Ip Group (LSE:IPO) — Technology Trailblazers

0:01 Good afternoon, ladies and gentlemen, and welcome to the IP GROUP PLC Technology Trailblazers Investor Presentation. Your screen. Just simply type in your questions at any time and press send. The company may not be in a position to answer every question received during the meeting itself. However, the company review all questions submitted today and publish responses where it's appropriate to do so. Before we begin, we'd like to submit the following poll, and I'm sure the company will be most grateful for your participation. I'd now like to hand over to managing partner managing partner of technology at IP GROUP, Mark Riley. Good afternoon to you, sir.

0:39 Good afternoon. Thank you very much, Mark, and thank you very much to everyone who's joined us this evening. I'm sorry that I'm coming from you to you from a virtual a futuristic virtual office building rather than our this is almost as impressive as office in Pancras Square just behind Kings Cross. We were hoping to have an in person event today. Unfortunately, being scuppered by train strikes, prevented which several of us, including our speakers reaching the venue. So apologies, this had to be virtual, but thank you for coming along. So today, we're gonna hear from two of The UK's most outstanding innovators, the innovators that we have worked with over the last several years at IP GROUP. We've invest in companies that Nikki and Tom have founded. But first of all, I just wanted to spend a few minutes just reacquainting those of you who know about IP GROUP and also introducing those of you who don't to what's happening currently in the deep tech portfolio.

1:31 My name is Mark Riley. I'm the managing partner of the deep tech side of the IP GROUP business. Those who are familiar with IP GROUP may know that we split into a life science division, run by my colleague, Sam Williams. And then we have our clean tech function, which we've recently branded Kiko Ventures, and they're doing very exciting work investing in innovation in the clean tech domain. But today, this event is organized by the deep tech team. And the deep tech, part of IP GROUP, hopefully, your speed reading is up to scratch to, cover you the disclaimer before I launch in. The deep tech part of IP GROUP is focused on delivering value through growing innovative companies that enable and secure the digital economy, create new human capability, and generate prosperity for all. We are an impact investor. IP GROUP has been an impact investor for all of its twenty years of existence since long before. It was the fashionable thing to be an impact investor. We have always invested in technology, targeted in improving human lives, improving our future, and that remains the case in deep tech. We do that in very broad ways in deep tech and investing in all sorts of different enabling technologies across the engineering and information technologies domain.

2:38 This is the current IP group team. You might not want to read through all of the individual CVs there, but the key things to take away from this slide are the background, the educational background of all of this team is is is an engineering, is we see a couple of PhDs here, these people with scientific training, and then of course, the depth of experience in venture capital. This is a very experienced team that we've built in the on the deep tech side in IP group. And as a result of that expertise, we've had a very successful couple of years recently. And on the left of this slide, you will see just some highlights from those last couple of years, some exits that the team has secured. First of all, we sold a company called WaveOptics to Snap Inc, a listed company in The US for about over over $500,000,000, which we believe was Europe's largest ever exit for a venture backed deep tech company. I believe DeepMind was sold for sort of similar amount, but this is kind of, we think, first or second most most valuable exit for the European deep company, so very successful outcome there.

3:39 A year or so ago, we sold our stake in YoYo Wallet to a company called Solpay International Payments business. We sold process systems engineering to Siemens, and just this last month or so, we sold a company called Reinfor to the global leader in robotics process automation, UiPath. And in addition to all of those exits, we've also got a whole handful of assets now that we've gone come in very early. We've come in at the sort of sub million pound value, very early stage of these companies and growing and being with them all the way through the journey up to past a million, 100,000,000 enterprise value, including Ultraleap, whose founder you will hear from later on in this webinar, including Garrison Cybersecurity, which got through the the £10,000,000 revenue threshold very quickly and is growing very fast. Featurespace similarly, which is doing great work and revenue growing very rapidly.

4:26 So it's been a really exciting period over the last three or four years in the deep tech portfolio, some great success stories there.

4:33 And as a result of that success, we've started to attract some really top tier venture capitalists to invest alongside us. We for for for a period of IP GROUP's history, we were plowing a relatively lonely thorough investing in in deep tech in The UK, but that's that's not the case anymore. People have seen the success that we've had, and people have seen the the benefits of investing in in early stage deep tech innovation. And so now we have investors like Dawn Capital, Insight Partners, Crane and Excel investing alongside us in the portfolio. All of these investors have come into our our portfolio assets in in recent times. So the deep tech team today is focused on four broad areas. Our focus areas are applied AI, particularly in the area of cybersecurity at the moment. We have a couple of companies in that domain.

5:19 We're interested in next generation networks. Our current communications networks are not are not set up to support the use cases of the future, the kind of virtual reality use cases that Tom will talk about later, the the the IoT Internet of Things, devices. Our communications networks need to upgrade to support all of those use cases, so we're interested in technologies that will facilitate and enable that upgrade. We're focused on the evolving interaction of humans and machines, and Tom's company Ultralip is a good example of that. And finally, we're interested and we've got some exposure to this this very futuristic area of evolving computing technologies. And we've invested in a couple of companies that are developing new forms of qubits to build a quantum computer. We're also exposed to some neuromorphic computing activity as well.

6:08 So those are our areas of focus currently, and we are open for business. These I just picked out three examples here of recent investments that we've done, new startups that IP GROUP has invested in on the deep tech side over the past twelve months. And, we are focused on on on early stage innovation and and looking for new opportunities in in that domain, for technologies that can have a positive impact on the world. Which brings us nicely to our first speaker. Professor Nikki Trigoni is the founder of Novenio, a spin out company from the University of Oxford. Nikki is still a, professor and head of the cyber physical systems group at the University of Oxford, well as being the CTO of Novenio. And Nikki is here to talk to us today about the use of artificial intelligence and how it's impacted many aspects of our daily lives, and she'll present some ideas where we should be looking for the next generation of opportunities.

7:04 Over to you, Nikki. Thank you very, very much, Mark. It's a great pleasure to be with you today, and thank you for the kind intro, and thank you all for joining. I would like to start just by, you know, reminding everyone how far we've come so far. We have all witnessed tremendous sort of impact of AI in different facets of our society, anything from using AI to increase efficiency and decrease costs in businesses to analyzing vast amounts of data, scientific discovery, all the way to making life more fun. And and we've come a long way, and we need to pat ourselves a little bit on the back because we've achieved what we call the impossible thing. We've taken myth, and we've turned it into reality. Right? So I was in Crete this summer visiting different museums and sort of talking to my kids about, you know, legendary automata in workshops of and sort of.

8:12 And here we are having, like, amazing state of the art warehouses with fully automated processes and robots, autonomous sort of vehicles carrying out all sorts of tasks. And, you know, from a legendary sort of robots that sort of can roam around areas, whole islands, and sort of protect places to being able to explore space. Right? And not just sense information there in but also collect this information and coordinate each return into Earth. So the question that we have is, like, can we, as scientists or investors or entrepreneurs, join in to try to turn today's myth? And what is today's myth?

9:00 Right? Today's myth is artificial general intelligence to potentially envision this becoming tomorrow's reality. Right? Can we get into a world world where we can barely distinguish if we have a conversation with a machine or a person when we are sort of asking for a certain service? Or can we expect a machine to just just enter a home and figure out how to make us a coffee? Or can we have a machine study for a university degree and or solve a complex financial problem better than any of us in this room potentially could do. I mean, this all sounds very futuristic. Right? And we are all optimistic, and we like a bit of futuristic reverence for AI. But, you know, there are really good data points that shop that sort of show there has already been massive progress in the last few years. And an example is one linked to being able to generate natural language. Right? So OpenAI GPT three, as an example, is one of the largest neural networks.

10:06 It has really moved the needle in term in terms of sort of language generation. And what it does is takes input text and pretty much predicts output text. But imagine, like, all the possible opportunities from being able to, for example, take a piece of code, legacy code, which is a piece of text, and predict what was the intention or what this code was purported to do. Right? Or take a a piece a small piece of text, the title, and try to produce an essay. I'm sure my my sons would love for this to happen. Or to basically ask questions and and get great service from a machine, a great customer service.

10:52 So there there are huge tremendous opportunities with that come with using AI for a generation of language. And but language is just one mode, right, of information. So it's just one piece of the puzzle. And future AI systems, I think, will increasingly use computer vision together with natural language. And so a good example of such a versatile system is Wudao two. It comes from Asia. Right? And it is able to connect language and vision. It uses a mega model as one would expect and, you know, mega compute power and a mega data. Right? But it can do quite a lot of interesting things. It can multitask.

11:40 It can take a piece of text and sort of generate, you know, photorealistic images or

11:48 even pieces of art. Right? Or it can take an image and and sort of create a description about it or a story about it. So these these are really, really interesting applications that could have implications and sort of bring progress in different sectors, including health care, entertainment, and so on.

12:10 But besides language and computer vision, mobile autonomy, I think, is one of those parts of applied AI that are really at the heart of what is going to to to sort of excite people in the coming decade. So we've seen massive commercial and government push lately, and we're expecting driverless cars to be with us in a few years. Right? And there has been huge backing and and funding for this to happen. And, you know, some of us would sort of fear or are happy that our kids may never have to get a driving license in a few years. And and a lot of people are working in that direction. So we have got a global commercial tech race with lots of different companies being involved. Right? But as an example, I was looking at the recent press release, a release on Tesla, and you could see, like, the incredible masses of training data that is is is continually streamed from hundreds of thousands of customers, right, and fed into simulation and then kind of used to to continually train models.

13:24 And, interestingly, we also see some of those companies investing in hardware and trying to come up with new hardware, for instance, the Dojo cabinets that sort of can replace, you know, thousands of GPUs with a few cabinets of those of this new hardware. But one of the more interesting things is that there are quite wild predictions. Right? That more more than half of the of companies in the future, in the next two decade decades, are going to be using mobile autonomous systems in some in some flavor.

14:04 And if you look at this, it's not very surprising. Right? Because autonomous systems have application in all sorts of areas be besides driverless cars. So you can see their application is in delivery vehicles, multi legged robots, soft robotics, and fruit picking robots, you know, all the way to space and to ocean exploration. And there are many, many challenges that we have to solve for for those applications to come to fruition. Right? So first of all, we have to solve challenges to move mobile autonomous systems from shared autonomy to full autonomy and to decide whether it will be acceptable to have a full autonomy or we always have to keep the person in the loop.

14:50 In our lab, we are working on robust general and adaptive models for localization of vehicles and controlled mobility. And this is very, very important because we need to be able to capture all edge cases, cases where the weather can be very unpredictable, whether it is sort of cases that we haven't seen before. And as Mark sort of alluded before, we need very high speed networks, high bandwidth networks in order to make sure that we have good communication between the vehicles and devices. So these devices could be roadside units or it they could be sort of devices, wearable devices or smartphones of pedestrians. Which brings another question, like, how can we innovate to make sure that some of those machine learning models can be squeezed and compressed and pushed into the edge and and so that they can work not just work even in the absence of network connectivity, but also work in a privacy preserving manner.

15:57 And, of course, we will see a lot of opportunities for companies that will then have to work on how to generate sort of verifiable models or explainable models or how to bring accountability to novel AI systems. And eventually, the most important thing is for people to find ways where this kind of new technology can be trusted by by individuals and put to great uses. Now in our lab, we are looking at the next generation of autonomous systems. And we look at the sort of more challenging scenarios, for example, emergency scenarios, blue light applications, where you may have very limited visibility, so sometimes complete darkness, a thick smoke, etcetera, etcetera.

16:46 But we also have almost often limited connectivity or no connectivity and no pre installed sensor infrastructure. So these are typically unusual situations where there is a lack of training data, you cannot easily trust sort of supervised learning techniques that have been trained in one with a few sort of specific datasets to work well in these new environments. This creates a new need for new sensor modalities as well that can see in the dark. So we are looking, for example, at thermal cameras since normal RGB cameras would not work in this scenario or millimeter wave radar instead of LIDAR. And these new modalities or a little bit less explored modalities have their own individual sort of issues they have. For instance, with thermal images, we we don't have strong visual features.

17:40 With millimeter wave, you've got very sparse point clouds. So there are a lot of opportunities for universities and spinouts to work on developing algorithms based on those noisy sensor modalities for localization, for control, for segmentation, and also for people detection. And, of course, you know, the dream or, you know, a wonderful thing for, like, a conference of mobile autonomy and sort of language generation is to be able to combine the outputs of those bespoke machine learning models that can do localization or control or mapping or semantic mapping of the environment to kind of tell the story of what happened in the emergency scene automatically.

18:26 Right? The team entered into the building, comes from semantic mapping, went up to the 2nd Floor, comes from a localization system. So being able to create narratives based on scene understanding. Besides, of course, creating narratives, an even more exciting direction, one that Tom is going to talk about later on, is creating new environments and participating in this new form of mixed physical and virtual world where it's using sort of techniques such as virtual reality, augmented reality technologies for digital twins, metaverse, omniverse, and so on. And here, there are huge opportunities for context creation and also for content used by participants.

19:16 And we need to think about new models of interaction, learning, creativity, collaboration. And we also need to think about companies that will address the challenges of, for instance, how do we do identity management in those spaces or interoperability across those metaverses or even sort of pushing the boundary to find that new financial models or new forms of government in this kind of kind of mixed environments and the ethics and regulation sort of frameworks that come with those.

19:53 But as fun as it can be to to basically entertain this possibility of mixed environments, I think one great responsibility for us at the moment is to care about our physical environment, and AI can really have a fundamental impact in this area. For example, we have great platforms, digital platforms, social networks where we we can receive a huge amount of data. And we've got AI models that can crunch through this data and find patterns and sort of

20:27 highlight issues around the weather about biodiversity, about the usage of resources over time. We can have AI that sort of can improve the efficiencies in companies that actually now are responsible for wasting energy. And we can have AI spinouts and bigger companies looking at sort of new processes and new materials for optimization for cleaner energy production.

20:57 Another frontier for AI, of course, is health care, and this is particularly close to my heart because of Navenio, of course, where we're using machine learning to build indoor positioning systems that are completely infrastructure free. And then we have shown that we can really optimize work in hospitals. We can sort of double

21:21 the the throughput of tasks in hospitals, making a huge impact on the NHS and hospitals more globally. But there are other many other sort of researchers and spinouts and bigger companies that focus on apps that can provide people with the ability to take control of their own health care. For example, this app for smartwatches, Apple smartwatches that is for dementia sort of patients. And, of course, the use of AI for diagnosis of disease, particularly interesting diagnosis of disease that can happen very early at the very early stage and which can sort of lead to curing the disease much faster. And, of course, drug discovery is an another big area of machine learning.

22:12 A couple weeks ago, I was involved in a really, really interesting exam of a thesis on physics informed AI. So was sort of exposed to a number of diff different interesting problems, how we can study complex processes, aerodynamics, thermodynamics, earth and space science by combining the laws of physics with data driven learning.

22:39 In chemistry, AI, huge, huge impact there as well. A very good example is simulation of molecules in order to study different types of molecules and how how their properties and, of course, interactions through simulation. And in biology, anything from, you know, studying three d structure and fold the and and the folding patterns of proteins to more

23:06 to to to sort of artificial antibodies and even, you know, study artificial bundles of brain cells and and their sort of development over time.

23:20 And in this sort of use of AI in science, but also in other applications, for instance, secure cryptography. As Mark mentioned before, there's great, great promise in trying to combine more classical computing with quantum computing,

23:40 both hardware and and and software. So we see opportunities for new hardware. Here's an example at the top from the Oxford ecosystem, but also algorithms and compilers and software libraries. There's a lot of big players in this area as well as new spinouts, very promising spinouts that could sort of change the way in which we address complex problems today. And and this could have implications to many, many other sort of applications, including security banking and even and and finance.

24:16 So ace AI's playground is definitely vast. I have only had time to touch on some some of the frontiers of AI, and I think we're going to have a very interesting next decade looking at a variety of different problems. Definitely fertile ground for innovation, definitely for investment as well. But one thing to bear in mind is that AI is not a standalone technology. There is the biggest opportunities will come when it's coupled with other emerging fields. And, of course, at all points in time, we have to bear in mind that, you know, humans have to be at the front and center. So accepting, trusting, regulating services and products that come from machine learning needs to be a priority for all companies and spin outs that sort of work in this area.

25:07 Thank you very much.

25:11 Nikki, thank you. Fantastic talk. I couldn't write fast enough for writing down all the ideas for areas that that we as IP GROUP might invest in and new opportunities for for innovation. Thank you. So no questions for Nikki at the moment. We're gonna do a q and a panel at the end, where there'll be opportunity to ask both the speakers your questions. So please do save those up and and put them in the q and a either now or the panel session towards the end. But for now, we're gonna move straight on to our next speaker.

25:42 Doctor Tom Carter is the founder and chief executive of Ultraleap. He is one of the brains behind the world's first and only mid air haptic solution and more of our lives now are lived online. And Tom will explore where the value lies in this virtual world and where investors should look to benefit from this emerging opportunity. Over to you, Tom. And then thanks a lot, Mark, and hello, everybody. So we're gonna talk about the metaverse, a playground classroom or workspace. But to give a bit of structure, a little bit of simplicity, we're gonna work through what, where, and how, and a breakdown of what this metaverse thing is. Because you can't really start a conversation or a presentation about the metaverse without answering the question, okay. Well, what is that? What is the metaverse?

26:31 And the reason for that is there isn't really one clear solid definition that's widely accepted. Lots of different people have different opinions and views and visions as to what the metaverse is or what the metaverse will be. So Mark Zuckerberg and Meta, formerly Facebook, they believe the metaverse is gonna be this virtual space where you can hang out with your friends and be social and have fun.

26:58 Microsoft believe the metaverse is a virtual space where you're gonna have meetings and work through PowerPoint slides together. NeonTech, the makers of Pokemon GO, believe the metaverse is not actually a holy virtual space at all, but is the real physical world that we all exist in today augmented with virtual content, information, virtual objects, virtual artwork to to sort of fuse the virtual and the real.

27:30 And then, of course, brands all have their own ideas of the metaverse as well. So Walmart, I think the metaverse is a place where our children are going to go and do their shopping in Walmart stores with their favorite cartoon characters to prepare themselves for shopping in Walmart stores in the future, I'm sure.

27:47 So there's lots of different opinions and views as to as to what the metaverse is gonna be. So to try and make a little sense of it, it can it can help to jump back in time and look at, okay. Well, what was the last time that we looked at something like the metaverse that had this kind of confusing view as to what the future is gonna be and sort of all encompassing paradigm shift? And a great example is the information superhighway. So this is a a magazine cover from Popular Mechanics. I think it's in January 1994, and their their leading story is understanding the information superhighway, trying to educate people as to what this information superhighway thing is all about.

28:30 And really, it's this it's this place, this technology that's gonna enable you to shop and bank and learn and be entertained. And it's gonna be delivered via interactive TV, which, of course, now in 2022, we know that, okay, yes, you can do all of those things. You can shop, you can bank, you can learn, you can be entertained. Some of them come through interactive TV, but we don't use the term information superhighway anymore.

28:57 Maybe you can look at that and go, well, yes. We call it the Internet. So let's look at the Internet from this is from 1995, so a year later. Here is the Internet. It's this little the the biggest blob in the middle, and as as they were trying to convey the Internet is part of the matrix. France accesses the Internet via Minitel. And if you want to interact with a commercial service or company, you won't do that through the Internet. You'll do it through this little commercial services bubble at the bottom that somehow interfaces with the Internet, but isn't the Internet itself. Okay. So even if we jump down to the technology level, sometimes it can be hard to see the future and see exactly how this is all gonna play out when you have a a paradigm shift coming through and how people are gonna go about tasks that span so much of their lives from shopping, from working, from from playing, and so on.

29:57 So with all that said, and having just said that nobody really knows what's going on, I've been asked to come here today and to tell you what the what the metaverse is. So here's a here's a stab at it at least. Rather than trying to define it into something specific, I I think it's useful to think about the the general concept of how it's gonna change what we do.

30:19 And, really, it's the next computing paradigm where the content, the virtual stuff is not constrained behind glass. So today, you have whole virtual world constrained behind glass that you interact with using keyboards and mice or peripherals or tapping on the outside of the glass. And the metaverse will enable either you to go through that glass and fully into that virtual world, or that virtual content to come back the other way through the glass and co inhabit the real world alongside you.

30:54 And really what this means for you as a as a person using this is it allows you to be present rather than something happening over there and you controlling it from a very separate place. It allows you to actually be fully immersed in that virtual world and have that feeling of of presence.

31:15 So if that's what it is, okay, to give a bit more color to that, where is it? Where are the applications? Well, we ask the question, is it the playground? Is it the classroom?

31:26 Is it the workplace? Kind of where is this going to be? And much like the information superhighway, it's gonna be all of those. So looking at the playground, it's going to really take entertainment and gaming to the next level. Even just looking at games and sort of things you can play in in virtual reality and mixed reality today, there's some really exciting content. On the right here is the most popular game in VR to date. It's called Beat Saber. You have these two lightsabers that you hold in your hand, and these three d blocks come flying at you, and you've gotta destroy the blocks with your lightsabers in time to the music. So it's kind of like an augmented version of Guitar Hero, but rather than standing in your living room tapping on a guitar while there's a rock band on the television, you are actually a Jedi with lightsabers there in the space doing your thing.

32:19 And then on the left is one of the most recent demos that Mark Zuckerberg showed this off last week, I think it was, or the week before. This is a new mixed reality demo that they've that they've produced where you can you can do fencing, you can fence your friends from separate places. So he has a headset on, the headset has cameras that can see his living room and beam the the video image of the living room through so he can see it. And then you can overlay the avatar of his friend who's in their own house in another part of the world. And then they can fence each other, and they can jump back and forwards and play fencing. So in both of those cases, rather than playing with your friend over a little headpiece and and looking at characters on the screen or sort of the guitar hero example, you're really there in the moment. You are present.

33:05 You can see your your friend in front of you, and you're kind of fully immersed and fully engaged. Over to the classroom as well, that same that same seminar, that same immersion helps you to learn better. So here is an example from Lufthansa, flagship airline carrier for for for for Germany. And Lufthansa put all of their air cabin crew through their safety training in virtual reality. So they've trained over 20,000 air cabin crew, and that means that the air cabin crew can be fully present in all of the different types of aircraft that they might have to fly. They can get on their hands and knees and check under the seat. They can put the seat tray tables up and back. They can go and do all of the activities and tasks that they need to do in the in the kitchen without having to physically put them on real planes that you would need to go and get, and you need to move the people and the planes to the same place.

34:02 But also, it allows a fusion of instruction and doing, a sort of different learning styles where some suit some suit certain types of people, others suit other types of people that you can go through your training, you can have the pop ups or a sort of a a bot and avatar instructing you, teaching you what to do step by step, physically placed in the environment where you can then actually physically do the activities as you're learning the steps rather than having the traditional approach that that I think most of us are used to have read about it in a textbook, and then go do the experiments separately. You can fuse those to get the two things together, creates a much stronger learning effect. So great for enterprise today, and you can see how that's gonna be great for the classrooms of tomorrow.

34:51 And then finally, in these examples, the workplace as well. There's a lot of examples for sitting in an office with giant monitors and huge spreadsheets, which are very, very exciting. But to try and take it into into something more practical, here's two two great examples that I love. On the left, a piece of software called r three d t, which is for designing industrial setups. So here, they've they're creating a new assembly line, and the person has been able to set up that assembly line, pick the the furniture, the the the machines, the equipment that they will need, the storage. And then rather than just looking at them in a catalog and measuring and drawing on a piece of paper, they can actually stand in that environment, make sure that the workers will be able to reach all of the different bits that they need to reach before clicking buy and then buying all of the stuff and and setting it up in their in their new facility.

35:46 And on the right, Autodesk v red software for enabling you to physically see a car in front of you during the design process where designers will use it to pick the colors and the trims and the the optional extras in the package. But then you can take across into the into the showroom. So when you're out there buying a new Porsche, you can stand in front of your car, customize it exactly how you like, look at it from all angles, and then it'll turn up on your driveway in a in a few days time. So some great applications. So that is what the metaverse is and where it is, now, how it is. And, okay, I prioritize triplicates over grammar here. But the reason I wanted to look at this section, which focuses on some of the technical challenges, to give you some ideas of where are the difficult problems that still haven't been solved yet and that might lead to to some great investments in the in the metaverse space.

36:41 So first, we have the basics. On the left, the actual headsets or the devices that you're gonna you're gonna put on your head, headsets, glasses, contact lenses, whatever they are, weight reduction is one of the number one requirements at the moment. Everybody wants these to be lighter. Anybody who comes up with any technology that goes in a headset that reduces the the mass of the headset, reduces the weight of the headset is hugely interesting. We also need an increased field of view. Some of the best headsets now have a really good field of view, but it's about a 100 degrees. And in real life, your real vision is more like about 200 degrees. So we still got a long way to go in being able to match the the the full visual field that you have in in the real world in these in these virtual settings.

37:27 And then lastly, at the bottom there, broader sensing, Not just being able to sense the room in front of you that you would be able to see with your eyes, but being able to sense where your legs are and how they're moving so your avatar can move in the same kind of way, and you're able to sense your face so you can mimic and and recreate the same facial expressions, and being able to to to sense your limbs when they go into all sorts of places, like behind your head if you're throwing a ball, all these other places that might be further away from the headset. So how you can create that full picture of your environment and what the what the user is doing from a wearable device is is is hugely important. On the right, the same typical tech stack that you need for most computing platforms applies to the metaverse. You need great content.

38:14 The best pieces of content are going to really change the game for for the whole ecosystem. You're going to need to places to buy that content from. There's a big push in the metaverse to have it open and interoperable, but you're still there are still gonna be platforms where you're going to go and purchase items or or software to use in the metaverse. And then infrastructure as well. Rather than having eight, ten, 20 people playing a game at the same at the same time, if you're in a large virtual world, you may have tens of thousands of people. Or if you have this fusion of the virtual and the physical in the real world, you may have hundreds of thousands of people all in the same space at the same time seeing the same virtual content. And that requires a a big advancement to infrastructure infrastructure and servers and how we actually deliver that content in real time to users.

39:02 So that's kind of the basics. And then I wanted to look at a couple of really acute technical challenges. And the first is spatial audio. Spatial audio is the accurate reproduction of sound in three d space. And one of the challenges with spatial audio is the shape and form of your ear affects the reproduction of the sound, affects what it sounds like to you. So if you and I somehow both stood in exactly the same position and a sound occurred over here, it would sound slightly different to each of us because of the the characteristics of our ears. And that's okay when you have Dolby surround sound because everybody hears Dolby surround sound with their own ears. When you start moving to in ear earphones that kind of push the sound into your ear canal and bypass the ear, well, that changes the changes the game. And because each person's ear is unique, we need to find a way of recreating, recapturing this sort of unique effects to each person so that you can get that really hyper realistic recreation of sound around you in your environment, and people can stop realizing that there's some sort of awkward virtual sound over here that doesn't quite sound right. Doesn't quite sound right.

40:19 That's one of the senses that's super difficult. The other one is for the sense of touch, which the feedback is called haptics, which brings me to my my favorite scene. Haptic is obviously very dear to my heart from from. My favorite scene from the Ready Player One movie, Steven Spielberg's story about a vision of the metaverse, which is here when Artemis asks, what kind of haptics are you rocking? Which is obviously a question that everybody's gonna be asking in the future. But really, haptics feedback for the sense of touch is essential for efficiently exploring what you're doing in a virtual world. It's essential for exploring the real physical world. If you go to pick up an object, you'll do 90 of that with sight. You can see your hand. You can see the object.

41:10 When your hand gets really close to the object, so you're you're limited by perception and how far apart the eye your eyes are in the front of your head. You can't tell with your eyesight alone whether your hand is just in front of the object or it's actually touching the object or you're kind of pushing into the object. They all kind of look the same. So that last 10%, we sort of switch off vision and do the rest on touch. Have I contacted it? Have I got a good enough grip to be able to pick it up? Will it come with me if I move my hand? And that's really important. You tend to reach and then your site moves on to the next task. And we're gonna need to be able to do that in the virtual world, but we're not gonna be able to perfectly recreate the feeling of everything in the real world. For many reasons, haptics is phenomenally complicated, and there's no line of sight to anybody being able to do perfect recreation of everything in the sense of touch in the next sort of twenty years or so. So we need a new language of touch to come with it, and then we need to be able to provide this with minimal friction.

42:11 People aren't going to be okay with wearing massive gloves as they go about their their their tasks in the metaverse. And then with these technologies and with this advancement, we need to think about how we measure progress. This is a big challenge for the industry because traditionally, really, we've looked at pixels per inch for a display, the flops of compute on a processor, how fast the visuals are refreshing. And there's a general general gauge for current generation technology that if you make those numbers higher, the experience for the user will be better. But broadly, those will help in the metaverse, but doesn't necessarily move us in the right direction because the goal is not to have a wider, bigger, snazzier screen. The goal is for the human to feel present in that experience.

42:58 And so we need to move away from these sort of highly accurate or specific metrics and more to how we do this kind of qualitative measurements. I think the tech industry is generally not super good at of is this just getting better for the sake of tech, or is it moving as to towards the actual goal of the of the metaverse? And then finally, if you want to invest or you're looking to help companies that you you think can have the the biggest impact on the future of humanity through the through the metaverse, then really, would point to ethics as being one of the the the key areas. I've mentioned the the the phenomenal amount of sensing and data these devices are gonna need to capture in order to create these experiences.

43:44 Well, we need to figure out how we're gonna do that and maintain privacy for the user so they're not just beaming up every unique detail about themselves to large corporations. They're also gonna provide these huge huge advantages in the workplace and education, and we need to make sure that there's some form of equal access to this, and we don't increase the difference between the haves and the have nots. And then possibly most acutely as well, if you're gonna release everybody into a virtual environment where they're not just on the other end of a keyboard, but they appear to be physically in front of you, they appear you you appear to be present in that scenario, then we need to look at abusive behavior and Internet trolls. And some of the some of the efforts in this area are what's going to enable us to make a better job of the next jump to the next generation of technology than we have to the current generation with social media and and the like.

44:35 So, hopefully, that's a super fast one zero one of the metaverse. That's fantastic. Thanks, Tom. Really appreciate that real tour de force of the metaverse. I think I understand that now a little bit better of the wheres and the whats and the whys, as well as some of the opportunities. That was fantastic. Thank you. And thank you again to Nikki and to Mark. So we're gonna move on to a question and answer session. There's quite a few questions already come in. If anybody would like to answer that list, you just need to click on the q and a, the right hand side of your screen, type your question. I will scroll through scroll through them and, answer them, and ask the questions to the audience, and everybody can hopefully hear the answer. The first one, I think, is perhaps more towards you, Nikki. It's a question about how secure will autonomous processes be from hacking. Is hacking a concern if something's autonomous, especially if it's mobile? Is is security built in? Do you have any views on that?

45:32 Very good question. Hacking is always a concern, and security is always a concern, especially for autonomous systems, for sure. But I think that there are a lot of different approaches to sort of increasing the level of security for autonomous systems. So one approach is to basically use machine learning itself to basically model the behavior, so of the true owner of the autonomous system. So if, say, someone else tries to interact with it based on biometric data or based on other data of the user, you can sort of try to to infer whether it is actually the the rightful user or not. You can also use sort of redundancy of sensor modality. So for instance, if you've got show in your autonomous system, like radar, lidar, cameras, and they tell you different things, you know that someone has really played with the input data, right, and has sort of changed and corrupted the input data.

46:34 It's kind of out of distribution. It's not what is expected to be seen in a certain scene. You could also look at sort of security through connectivity. So for example, you've got cars. Right? You could one car thinks that it's only, you know, a few meters away from another car. But if it is connected with a high bandwidth network, you know, these cars could share this information and and any kind of disagreement because of sort of hacks could be sort of identified. So there are I mean, it's not just only standard security and encryption and authentication. There are many mechanisms, and machine learning itself can be, like, a big tool to sort of enhancing security. So that's quite interesting. So there's some there's some of the classic methods to protect against it, but there's also there'll actually be new methodologies to actually circumvent the hacking and stop it happening and and the consequences of it.

47:30 And that's an interesting point. One for you, Tom. There's a question around how much, the metaverse development will will be dependent on The UK's broadband infrastructure. Obviously, that that's debated a fair bit in politics, of course. Do you have any comments on that? I mean, that limiting? Is that bottlenecking? Is there a dependency even? Yeah. Absolutely. It's it's it's super important both in terms of broadband to home, but also, you know, phone mask signal. Like, five five g is kind of a is a key component for the metaverse, and it's it's not really necessarily super focused on bandwidth. You know, today, for Netflix, for example, bandwidth just defines what resolution your video plays in and how many people in your house can stream a video simultaneously. The more important factor, in my opinion, is is latency.

48:17 Because if you've got these kind of activities happening with other people and other parts of the world, the I talked about haptics earlier. The the time that you have from doing an action to feeling the response to that action, do not to notice the latency is twenty milliseconds, around twenty milliseconds. So, yeah, latency is one of the the the real key bits, and that's gonna require future generations of of programs to and five g to really realize the potential there. Yeah. So it's not just about data rates and and ability to transmit information. It's about the quality and the speed of it as as much as the the bandwidth. That's interesting. I have a good question here that's come in about interdisciplinary teams. So people there's a comment. I mean, I think we know this. There's different different specialists, different teams of people working together often produce quite innovative solutions to problems.

49:13 I mean, to both of you, maybe to to to you first, Nick, know, given your experience at the venue and and elsewhere, have you seen or are you seeing or or you're planning to work in areas which are more multidisciplinary to look to solve particular problems and challenges or create new inventions, if that's not too broad a question? I mean, cyber physical systems, which is our research area, is multidisciplinary to begin with. We have in in the sense that it combines sort of sensor technologies, signal processing, machine learning, robotics control. It has multiple components. But beyond that, we've worked with other areas quite a lot. For example, we have worked with zoologists in the past to track badgers underground and came up with sort of new technologies that work underground.

50:06 As you know, Wi Fi, Bluetooth, and so on wouldn't work and use sort of low frequency magnetic fields there. Or we are having a collaboration with people from psychiatrists to see how mobile wear like, wearable smartwatches can be we can collect data from people that have early sort of signs of dementia and see whether we can sort of identify interesting patterns, mobility patterns from from inertial data on the smart watches. So there there are a lot of those kind of cases, health care obviously and No. That's that's interesting. How how about you, Tom? Sort of the the cross disciplinary approach, how how big is that for you? I think it's entirely essential.

50:55 I I I I sort of think that you think what is innovation? And, really, innovation is just taking two things that people didn't realize were connected and connecting them or finding out why they can be connected. And so all of the best inventions, all of the best eureka moments are usually sort of the joining of two different fields or two different disciplines. Mhmm. Mhmm. And, you know, at UltraLink, we have an insane number of different disciplines. We very rarely hire the same job spec twice across computer vision, machine learning, ultrasonics, acoustics. We do the sort of real deep tech hardware nitty gritty stuff right through to we are ultimately the bit that comes in contact with the humans, so we have to have the sort of psychophysics studies and the creative design. And it's only once you put all of those bits together that you can really see what you have and where you need to go as well. So I think that's why you're trying to know the best companies have that deep deep tech stack and that ability to understand top to bottom what they're doing.

51:49 Yeah. Okay. That's interesting. Here's another one for you, Tom. Actually, it's quite a good one. Metaverse, will there be one metaverse or many metaverses, if that's the right plural? How will how will that evolve, do think? Will it be corporate metaverses, individual metaverses? Or Yeah. So it's a it's a good question. I I think there's there's there's two elements to that question. There's the sort of the terminology aspect and then what will actually happen. So terminology wise, I think there is only one metaverse. So the the whole concept of the metaverse that we have the real universe, and then the metaverse, the sort of virtual universe that is sort of attached and sort of intertwined with the real universe. So there's only one metaverse and just different places you can go to on that. How that actually maps out? I and then we've got a everybody in the industry, we've got a push and fight for similar with how the Internet was created, that there are open standards and everything is interoperable, and you don't have to pay to go to one website and not another, and your Internet service provider doesn't tax you for going to Netflix and allow you to go to Spotify for free and all of this kind of stuff.

52:53 But let's sort of also understand that we're in a world with lots of very large corporate entities who will no doubt try as hard as they can, same on the Internet, to try and lock you into their Yeah. Their garden and keep you in their their bit of the metaverse, if you like. That's interesting. Fascinating. So, Nikki, another question for you. Robotaxis. As somebody you're you're an expert on NAA. So how far away is it really before we start using robotic taxis? You know, we can just hop in, tell them where we're going, and they take us, you know, take us there. Is that really two years away, five years, ten years? Is it still thirty years away? What's your views as an expert in AI? Very, very good question. I mean, I want to say a few years, but we've also said a few years and, you know, several years back.

53:45 So, I mean, I'd like to remain sort of optimistic. I I think it's a matter of sort of less than a decade. So so, definitely, we we already see quite a lot of sort of autonomy in vehicles already. I can sort of three three to five years might be, you know, slightly optimistic, but feasible, I think, achievement. Fair enough. What do have a view of that, Tom? I mean, you you're not quite an AI, but, yeah, I I guess, ultra haptics in in you know, you're you're part of this world as well to a degree. So what what's your view? It's it's it's complicated. It's another one of these sort of yeah. Bet bet your money on the future. I don't I don't envy Nikki.

54:33 I mean, I'm trying to a pin on that because I think there's that there are still some technical problems to be overcome in order to get to get the side of it. Right? So you have if if you knew what the solution was, you'd be able to put a timeline on it. It's how long is it gonna take to solve those problems with unknown answers, and then you'll be able to see the other side and how many problems exist behind that. Right? That's very true. I'm I'm gonna ask you, Mark, that same question. I mean, we've got investments. There's obviously, like, the group in this space. Well, that's right. Yeah. We're a a shareholder at Oxford University spin out that is commercializing technology to to build autonomy into to vehicles, and they're world leaders in that that domain. And they would very readily tell you that we're still a some distance off a sort of climb into a car, expect it to take you anywhere in the world or any on the road in the world. And the problem is, of course, we can solve sort of 98% of the challenges that autonomous vehicles face. We can't solve the last 2% very easily like that. One they particularly give an example of is a plastic bag blowing in front of the car, a human would immediately recognize what that was and that it wasn't gonna damage the car and you shouldn't slam your brakes on. A computer would have a great deal of difficulty interpreting this sort of unpredictable movement of the bag and and interpreting what's in front of them. And so the edge cases like that are very hard to solve.

55:50 So I don't think we'll have some absolutely pure complete takeover cars with no windscreen in a short space of time. What I think we will see a lot more of is semi kind of autonomous vehicles in fairly controlled settings. You already see this at Heathrow where they've got a kind of track and an autonomous vehicle taking from terminal to car park. We'll see that type of thing where city centers also or roads are designed in such a way that there's kind of a track that autonomous vehicle is not gonna come up up against too much that's unpredictable, and and it'll be used in more settings like that. No. It's a it's a good good response. A few more questions. One each, I think, actually. Somebody's asking about impact studies. More more for you, Nikki. Do people before they release, you know, new AI algorithms, what what happens? Do is there some impact case study about those? Do people say, well, this might have an impact in these areas that are sensitive or difficult?

56:41 Do people look at that still currently, and and do you think they need to look at that in the future? Do do people need to know what control them before these algorithms are released now and and how they might evolve in into the future? Impact studies are super important, but I don't think that they happen at a specific point in time. I think they happen continuously. Right? They happen by people who actually do the research and produce the systems. Then they happen afterwards when sort of the system is commercialized. You may have sort of government sort of approaches to trying to take and scale a technology. For example, for us, NHSX has funded scaling the technology to several hospitals and seeing whether we can get when whether we can check how well the technology works in a in a certain amount of places.

57:33 And and then you actually you really need to put it in practice to see how things how a system really works. Right? So I I don't think you can have a foolproof case where you can say, I I do a very content very comprehensive sort of study, even independent study, and I give it a a tick and off it goes. You continuously discover when something is used in practice, edge cases, and things where it goes wrong. And Yeah. I think where machine learning can help there is that instead of seeing it as approved or not approved, you you can imagine sort of algorithms that can take input from the users continually and improving sort of the systems rather than having having them pretrained and sort of fixed with fixed behavior. So I think we need to move towards that direction rather than sort of have very, very strict sort of points where we let things in the wild or not. It's interesting the commonality between answers there of use case of edge cases rather.

58:37 It seems to have impacted. It's an obvious one perhaps, but but when you think about it, it looks quite interesting. It's a commonality between them all. Also, I have one last question, which I prepared a little bit in advance, and Mark already has the question so he can answer a bit. What which but I quite like this one. What technology would you like to see commercialized or available in everyday use by 2030? I'll ask Mark that one first. But if there was one thing you would you say, actually, you know about now, actually, I would like to see that out there. Not not from your own companies, but some something else. What would you like to see? Well, you rather narrowed it down by saying not from my own companies running the portfolio of 40 or so companies in different areas. I mean, I'm a I'm a geek. I'm in a hurry for all of these these technologies to be surrounding us. I'm I'm ready for service robots that are much more advanced. I want to drive a less car before my son turns 16. That's for sure.

59:26 Flying taxis and things will be great. But I think, you know, every day, I think of a use case for very high power and and and and lightweight augmented reality spectacles. I think that will we we haven't thought of all the applications of those so far. And I think, you know, being able to shop for a car where I can walk around it and explore it and visualize a new extension to my house, and I can just every day, I can think of somewhere where I would use that that technology. And so I think pervasive communications that are kind of seamless and and high bandwidth, I think that opens up a lot of new human capability. It it it creates a lot more fairness across the world. I think, you know, the technologies that enable that would be be good. No. That's good. Tom, if it's you next talk. Yeah. It's a really good question because I can't say I can't say what I do.

1:00:16 So if you'd have asked me this question six months anyway, there's a little bit of a cop out. Maybe if ask me six months ago, would have probably said the sort of generative AI that Nikki was talking about. Mhmm. You give it an input, and it creates an output because the some of what that gives you, I have very low artistic skills. I would love to be able when you're creating presentations or when you're or even just for for sort of for pleasure to be able to have that feeling of being able to create something very artistic where it's my input, and I'm sort of tuning it and tweaking it, but I don't have the brush skill or Okay. The art if I make that happen, and you can have an AI that can do that for you. And because of it, I think six months ago, maybe not let's say for Nikki, but certainly to me, that seemed a very long way in the future, and everybody was talking about AI is going to replace the functional jobs and the creatives of the the last bastion of humanity.

1:01:10 And that's flipped super fast, and now it's like a new a new advancement every single day by the scenes of it by the sound No. It's an interesting point. I not thought of that, but, yeah, creativity. I mean, I guess, can machines really be creative? That'd be a good question. Maybe not for now. But Yeah. But, Nick, what what do you think? What what would you like to see in existence in 2030 technology wise? I wouldn't like to narrow it to specific sectors or even technologies, but I would like to say that I'd love AI to come to a point where it can learn from humans and take input and and convert with humans, not in this sense of, you know, language asking questions and answering answering them, but in in the sense of explaining what it does to humans and taking hints on how it can do it better.

1:02:04 So and this happening sort of in real time and to us on the job. I'd like AI to come to that that point. That's interesting. It sounds like there might be some overlap there with augmented reality as well, of course, as a display mechanism and then an AI to to answer the questions. That's great. I think we'll leave it there. Again, thank you both well, all to all the speakers, but, particularly, Nikki, I, you know, I really enjoyed hearing about some of the points you made specifically around, you know, connecting language, envision with AI. I hadn't really thought of those kind of concepts before in in mobile autonomy and how important and pervasive that is, and then opportunities around new sensing. I thought that was that was really interesting and insightful. And then thanks to Tom. You've opened the lid quite a bit there for me, and I'm sure lots of others around the metaverse as to the the what's and the whys and the who's and the and and such.

1:02:56 And then, of course, opened up the different use cases and the activities that can be performed within it and back to the opportunities that arise from that. So so all really exciting. So thanks again for the speakers. Thanks, Mark, as well for joining the the panel discussion and for the the introduction at the beginning. I'll just then say thank you all for attending and for participating. Any questions or follow-up, there's an email in the invite, so please just just follow that email link. And I'll hand it back now to Mark who runs the investor meet platform. Thank you. Great. Thank you, Lee, and thank you to the rest of the presenters this afternoon for updating investors. Can I please ask investors on the call not to close this session as we're now automatically redirect you for the opportunity to provide your feedback, and all the management team can really better understand your views and expectations? This will only take a few moments to complete, but I'm sure it'll be greatly valued by IP GROUP. On behalf of management team of IP GROUP, we'd like to thank you very much indeed for attending today's presentation, and may wish you all a very pleasant evening.