Welcome to episode 67 of Lost in Immersion, your weekly 45-minute strip about innovation. As VR and AR veterans, we will discuss the latest news of the immersive industry. And Seb is back among us. Welcome back. And no, we will still start with Fabien. Hello, thanks. Okay, so I have two quick topics today. The first one is about Magic Leap. And so it's unfortunate news for the ones that have been laid off, but it seems like there was a strategic shift in Magic Leap. Seems like they have a lot of, you know, we know the history of Magic Leap is a bumpy one. And so again, it's a new step in their story. And the announcement that you can see here in the article. So it's a bit, it's a lot of, you know, nice, embellished words that describe usual strategic change and layoffs. But what I understand from that, and I'm actually curious to know what you guys think, is maybe it's like a hint as they are done or they will be done soon, because they say that they will focus on existing customers and their ecosystem. So like developer support teams. So I don't know, maybe I'm a bit overly pessimistic, but it doesn't look good for Magic Leap. So yeah, maybe let's start with that and see what you guys think. And let's start with you Seb, what do you think? Yeah, I wonder what will be their partnership with Google mainly. Will there be focus in helping them building application or from what they learn on how to build Excel application? I don't know. But yeah, or their software, the tools that they build on top of Android to make the application. We'll see. Yeah. I think Magic Leap is really the example of a company that made a lot of buy choices and mistakes along their history. If it's the end for them, we know that they've been living on IP for the past few months, maybe years. They are just selling some IP to other company and Google especially. But when you are looking back at when it started, their first Magic Leap product was for me better than the HoloLens. And because they just focus on artistic and entertainment, they clearly didn't make the best choice. Because we know that this is not a market where there is lots of money. And obviously, they found that it was not where they should have been doing business with. And it took them several years, I guess three or four years to understand that. And they shift to the industrial field just to be copying HoloLens. I guess they were trying to get one part of the market that the HoloLens was not working with. And this was their main mistake because at this time, they had the best device, especially in the field of view and economics as well. And then they just keep on going and keep on doing those strategic mistakes. And now for a company that could have been one of the leaders in AR, they are just like... I'm like you, I think they will be closing their doors in a few months or just being bought by Google. And if you can remember well, they had a whole AI branch back in the day. I guess it's 2020, 2021 when they had this artificial intelligence avatar or persona that was able to interact with you. And they really work on the presence of an artificial intelligence agent. And from one day to another, all those work just disappear. It's really, really hard to find their work back. I guess it was Leia, it was a woman that was in an AR environment and you were able to discuss with her and see that she was adapting to what was going on in the environment. Either it was staged or they were very in advance in AI, but all these completely disappeared. So I guess it was more staged at this time. Yeah, it's a pity to see this kind of company just disappear. Oh yeah, there's the whale effect as well. Very, very bad marketing presentation to oversell clearly what the device is able to do. And yeah, so lots of mistakes and bad choices. Yeah, I completely agree on the strategic error on placing that for gaming and theme park and events, where it should have been right away a device for industry. And yeah, one of them, also one of the issues is the form factor. It's way more complex to wear it compared to the HoloLens 2. You can't wear glasses underneath and you still have this backpack that we know also for Vision Pro is kind of an issue. So, but the Magical Leap 2 is a great product when you have your prescription lenses inside of it. It was well and the way they did something that nobody does right now. They changed the opacity of your mixed reality object by changing the opacity of the glasses itself behind the 3D model that you display. Which I'm sure a lot more the 3D model in your environment. That's something nobody does yet. So they have no competition on that. And yeah, the PC that you carry is way more powerful than the one that you have on the HoloLens 2. So also you can render much more. So yeah, it's a shame that all goes to waste. They are not improving their design. But we'll see how Google reused what they did inside the new product. I'm also curious to know what you guys think of this trend. I don't know. Quest 3, Apple Vision Pro, Pico, all these devices and cameras to see the reality instead of using passthrough. And I don't know, are passthrough devices like the MagicLip and HoloLens not the good solution until some industry figures out the brightness and the field of view? Is it like a dead end for now? Or are there still experiences and maybe advantages to use these kind of devices? For me, it's more a question of cost. I think those devices, we start to see some serial, for example, glasses. That seems to be able to do quite a nice rendering inside the see-through glasses. But I think the cost right now is not an option. You will get an headset that will be like $10,000. So if you want to open a market, you also need to consider how much the device will cost. And so you make sure people will buy it. Otherwise, it doesn't make sense. I think right now, the camera and screen are evolving a lot and getting cheaper and cheaper. So it makes sense to make this kind of headset until, like you said, the see-through devices are cheap enough to be produced in these kind of devices. Yeah, Apple has been very, very clear with this, meaning that they announced that the video pass-through is not the goal of their device. It's just a temporary technology for them to be able to master and bring up some concepts like spatial computing. And yeah, I don't think that the video pass-through is what people or the public in general are waiting for. They are all asking for those pair of small glasses that can be able to create this AR experience. What I'm a bit worried is that the see-through technology is completely disappearing. Microsoft told us that there won't be an HoloLens 3 for now, and maybe the 2 will be stopped during the year or at the beginning of 2025. So it will mean that maybe, you know, we all know that Meta is preparing a new pair of glasses that would be different from the Meta 3. Maybe they'll be doing some see-through, but I'm not sure about this. And yeah, it will be meaning that the see-through technology will be completely off or it will completely disappear in 2025. And what is maybe the scariest on this is that it should be the technology towards where companies should be going. But it's been like nearly 20 years that we've been working on this, and we still have the same issue, meaning the field of view and the brightness of the augmented reality objects. So the question is, as you mentioned, Fabien, are we in a dead end because we can't go over those issues because we don't know the technology yet? I'm very, very curious. We know that the see-through has a lot of issues, especially for optical one, focus, and so on. So maybe they are just taking the easy road, as you mentioned, Seb and Fabien. It's the cheapest one. But I'm not sure that public will be catching the ball of using mixed reality through a video pass-through. So we'll see. Okay. Well, speaking of devices, I saw this news a couple of days ago about a new headset. The informations are very, very sparse. So it's only like a copy paste of this on like all social networks. And that's it. So I don't know how real is it. But what was interesting is kind of this form factor that they have chosen is similar to, you know, what basically like deported the processing unit in like a backpack or a front pack. Here they have this device that stays on the shoulders. So it's a nice idea. I don't know how practical it is, you know, how stable it is with movements and stuff like that. And how real it is. But I just found that piece of news interesting. So, yeah. Seb, what do you think of this like form factor here? That's definitely interesting. We all have shoulders and neck. So that kind of makes sense. For a large public, I wonder if the goal is to wear that a long time. It will depend on the weight and how it feels to wear this. I don't know if it's adjustable also. Because depending on where your bones are structured, you may want to readjust it. But apart from that, I'm wondering what is inside the glasses. I see two cameras on the side. Do they have special tracking? I don't know. But I guess there is no video yet of what you see inside and what kind of experiments you can do with it. Yeah, it's just very, very small information here. Even the OS? Is it Qualcomm behind that and Android? I don't know. I didn't find any information. So we saw the same info, like you said. Everything is the same on each social platform. Is it a Chinese company? I think so, yeah. They talk about Shenzhen, so I would guess it's a Chinese company. Yeah, because it's been a lot of headsets mixed with IT and so on that are just popping around these past few months. We just have the information, like the launching or just preview of what the headsets are doing and what it looks like. But after that, we don't really have reviews or we don't really know what those companies or headsets are becoming. It's very strange. Yeah, popping those prototypes. It looks like a finished product. It's very interesting. Or maybe it's just for the Chinese market because they are doing lots of those, especially in artificial intelligence. You know that there is this video generator that is only available for Chinese residents. Maybe they are doing the same with some technology. I don't know. Yeah, so we will keep an eye on it and see if we see some more details or test videos and stuff like that. But yeah, we'll see. What seems weird is that if the position of the camera is here and they are calling it an MR headset, I wonder what kind of... I don't think it's see-through, it's pass-through with screen inside the headset. I think so, yeah. They are talking about OLED, I think, screen. So yeah, the deformation that you must get with this kind of position of camera makes me wonder what you see in mixed reality. Okay, well, we'll see. Stay tuned. Okay, well, Seb, over to you. Yes, so a couple of weeks ago, we talked about the Vision Pro having some Vision Pro 2 OS allowing to do some 3D tracking of objects in your space. And we show that to the example, which seemed fake at the time on the Apple video, but someone did some tests on their side and that's what I wanted to show. So here you can see a couple of objects being tracked in a kitchen and see what kind of interaction you can do with it. So here the guy is showing that when he's touching the object, the headset is able to recognize that and change the color or change the behavior depending on how the user touches it with one hand or two hands. Here he's just changing the color of the wireframe. Here he's showing how several boxes can be, several packaging could be tracked with a visual effect on top of it. And also interaction with his hand. When his hand is getting closer, the shadow is changing. Then a bottle of milk inside the fridge is also tracked. And also a visual effect applied on top of it. What amazed me is the accuracy of the tracking. It seems to be a bit, there is a bit of delay when you move it, but when you stay still for seconds, it seems to be really, really accurate. So yeah, I don't know if you saw that, what you think of that, maybe Guillaume? Yeah, I find it very funny. I guess it was maybe one of the first use case, but the most common or the one that people are thinking at first is the, how can I cook with AR? So I don't think we need AR for preparing a bowl of cereal, but yeah, we get the idea. I'm very surprised by the overlay of the blue bowl. It doesn't seem to match nicely, but, and it's kind of shocked me because it's kind of, yeah. Remember, you don't see it in 3D. What we see here, it's only a video, but when you see it in the headset, you see only one eyes. So when you see the boss, I think it makes sense. Okay. Because those kinds of, yeah, those kinds of render, we are now used to get the perfect match, the perfect match between the object, the real object and the virtual one. So maybe Fabien could try this on his Apple Vision Pro and tell us what is the real deal here. But it seems to work on the cereal box. That's why I'm, yeah, we'll see. Even the occlusion of the milk bottle seems to be good, so yeah. Yeah. But yeah, it's clearly what we are waiting for. And we just have to see what is the capability of this algorithm to the number of objects you can remember and can track at the same time. I think this is, this would be the bottleneck at some point, because ideally we would have an infinite number of objects that could be interacted with. Because I don't know how many times it requires to get this 3D recognition, 3D object recognition, because back in the days you had to scan the object and it really depends on how you scan it. The result was very, very different from one scan to another. And it really depends on the object itself as well. It seems to be an object that you provide as an input, like a 3D object. But I need to dig more into the... So you would have to scan it at some point. Which makes sense. Yeah, you have to train the model actually, from what I see. Yeah, I'll dig into it and reply to that next week, I guess. Okay. Yeah, Fabien? Sorry. Yeah, there was not much to add to what you guys said. Yeah, I will try that on the Vision Pro as well, to see how accurate it is. Okay. Oh, and something that could be interesting as well is, as there are no controllers on the Vision Pro, it could be cool to have like, you know, you use something like a prop as a controller. I know I've just got the idea, so I don't know really how that would work, but that could be interesting usage if the tracking is really good. I saw someone tracking controllers directly and being able, depending on the orientation, to control something with it. So yeah, that's something to look at, definitely. But also for training complex tasks in the industry, that could be really helpful to get the user on each step and really see that the step has been done because you can catch where the hand of the user has been. Yeah, I see a lot of use case for that. All right. The next topic is this one. It's a paper that went out for SIGGRAPH 2024, which is in one week. Well, they're showcasing a way to capture, walk in the street, capture directly the street and remove, get an output where all the persons and all the plates are hidden. So basically copying the street view, but being able to really move from place to place, not from one picture to one picture, but really move in between and place yourself wherever you want. Without the need of having complex algorithms to remove all persons from the view you have captured. So yeah, Guillaume, if you have any comments on that? No, not much. Very interesting to see the radius of their rendering, because when we tried this, we were limited to the number of points and take. It seems to be improving a lot faster. I did some research about the geospatial application of Gaussian splatting, and it's still very, very small because they are more focusing on the NERF, which is older, of course, than the Gaussian splatting. But this kind of paper could accelerate the use of Gaussian splatting in the geospatial field, of course. Yeah, it's almost game-like, the rendering they get. It's amazing that you can really move wherever you want. They capture with a 360 camera, I guess? I guess, yeah. Very nice. It's very impressive that it's an empty city where you're working. I think we will see a lot more from SIGGRAPH this year. That's just the beginning. There should be some NVIDIA announcements. They are usually doing some workshops and stuff at SIGGRAPH, so we'll see. Especially in the meshing and 3D rendering. Fabien, do you have any comment on that? Not much. I think I say it every time, but it's really amazing to see how fast these technologies are progressing. If I'm correct, I think it's one of the team that published this paper is the team that created Gaussian splatting. I mean, initially published the first paper on Gaussian splatting, the INRIA team. I might be wrong, but I try to remember. It's nice to see that they are improving on it and optimizing. Very cool to see and curious to see where this will go in a couple of months. What is really impressive is that every time they are showcasing the previous version, which is like the six months ago version. The improvements, you can see it right now, are quite amazing in that short amount of time. The last subject was this one. It's a quick one. Meta changed the name of their mobile app to MetaHorizon. It will contain the same thing, but they said it will be easier to also share content and interact with the MetaHorizon platform with users that are using the headset. They seem to move forward with adding a lot more inside the mobile app so you can even interact with people that are wearing the headset at the same time. Which is something that was missing right now. It seems like a good move for them. I don't know what you think. Yeah, they kind of have the habit now of getting the name of one of their products and spread it around their ecosystem. So maybe Meta will be changing their name in Horizon in a few months. Yeah, they're trying to get something here with their Horizon OS, Horizon app, the Horizon Metaverse. I don't know exactly what is their vision now about all this. We all know that Max Gerberg is believing in this with all his heart. We'll see if he's true or not. Yeah, not much to say. Just this whole branding process for them to make things clearer for the public and try to get more users, I guess. Alright, Fabien? Yeah, so I don't have the updates yet, but if we look at the app, it really looks like a social network now. You have posts and sharing and likes just at the front page of the MetaQuest app. So I guess it will be pretty similar in the next update. So, I don't know. So, yeah, Horizon is like they're trying to do the VR social network, like to push in that direction. I don't know, we'll see. Right. And by the way, I tried the V68, version 68 of the Oculus on the Quest 3. And same as you, Fabien, my window, as soon as I removed it or even if I launched an app and I stopped the app, the window I placed in the space was completely messed. So, one thing I will try is to redo a scan, remove everything that the Quest 3 learned from MySpace and redo a scan and see if there has been any update that they forgot to mention that you need to re-scan your space to position stuff. That might be the thing so they learn now what is a world, what is a virtual world. I will let you know if it's working better or not next week. Yeah. And I even had a hard time getting back the app. Like, it was so far. So, I wasn't able to find it looking around and I wasn't finding it. So, yeah. Yeah, it doesn't seem usable as it is right now. I started it, I placed a couple of windows and then it was completely messed up. So, I stopped using it right away. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah, it doesn't seem usable as it is right now. I started it, I placed a couple of windows and then it was completely messed up. So, I stopped using it right away. It doesn't make sense. Yeah. All right. And to your subject now, Guillaume, I guess. Yes. So, it's a bit of mixing between those topics that you talked about. There is a kind of a battle going on between Google and Meta. As Meta is trying to acquire nearly 5% of the Essilor Luxottica company which is responsible of the Reban brand. So, Meta obviously is trying to secure their partnership. And yeah, it takes a bit of the company spending a few billions here. But at the same time, Google is trying to do another partnership with Reban especially to put their Gemini AI inside those glasses. So, something is going on here. I'm maybe surprised that Meta is trying to spend that much in the Reban brand as they announced their own pair of supposedly lightweight glasses for next year. So, probably it's not like we thought. Meaning that they would be doing a Reban++ with the Spatial and X-Reality and so on application in it. So, yeah. I don't know what to think anymore. Are they trying to do a Reban 3? That would be just an improvement of what they did in the past. Meaning that it would be smart glasses and not mixed reality ones. And they'll be doing like simply an Apple Vision Pro like for next year. Meaning it would be a video pass-through and just a cheaper version of the Apple Vision Pro, I guess. And in that way, I don't really understand where the Quest 3 would be positioning itself. It's like if they were doing the Quest troll, not so good idea that they did in the past. Trying to create a new product that is doing competition to their own products. So, they are dividing the market between their own products and then they are just realizing that the number of users is not increasing. They are just replacing the user from one hardware to another. And yeah, we'll see. And I hope because Gemini is really not the best AI on the market right now. So, I hope that the partnership is not going further with the Reban for their own sake. Yeah, that's pretty much it. But if this partnership was with, I don't know, Cloud or OpenAI, I would be more confident for them to get this kind of deal. But yeah, we'll see. What do you think about this? Fabien? Okay. Yeah, it's... As you mentioned, I think the interesting move is... Sorry, this move is interesting knowing that they are planning to release something later this year. So, as you said, is it a completely... Just an upgrade or a completely new product? And their partnership with Reban for, I don't know, like the form factor or maybe prescription glasses, I don't know. So, it's as usual. And as we saw also with Magic Leap and Google, it's a lot of like pieces moving on the chess of mixed reality. And all the giants are really completely involved in that. Now that Samsung is also in there. So, it will be an interesting end of the year, I think, for XR. And not to mention the Quest 3S, which we don't really know about it. We didn't hear about it since like the leak that was published a few months back. So, yeah. As usual, I think you said that often, Guillaume, the roadmap and the strategy of Meta is really unclear on that. So, that's not helpful for us. We don't really know what's coming. Yeah. Yeah, I agree. So, it's an impressive time where everyone seems to be in a war and trying to catch the latest technology and buy the latest technology from others. I think it's also amazing what we see about Samsung, Google and Microsoft all working together. I think that the first time three big companies like this try to compete against others. And about Meta, I think they announced the Quest 4 for 2026 and a new Quest Pro for 2027. Or the reverse, maybe. Quest 4 for 2027 and Quest Pro for 2026. So, they're still announcing a lot of things in the future, even though the market doesn't seem there. So, it's amazing the amount of money all of them are spending, even though there is not a lot of people buying to do that. So, I hope it will go forward and not crash down on themselves because they spend too much money on it. We'll see. Yeah, about this kind of adoption. I think it's been a few weeks now that the Apple Vision Pro is officially available in France. Because I've seen some reviews popping up of people trying on the Apple Vision Pro. And it seemed like old news, but they were like French. And apparently, there is not another wave of new users. I guess we predicted that, meaning that all the users that were interested in the Apple Vision Pro got it on day one with the US launch. They just got someone or a company for buying it in the US and shipping it in all the world. So, I'm not sure that this delayed launch is very strategic at some point. Because basically, it's not doing anything, despite the fact that people are trying this six months after, barely. And yeah, I'm questioning the strategy here. But it confirmed that when officially a new device is available in the country, the number of new users is not increasing. And it's worrying for the future. Meaning that we have a very strong basis of passionate people about VR and mixed reality. The larger public is not getting it. But yeah, will tell me that it's still the Apple Vision Pro with a $3,500 price tag. So, it's maybe not the best choice, best example here. I don't know the real price in Euro. Just looking at it. It's 4,000. Okay. So, it's more expensive because of, I guess, the US dollar and the Euro is the same right now in terms of change. So, they are just adding a 500 Euro for shipping. So, even there, it's better for you to get it in the US at day one, instead of waiting for it to be available in your country, which is very strange. If you want the cost to travel. Yeah, it was the same in Japan. There was a queue in front of the Apple Store on the day of the release, of course. But not that huge one. It was raining. I don't know if that kind of things. But yeah, it's also very expensive, especially because the yen is really low right now. It's been for years, but it's really low right now. So, I think it's 600,000 yen or something like that now. So, it's pretty expensive. Okay. So, do you have any more topics to share? No? Okay. So, that's it for today. See you next week for another episode of Lost in Immersion. See you guys. Thank you. Thanks. Bye. See you guys.
