about like, we've invested in fund one heavily around kind of like 3d and like game engines. Now you might say, well, hang on, well, what's the link between like 3d and VR and like, game engines and robotics and actually think about the example where okay in order to understand like today robotics it you know is very much kind of on rails it you know it's very kind of complex task it's very programmed right when we're now moving towards this like chat gpt moment for robotics where actually those robots have trained or kind of generalized models in order to take action in the real world but how do you actually get the training data for those models like yes you could take a bunch of video from the real world but actually um people like nvidia are building projects like cosmos where they're synthetically generating uh video in in 3d engines and then using that to train the data and then actually if i think about it now i've got that like robot out in the real world ai probably isn't still like being able to allow it to do everything so you're still having to fall back on things like teleoperation and that's where something like vr comes in like we've spoken to a number of companies and have one in our portfolio we were thinking about like once i have my robot out in the field well i may need to fall back to a human operator who is in some form of kind of like vr or kind of ar setup in order to experience what the robot is thinking and you just can't do that on a 2d screen if i'm going into a a nuclear site to you know do something that's very high risk some kind of dirty or dangerous task you don't want to risk your operator having to kind of like be in 2D and not understand that like the depth perception of the thing that the task that I'm doing is.

17:13So we're really excited. And what we've seen is just this kind of acceleration and the convergence of these technologies. And it's really kind of at the center of all those, which is the sweet spot for FOV. Yeah. When I think about specialist VCs, I am always like I'm digging into understanding understanding exactly what are you seeing technology-wise, society-wise, basically what I call almost inflection. So what are the things that are happening that are lining up perfectly to make the sector, the vertical, the technologies you're focusing on be perfectly timed right now? But that's the one side to a thesis.

17:56There's the other side, which is how does it align with the firm. And that to me is just as important. You might be focusing on this, but if you have no credibility, no empathy for the founders building in that space, that's going to be super difficult. So I'd love to ask you, how does this side of the equation look like? So how does you and your partner and the whole team at FOB, how do you, so to say, match up against this thesis? yeah and i should say just on the on the specialist thing it's interesting because like we think of ourselves as a specialist fund and that we have a very clear thesis which kind of rules out maybe 80 percent of all of the other deal flow that we could be looking at when we see a deal it's like right this is in thesis we can go very deep we see how the all those things relate to each other but it's funny because i've spoken to other gps and some lps where they say but hang on this doesn't look specialist because you're talking about on the one hand like you know the the game industry on the other hand you're talking about construction and now you're talking about robotics so and that's probably a topic in itself but i i kind of see it as there's definitely some funds where you are like right i'm fintech focused or i'm climate focused and you're just in those industries and sectors whereas it might be like hey i'm a i'm an ai fund or i'm a deep tech fund and then you're probably looking at right we have these technologies but what are the areas that it can be applied to now so what we were investing in in 2022 may be different to what we're investing in 2026 i think we had in our thesis that robotics was going to become more powerful because of this tech stack but it wouldn't have been the right thing for us to invest in then we've had to wait for those technologies to mature but i'll pin that to answer your original question and yeah i think that's another thing that you know founders and And VCs always think about this kind of founder market fit.

19:46Why are these founders uniquely placed? And I think the equivalent in VC is like a GP thesis fit, right? Why is this person doing it? I think that there's a few things there, but the simple answer is that we think this is a kind of 15 to 20 year trend. We're probably in year five of it being kind of investable, but actually between Petri and I, we've actually been looking at the space for, you know, for kind of getting on for seven, eight years now. So when, so I started my career investing, I was a founding team at SoundCloud and moved into VC in 2014, moving to Seedcamp. And I think for both AI and VR, it was still like the previous cycles of those things.

20:28The AI was still breaking out. We just had like ImageNet. We had the early reinforcement learning and machine vision. And we made a few bets around that. We also had the kind of first go around of VR when Oculus was sold to meta. And I made a few investments there in that space, which are actually still going and doing really well today. But it was still very early. Not many other people were looking at those spaces. I actually then moved after Seacamp. I moved to CVC. So I moved to HTC. HTC had kind of got to this kind of, you know, next era, like maybe a little bit early, maybe too early, where they'd sold their smartphone division to Google and were investing in things like VR and AR and 5G.

21:11so whilst I was there like Petri was at a an accelerator in Helsinki called Nordic accelerators NXS or Nordic XR startups so we we kind of figured that actually we're probably the only fund at that point in time it was still early but we had strategic reasons to be investing in those sectors and we would just see the same deals and we would see a lot of other VCs just not kind of wanting to kind of dive into those slightly more frontier topics they were much more comfortable We had the era of B2B SaaS with playbooks that made sense for most of the VC community in Europe. And then we noticed that actually in US, you had already maybe four or five other funds that had these kind of specialist lenses on similar topics to us.

21:55Like you had Evernote LDV, that is a computer vision focused fund. You had people like Gregor Anorak. Gregor was one of the first investors in Oculus. us you had a specialist vr fund called the venture reality fund um that was kind of co-investing us in some of those vr companies but europe really didn't have like a go-to like those founders were getting on a plane and going to kind of san francisco or they were just having to do a lot to kind of convince that like hey i'm slightly different like you know like do you want to give me the first check so yeah so although i think that that's where the gp thesis fit has come in Like we've just spent all of our time looking into those areas.

22:33And I think over those years, some of these other trends have evolved and we've kind of, I think what started off as a bit more of a narrow focus, there's definitely a kind of broader surface area that we call kind of spatial computing, or we generalize into saying, hey, this is the next era of computing. When we launched the fund in, well, when we started raising the fund in 21, I feel like most of this stuff was kind of non-consensus. is kind of good and bad that I feel like today, some of these, I mean, VR is still a bit of a non-consensus area, but I certainly think things around like physical AI, computer vision, robotics, those are much more common sense.

23:09And now it's like, okay, well now the rest of the world like kind of believes us too. Now it's like as a fund, how do we kind of add value or how do we get to those deals maybe before others do? It's funny because on that part around the market, I always describe it in venture as you almost only want to be in one quadrant of the competition matrix, right? You only want to be in the one where everything is taking off and there's a lot of players. If everything's not taking off, it's not going to be a good place. Those are not going to be great successes. But on the other hand, if everything's taking off, but there's not a lot of competition, you also don't have co-investors and follow-on investors.

23:49So Venture is this weird space where it's very hard to go outside of where the majority of the investors are, which pushes us into a competition situation at the same time. You, of course, can't do it. It's just incredibly hard and you have to have a fund model that then allows you to protect your investments in the follow on route. Exactly. And I think, you know, Fund One was a 20 million euro fund. It was very much like we didn't want to oversight the fund for the opportunity at that time in 2022. Obviously that's growing because the opportunity is growing and we've proved that we can get really great deal flow in this space, but we have to be a very collaborative fund, right?

24:27So we're often kind of following in seed or kind of leading a pre-seed and being able to bring other angels and other kind of early believers. But it meant over the last kind of four or five years, we've really had to work hard to know, right, who are the specialist funds that we can collaborate with? And now there is a nice kind of number and there's like, you know, some of the gaming funds have been the ones that are investing more in like gaming tech like we've done deals with those guys um like Andreessen Horowitz builds at has a program called speedrun where they do games investments but they love things that aren't just games but it may be like game adjacent now we also have in Europe I think there's a really kind of great pool of like deep tech vcs so when we're thinking about the bets that we've made in optics or semiconductors or power those there's a new bunch of funds they're willing to take those slightly kind of longer term bets and then the challenge is just going back to the earlier thing it's like timing so we believe in a world where in 10 years time like ar glasses will be kind of mainstream i actually wear the meta ray bands probably like 80 percent of my uh of my week i'm not wearing them now maybe i should have done we we hold that belief but then as as early stage pre-seed investors i now i have to make a bet on like okay well is there an investable opportunity and actually like you know three or four years ago there was a company called wave optics that got started there amazing technically talented team out of um out of cambridge um making components for air glasses right and they had investment from octopus ventures that's now sold for 500 million to um to snap if i'd been investing in like ar applications and i can name quite a few that have pitched us over the years like those are two market dependent like these require like you know hourglass to be mainstream so you can invest and it's the same with vr like there's certain categories like gaming is booming like the amount of revenue that can be made on on the quest store is is insane the biggest vr game was actually made by two two guys in in uh in the czech republic and they just made something that was very unique to the platform that ended up getting sold to meta and those guys are now very you know much better off for it so yeah we just have to i think that the trick is just understanding the timing and knowing where the kind of downstream vcs are willing to bet and who those believers are.

26:42There's no point raising money from somebody who just doesn't have conviction around this space. Just move on to the next investor. I want to just ask you on the AR glasses and you're obsessed with this space and really believe that this is going to change things. We now have Dave changing his glasses. For this special segment, that makes a lot of sense. And Dave, I want to ask you as a power user, using the glasses 80 % of the time. I've many times thought, I'm goddamn, I'm using ChatGPT a lot. I wonder if I had it plugged in, or in this case, Claude plugged into my ear all the time, how much more would I be using it?

27:21So tell me, what is the experience? Yeah, indeed. Well, I think it's kind of ironic because I'm already a glasses user, right? So this is already a form factor that I have on my head every day. That means the utility I require from it is probably less than what you may need from it and ironically it's like people are people spend a lot of money to have laser eye surgery so they don't need glasses i kind of somehow wonder if there will be a point in 10 years time where actually people with perfect vision will just be opting to put glasses on because it helps them so much in their day-to-day lives so yeah i'm going to be interesting to kind of trap that one but for these i have the i have the context for people can't see the video i have the meta ray bands on and actually the big thing here is these aren't ar glasses but they are smart glasses so the key things that they do is like i'm wearing these anyway they have my prescription lenses and they have transition lenses so i also use them as my sunglasses but what's great is i just have audio like this is this is kind of the the best and easiest way just to kind of be able to consume audio i have a 12 year old and a 16 year old son and my 16 year old pretty much lives most of his life with his airpods on and then when he's with us he has them in transparent mode and when he's like on the bus or at school he has them in in closed mode so we've already got to the point where there's a generation who just are permanently used to kind of having audio and then interacting with audio and actually the the microarray and the speakers in in this device are phenomenal so even if you just didn't have like cameras or any ar functionality then then that would be great and then i think the reason why met to release these is because they had ambitions towards full augmented glasses or what they call their project um orion which we've been tracking for years but i think they almost accidentally stumbled because they have big ambitions in in ai they actually realized that the glasses are in a way the most human interface for ai i could like you said you could i'm using chat gpt they have a like if you're a premium subscriber they you can interact with voice and you're just having a conversation with the ai i can now do that when i'm walking down the street if i'm walking to meet a founder i can say hey like tell me about this company and i can be you know those same things i'm putting into chat gpt i could put i could just ask through voice now that's great i could also do that through headphones right so what's why why is glasses also useful i've got them on all the time but also these have cameras right so i can i can take a i can say i meant to take a photo or i can just kind of press a button it will take a photo and now I can interrogate that photo using that as well so I could I could look at a sign I can say hey what does what does that tell me what's even more interesting is now you'll have persistent like video feed so I can just be walking around in the outside environment now at the moment it relies me to prompt meta to do something useful like I'm walking down a street and I kind of say hey based on what I'm looking at what can I do but pretty soon these things will become context aware right so pretty soon the ai will understand me my preferences where i'm going what the environment is around me and it can start contextually just providing me information that i didn't even quite realize that i needed this is before we've even touched ar then and we have this this is i also have the snap spectacles on my desk and these are now a standalone our device once you then add even if you just add simple visual augmentation um like if i was able to read that like if i'm on the tube i don't want to be talking to chat tpt but if i could be you know kind of like prompting it and then just reading the readout of it or if i'm walking to a new location i don't know where i am just having a single like single color left or right in my field of vision is killer or imagine i'm doing a skype call i'm not skype calling on to skype because i'm not a zoom call and i just have zoom up in my thing on my kind of like 2d peripheral like and this is just we're already like some people wonder well is this going to be socially acceptable but why is it sociably sociably acceptable to be walking down the pavement by looking at my phone i'm bumping into people in crowds i'm i'm not interacting at the dinner table so we can get out of our phones like there's been lots of experimentation with like the humane wearable and things like that but i think actually when you think about it the glasses form factor is is probably the most human and the most natural device to center some of those these kind of digitally augmented superpowers around on the day-to-day how many prompts do you have to your glasses so i'm usually asking very simple things and it's funny because in in europe at the moment a lot of the most powerful ai features thanks to the lovely regulation we have here in europe a lot of the ai features are still only available um to the us but yeah i'm using the glasses like most of my waking days there's still some you know if i go out for drinks to the pub i don't take it because there's still a well i think there is still a question about the weather society accepts people having cameras on their face i think that's that's still one thing to be tested and it will be tested later this year when when meta released the meta version but yeah i'm i would say i'm probably using the ai functionality only maybe like two or three times a day but i am doing like a lot of my calls uh and a lot of my like listening to podcasts um directly on the devices it's already providing utility and then we can get it out of the box and i think that's why like you know a lot of people say when is the iphone moment for something like vr i think the challenge with vr so far has been that you you need a really good reason to put like a fairly heavy thing on your face and there are lots of really good reasons to do it but like it hasn't got that like with with the glasses form factor this is this is a device i'm already used to using on a day-to-day basis i it's already worth spending 300 euros on these blouses just for the audio and now we can start to build functionality i have to say i am personally most excited about the future that seems to be quite close around the corner, especially considering if you put on glasses that are able to capture everything you do and see during the day.

33:33And then for me to be able to do a wrap-up conversation that lasts 10 minutes of the day, would say, tell me based on today, what should be the things I have followed up on and made sure actually happened based on today? Because I think we have so many conversations during a day where there are incredibly powerful follow-ups or actions to be taken, but we just don't get around to doing them. And when you then match that with, if you, as an example, do it end of day and you then say, all right, please write up the to-dos for me in the morning for me to start tomorrow with. Like that would be so powerful.

34:12And I see no reason why some AI that had been fed my inbox, fed my WhatsApps, plus also what I experienced during the day, it would very quickly be a much better Andreas than I am. Exactly, exactly. And, you know, I think, I mean, one thing to be clear, I think if you look at the portfolio of just over 30 companies we've built in Fund1, actually, if I count the number of companies that are pure play AR and VR, it's still probably three or four, right? You can count them on one hand. And those have been in bets like with VR, like we've made a bet in gaming. I think most of our bets in Fund1 in this space have been more like, what are the kind of stages?

34:50Like, what are the developer tools? Like I was saying earlier, what are the picks and shovels that the platforms need to bring these ambitions like closer to market what are the developer tools that developers are going to need to actually build these you know these are very complicated like applications that we're talking about that kind of understand the real world and can blend your digital world with it i think when i talk about our thesis people may be meeting with a little bit of skepticism but this is this is a future that now like the five six big companies like NVIDIA, Google just released Android XR, Apple released Vision Pro, which is more of a long-term project over the next five to 10 years, like Metra there, like Samsung are coming back into the market.

35:31So again, this surface area, it's still, I think there's still lots of questions around what is the startup opportunity on the application side, but yeah, when it emerges, we're going to be there and be getting to those entrepreneurs first. I think this is the perfect pivot into talking about something you spoke about earlier as well, which is, of course, the criticality of timing. Timing is absolutely crucial in venture. You started out saying, well, when I was at HTC, maybe we were a little bit early. It was more for strategic reasons than it was financials. Tell me, how do you think about timing risk today?

36:08Yeah, really good question. I mean, yeah, I think the first thing is, the example I gave earlier of WaveOptics, you know understanding where we are in the technology cycle and understanding kind of where where the value is um i think that is that the first thing the second thing is i mean we're a relatively small fund i think we as a smaller fund we definitely have more appetite for companies that may end up becoming m &a targets um so in the case of wave optics if you're a pre-seed investor and maybe you've got like you know four or five percent exit for a 20 30 million fund like that's going to return the fund and then some so i think we also have to think about okay if there is a likely if this is more of a kind of component where there's going to be an m &a path with one of the bigger tech platforms just making sure that we have the right economics and that we're in at the right time so that's why we really focus on this kind of day um day zero investing and then we also have to think about like what are the opportunities where okay there is a startup opportunity here but you know could apple just build this themselves it's a bit of a cliche like most most companies you know one obvious vc question is like will google build this but it is you know we've certainly seen companies pitch us you know things around like kind of you know generative ai or kind of the visual side of that where you know they'll pitch us this amazing application but then actually it turns out that that's something that like meta or snap or whoever it is want to do in-house because it's more of a core application so yeah definitely challenges But I think, again, that's why you kind of need to be specialist in this space, because that's quite a nuanced discussion to have with a founder.

37:47You can't just suddenly say, hey, maybe we should add some AR companies to our portfolio because you need to understand how to navigate that type of business. In your notes for the conversation, you had a statement that said that most founders are living six months in the future. We're sometimes living six to 12 months in the future. I'd love for you to unpack that because I think there's something very important in that. Yeah. So I think all of the great founders, there's a certain element for them. They've seen something wrong in the world or they've seen a future opportunity in the world. right so most of the founders we work with are more in the kind of you know they've seen the opportunity um the really great example is is someone like um you know justin can who built justin tv that turned into twitch and then became a really huge acquisition from from amazon and if you hear his early investors say well like justin tv started out as this crazy guy who had like a live streaming video and would just kind of he was just doing like live blogging because he knew that cameras plus a 4G connection suddenly made a new behavior possible, right?