Are all the live streams on? Are you sure? The X ones, too. Okay. All the streams. Well, it's nice to see you all. Thank you for being here. A lot of people wanted to be here tonight. We didn't have a lot of space. Even the majority of the employees are all watching from the office right now so we could make room for everyone. Yeah. Are we ready or do we wait a minute? What are we doing? What's the hands mean? Are you guys ready?
Now are all the camera and technical people ready? Okay, good. Cool. That's another thumb. All right. Gosh, that means it's time to start.
Scary. Anyway, well, thank you for being here. I'll probably, yeah, we're gonna probably start in a slightly unusual way because we're a slightly unusual company. A lot of people may or may not know that we actually, we have no investors. We're kind of a weird thing. We're kind of a community-supported research lab. We have a lot of people around the world who love using our stuff and there's enough of them that we just have a lot of freedom to build a lot of really great things. And of course we feel an obligation to build a lot of things.
But before we say what we're building, it probably helps to talk a little bit about, you know, maybe a little bit about David, a little bit about where I came from and where a lot of the employees came from. And we'll kind of give a little history, a little backstory. Because otherwise, if you know us for our cat pictures, you're about to be very surprised. But we make really good cat pictures.
Cool. All right. I don't know, I have two laptops. This is a bit cool. So yeah, we're Midjourney. I'm going to start off talking about me, which is weird. But yeah.
Yeah. I grew up in Florida by the beach. My parents worked in medicine. My dad had a dental office on a sailboat. So I grew up a lot around sailboats, a lot on the beach. There's me taking, you know, I was spent a lot of time sort of very in a somewhat solitary way like walking out in the night like this. There was a moonrise, sunrise. I'm just walking and thinking about the world and I'm seeing giant sea turtles come up and laying eggs and big beams of light from the lighthouse sweeping across the horizon. If you ever wonder why we have a sailboat logo now, you know.
And yeah, this is sort of, most of my friends were on the internet. So, you know, this is around the time where I kind of wanted to tell people what I was experiencing. Like if I say I was on the beach, I saw giant sea turtles and red lightning and no one was like, 'What is that?' you know, so I got cameras, I started taking photos, and yeah, I mean, then I started, you know, but then it's me and it's something about being out in the beauty of nature and trying to communicate this thing and then getting into technology to communicate it.
And then obviously I'm doing photos in Photoshop. I'm making space art. I'm really into space. And then because I'm into space and I'm into the universe and sort of how all this works, I'm trying to figure it out. You know, I go into physics. And it seems like the physics people kind of knew something. I was like, a lot of people seem to know things. A lot of people seem to not, but the physics always seem to know something and I want to be a little bit more like them.
And along the way of studying physics, I also studied a lot of math. And at some point I'm like, I might as well get a math degree while I'm at it, too. And I'm like, math is pretty cool. And somewhere along the way there is a sort of weird realization that in math and physics you kind of just want to be able to predict the world and you do that with math but the math is somewhat secondary to the predictions. You know you can say things like there are pi e to the six seconds in a year and I'm going to cancel out the pies in the equation with that which is not true but it's like totally okay if you're a physicist.
And then in math the math people scream when you do that. Like it's not pi e to the six seconds in a year like what are you doing and then a lot of times the physics will be like oh the maths guys will figure this out later clean it up make it nice. And so I felt good doing both but the more I got into it there's a sense that like with physics it was about making predictions even if the numbers were kind of fudgy and things were not totally clean and we don't really know what's real but we can make predictions and then on the math side it was like all that's the only thing we really know are the numbers and the sort of raw structures of axioms and the things that come out of them.
And there was a sense as I was going out into the world that I had to choose between these two things. It was either a choice between math and the sort of pursuit of absolute truth or like physics and the sort of pursuit of absolute reality. And I was like that seems like kind of a messed up choice. Like I just, this seems really hard, like no one told me about this. And so I kind of, you know, I did go to graduate school for a little bit, but I think I realized very quickly that I wasn't going to be able to do these things. I wanted to back away and just try to focus on people.
And so, this is when I started to work on my first company, which was Leap Motion. Some of you, I'm sure, remember this company, some of you don't, but we'll talk about it briefly.
So I'm maybe 22 or 21. I'm starting this company called Leap Motion. So, Leap Motion was about this idea that the limit in technology isn't the size or the cost or the speed, but often just how we interact with it. Doesn't matter how powerful computers are or how smart we are, but what we do kind of comes from the ways that we interact and that the interaction itself might be the biggest limitation in people and technology in the world.
And so, my first company was all around this sort of user interface problem. And so we built this thing called the Leap Motion Controller which was a little device and it tracked your fingers and it was pretty remarkable. We put it out there probably at 22 and we sold about $40 million in pre-orders in the first 48 hours. Which at the time was a pretty shocking thing. We didn't tell anyone, we probably should have said it but it was pretty shocking like that doesn't happen. I don't know if anybody knows, it still doesn't happen. I don't think anyone, that's happened. It wasn't a Kickstarter. We had a website.
And from that point on, things went a little wild in my whole life. Everybody had advice for me. Everybody wanted things to be different ways. But I just kind of had this problem that I wanted to solve and this thing I wanted to explore, which is sort of our connection to technology and how much greater we could be with the two together.
Yeah, over time we made a lot of hardware. So here I'm, you know, I'm working with cameras again just like I was as a kid. There's a lot of weird physics. We're doing weird LEDs and optics and it's kind of coming from what I know. But it's sort of focused on people. They made, there was a lot of work making it smaller. There are these little cameras made even smaller over time. We had like hand tracking running on like 30 mils off the DSP. And it was pretty cool.
I think one thing, something maybe more people now know of is we went into virtual reality because this all happened before Windows supported touchscreens. So this is a really long time ago and we were really, they say the future's not evenly distributed and I hope you'll see that a little bit right now. But we did a lot of VR and VR was just really cool. We kind of defined the concept of hand tracking in VR. For better or worse it's probably still better than any hand tracking you can get today. And yeah, it was just really cool.
And something that people didn't know, I mean, so I think people knew it was cool. It's still probably one of the coolest technology demos. But something that people didn't know was actually how it worked. It's been almost 10 years, so I can talk a little bit about that now. All the patents are published. All the stuff is out. So what was this?
So this system was actually, in modern terms for those of you who know AI, a 600 million parameter mixture of experts with a 5% sparsity. Which is wild for 2015. Absolutely out of its time. It was trained across a CPU cluster not a GPU cluster because we didn't do GPUs back then. It was all hand-rolled. TensorFlow didn't exist yet. And yeah, again, the future is not evenly distributed, but it did a lot of really cool stuff.
So, in fact, not only were we doing hardware and cameras and sensors and interfaces, we were doing a lot of really weird AI stuff. And I wish we, you know, we didn't really talk about it a lot because we were kind of honestly a little scared that someone might catch on to the whole neural network thing. But eventually it happened. And then, this is, we just made an AR headset too. We did open source AR headset. It was called Northstar. Thousands of people around the world in Japan made their own headsets. It was also really really cool.
So we're building all, and so you know, let's see, here's more crazy things. All the CPU, this is our infrastructure training the model, the neural network models with the data pipelines moving across all the servers. This is still a pretty weird image in 2026. Like I don't think any AI companies have weird flow graphs of their data pipelines like this still. This is still pretty weird.
But anyway, coming out of all this, I learned a lot and, man, there's a lot of things I could talk about, but, you know, little things from, for example, like, when you start a startup out here, there's a certain way of doing things and the general desire and the pressure and the structure, which isn't to say it's bad for all people. I would say it's generally pretty good for all people, but it involves this concept of having a plan that has a 10% chance of working with a 100x returns if it does.
And I think as a kid it's really cool, unless you take a huge swing you can do all this kind of, you have all this opportunity to do stuff. But for me as I kind of went through this process over years and years, this is over 10 years, I realized there was, didn't quite feel right for me because I didn't really want a company I just wanted a home. And I didn't really want to like make 100x returns. I wanted to solve problems that I cared about. That they were like big missions in the world and I wanted to like compound things over time and just see what I could do.
So after Leap Motion when I went to start Midjourney I wanted to do a lot of things differently. And the first thing I wanted to try to do was, well maybe I can do this thing without investors which was scary. I only had like 200K in my bank account at the time. I did not come out of Leap Motion particularly well. But I had a particular mentor, Bill Warner, who said, you know, David, you could do that. And I was like, Bill had this thing where he was one of my mentors and everything he ever told me ended up being true. And I usually didn't listen to him. And so this time I'm like, oh no, he's probably right. So I'm going to give it a shot.
And so, that was one thing that was different. Another thing that was different was although I still really am passionate about people and technology and the way they work together and I hope we see that a little bit in Midjourney today. There was another thing that kind of came out which was that I realized even if I gave everybody VR headsets and AR glasses and crazy hand tracking they weren't necessarily going to still think or be able to do a lot of the things I did or the ways that I do things would still somehow be inaccessible and something felt wrong there.
I felt like that there was, it was almost as if the very powers of imagination themselves needed to be stronger. There needed to be some infrastructure to be something that we had that was more than what we had to be able to make things and imagine what we wanted or what could be and something that would encourage us to believe that we could do those things.
So yeah, that was sort of the theme that I wanted to think about when I was starting Midjourney. So there's a lot of things we were playing with. And one of them that I was kind of, find is the year now, the year 2021. Is the year 2021. And one of the things that I was looking at was this new thing called diffusion models.
And so diffusion models were this sort of interesting thing. They were like a new kind of machine learning thing, but it wasn't like the normal input output. It was more like this idea that you could kind of hallucinate things out of noise and that the shapes or the sounds or the motions or just anything just sort of like condense out of the noise like water vapor into shapes of clouds and it was really beautiful.
I didn't really, so like it pulled me in and I started playing around with things. I'm just by myself in my apartment. And you know, one day I asked, I was playing around with some code and I asked it for a, what was it? An arcane lighthouse. You saw lighthouses. I grew up with lighthouses. And I got this picture. This was the original picture. And I looked at this and I was like, 'Oh, okay. This is interesting.' We got like a floating lighthouse with a lightning beam, you know, like light, and like auroras. A little, like maybe golden, like a little, it's a green sailboat.
Like this was more than I asked for and I was like okay actually for this brief moment even if it was just a mistake this thing was actually more imaginative than me and my original goal was like hey I just want people to be more like, they want people to do the things that I do and I'm like actually wait no wait wait wait like this could be better than me or maybe with me and this like we could be just even better like I don't know what this is this is like day one and I'm like okay okay this is pretty cool.
And then I mean this would happen all the time. There's another one where I asked for a picture of like a father and we got this like suit with like a child as a tie and I didn't ask for that. That's not in the data. Like there's now, if you really look into why, how these things work it kind of makes sense that you could get a picture like that. But again like very creative, very interesting, very unexpected. There was just something, there was some juice here.
Everything I made whether trying to ask for things that I hadn't seen before I was like this is really cool. Even like the sort of watching the noise sort of come on, come on again. Even just like watching the noise of the thing sort of bubble and churn and the images form, you could feel like the thing kind of like some kind of breathing aliveness of the thing. And it just felt really sort of profound and interesting. And I was like, 'All right, there's some magic here, and I want to get it to more people.'
And it's sort of, and maybe it can do some of the things that I care about doing. And so the big problem from here was it's now the year 2021 and I calculate out, you know, how many, like this is before ChatGPT is launched. This is really early in the field. And just trying to figure out okay like it takes like X minutes to make Y models and if you have this many people making this many images like what do you need? And it turned out to be like I need like 10,000 GPUs.
And so I called, luckily, you know, because I did this cool company called Leap Motion, I could call up friends at Google and be like, hey, it's David. Like, I'm doing a thing. Like, oh, David, he's doing a thing. Like, it's like a new thing. It's gonna be like cool, like the old thing, but like this time instead of needing cameras, I just need like computers. Like, we got computers. Like, no, but I need like a lot of computers, guys. Like, oh, I need like 10,000 computers on day one to launch, which is like at the time, this is again, this is before all the GPT stuff. This is like that's like a crazy ask and I'm like don't worry like I'm good for it like I'm going to find a way to pay for it but like I just need like, you just need to like give me 10,000 servers. And there was sort of this moment and they go, okay.
And I mean there's some questions like, you know, how does David bootstrap an AI lab, you know, part of it does come from when you're out here for a long time there's a certain degree of trust you can get and it's really wonderful that I could just ask for that and I had no investors and they gave me, you know, it's pretty cool.
And so yeah, so with those 10,000 computers, we built a little product and we launched it and it was the very first time that people could use diffusion models, any kind of image model at scale. You could run on your computer before that, but there was like maybe one or two places where like 100 people were using it, but this is the point where like, okay, millions of people can use it. There's like no limit. If you put your credit card in, we will buy servers and we will make this work.
And there was a certain, because we had no investors, there's a certain weirdness to this where it's like, you know, somebody has to pay for the computers and I've got like a couple hundred thousand in my bank account. So like, you know, I'm only going to last so long. So, you know, we immediately, we kind of had to build this thing as like a business from day one. Which is cool because it means like hopefully we're providing value to somebody. That's why they're paying us money. But it was kind of a weird thing.
And I didn't really care about money that much this time. Like I still, but the way I thought about it was like, well, something has to pay for the servers. And then in an ideal world, this, I would have some kind of business where like I could have enough money to do one Leap Motion worth of R&D on the side because like Leap Motion was this crazy thing I did for 10 years. Didn't cost that much. Was like 10 million a year, you know, roughly was what it cost to do all that. And I was like, well, that seems doable. Like it's a lot, but like I could probably swing it.
And then if I just had that protected thing with crazy R&D for just years and years and compounding and I could just do crazy things. I knew I could do crazy things because I had done it before. And the only difference was before I had to also do the 100x, like do it through, do a certain type of dance which again is neither good nor bad but in this case didn't always let me sort of just build the things I wanted to build or solve the problems I wanted to solve.
And so yeah, we launched this thing and it did good. So over the first nine months or so, we got about 20 million people coming in. This is 2021, 2022ish. So I mean the way I thought about it was I was like, okay, ideally we get like 10 million on the side for R&D budgets. But you know, I could mess this up. So I should probably have like a plan where it's better than that. And so I kind of designed the thing being like, you know, maybe if we get lucky, we'll do like a hundred million or something. That way if I really mess up, I'll still do 10. You know, that's the really thing I only care about.
But it unfortunately we didn't mess it up. And it kind of kept going and it kept going to the point where it was like 100 million in nine months and then 200 million in 12 months and it kept going even till today. And it was at the point where I'm like, oh, okay. We're going to do more than one Leap Motion's worth of R&D now. And in fact, I'm going to do the thing that nobody ever is allowed to do, which is more than one thing.
And it is hard. It's hard. It's actually really hard to do that. So we started kicking up a bunch of teams, hiring a bunch of people probably around 2023, 24, right after around this. And we've been working on a lot of projects. Eight projects. And four of them are hardware projects and four of them are software projects. Two of them are hardware projects that I hope you will be able to buy relatively soon and hold them in your hands and the other two are something else. I would say they're more like these sort of giant awesome machines. And I'm going to show you one of those today.
Yeah. But what about today? Like what are we? We're now like we've gone from like 2011 to like 2026. Like what do we even know about 2026? Like what can we even say about the future anymore? Like what do we even build anymore, right? And I think there's one thing we know more than anything, which is that just everything's going to change. So the big question, maybe the only question is like how do we want things to change and like what do we want to build and eventually like what do we want to become? It's an even bigger question. Yeah. So I think that's like the most important thing to all be asking ourselves at a time like this. And today is us giving you one answer to that question.
Yeah. So I think we might be ready for the first video.
There's more than one video.
What was that thing? So we are going to show off the first new whole body medical imaging modality in 50 years. We're calling it the full body ultrasonic CT or just the full body ultrasound. We kind of dreamed of something as powerful as an MRI and as casual as a trip to the spa. And so that's what we're going to kind of show you today.
And I'm sure you're kind of wondering how's that even work and why hasn't anyone done this before? And so I think in order to answer that, we're going to have to do another video, which is a bit of a technical deep dive. And that is the next guy. So, we're going to start with something called the transducer.
When you put energy into it, it vibrates like a speaker. And when it hears a sound, it makes energy like a microphone. These transducers thus act as both speakers and microphones. We have 8,960 transducers arranged in a grid. Each transducer is 200 microns wide, like a fine grain of sand. They vibrate over 50 nanometers, 10 times smaller than the wavelength of visible light.
Imagine a choir of 8,000 singers and an audience of 8,000 people, all listening carefully. The transducers are controllable at a rate of 100 million times per second. Together, they fire in patterns that send out structured waves. The waves travel across the water at over 1,481 meters per second. We construct a ring of 40 of these systems, 70 cm in diameter. The ring contains 358,000 ultrasonic elements.
The chips take turns sending out waves. As the waves have a chance to dissipate, we fire the next one. One by one, they fire at a rate of over 1,000 times per second.
The waves travel across the tank in 480 microseconds, about 1/12,000th of a second. Hundreds of thousands of transducers listen. Each sensor resolves motion smaller than the width of an atom. Not micrometers or nanometers, but pometers. Working together, hundreds of sensors can even push into the subatomic femtometer range, finer than the scale of atoms. A scale I don't even have anything you've heard of. I'm sorry, guys.
The system captures data at a rate of a leisurely 17 gigabytes per second. Gigabytes of sonic reverberations flow around your body and from the vibrations we will form images. Each probe sees different angles of your body. We combine thousands of waves and thousands of sub-images. Over 40 gigabytes of data moves through the system to see just one slice of your body.
We analyze these images. You can see them little shadows of the organs and everything to make out organs, structures, and tissues. Here we'll show you a slice of 25 different biological structures. This is a real image.
We do this again and again as you move through the ring. The lift will lower you at a gentle rate of 4 cm a second with a precision of 0.1 millimeters. Over a period of 60 seconds, our goal is to get several hundred slices of your body, reconstructing across 21 servers with two petaflops of compute power and 806 terabytes of raw data.
The slices form a 3D map with the ability to resolve internal tissue details as small as half a millimeter. Less than a dozen of these machines operating together at full speed can do more full body scans than every MRI machine together on Earth.
Our goal is to build a fleet of 50,000 of these scanners capable together of doing a billion scans a month. Enough to bring full body imaging to everyone on Earth. This is a new kind of infrastructure. It's full body ultrasonic computational tomography. No such device has ever been built until now. And yeah, we're calling it the Midjourney Scanner.
Yeah. So, they're kind of hard to make. But you can make them actually. Like we're in the year 2026. We have access to pretty spectacular technology and I know it'd be normal to just be like, you know, we're making a scanner, but actually the details are what makes something like this. And I think it's very easy to talk about the world in terms of vague problems but the world is made up of actually lots of really interesting details and the details are what you can solve them with.
Like all the little things like 1,400 meters a second, the speed of sound through water. You know, I don't have to say that but if it was the same as the speed of sound through air it wouldn't work. The fact that sound goes through water, the reason why it goes so fast through water is the reason why you can build something like this and there's many other details like that and all the details are all just there. And there's more of them, yeah.
And so yeah, that was cool. Can we go back to the slides? All right, one of the projects. What are the others? Not going to find out today. I'm sorry.
So yeah, I'll show you some quick images. So here, these are real slices. Lots of little things. We got radiologists helping us label them. It's kind of new looking, so it does look pretty different than anything you've seen before, but you can definitely see all the stuff. This is day one of a totally new medical imaging paradigm. We're not even using any AI yet. It's just really cool hardware and software and a lot of great people working to make something that we haven't built before. But obviously, these systems can be much more sophisticated over time.
This is not really the product. This is actually just the first prototype. No one's built one before and we just did. So what else we got? Yeah, here's another one. If you, I thought the slice thing was cool. Here's one more slice thing. Ain't that cool? Ah, so cool. Little feet and everything.
So you know, again, it's the first day of MRI. That's what it looked like. And the first day of full body USCT, ultrasonic computational tomography. It looks like this. And obviously MRIs look better than this today. And the ultrasound will look really incredible over the next few years.
Here's an example. Even right now. So, you know, I'm going to say this is both a fair and an unfair example. So, this is an example of your thighs with a normal full body MRI and this is the same sort of thighs with the ultrasonic CT. And you can see already there are certain ways where you can see the different boundaries between the muscles and some of the fibers and the veins that are actually already better than an MRI. We're not better in every way yet, but at day one, which is this is day one, to be better in a lot of really interesting ways is already really exciting.
It means that at the bottom, things are already easy. Sometimes when you try something new, it's good just to push and see what gives. And in this case, we're pushing and the universe is giving.
Yeah, there's little segmentations, seeing the organs. The scan we showed earlier, this is an ultrasonic phantom. So it's like meant to mimic a body as closely as possible. It's like made of gel and fake organs and everything, but you can see all the way through you get all the bones and it gets really great imagery. And it's only going to get better from here.
So now you're wondering, what does it look like to get this out into the world? We say we're going to make 50,000 scanners, but it's not going in here. It's not going at my office like that's something else has to give. So when I think about how I would personally like this to be in the future is I don't want to feel like I'm going to a doctor's office. I want to feel like I'm going somewhere that's nice. And ideally the place is so nice that even if it had no Midjourney scanner, I would still go. That's a bar for what does it mean to be nice, right? It shouldn't be nice just because there's a scanner. It should be nice.
So what does that look like? Well, you're gonna have to get wet. So, you know, that's a thing. That's one of the trade-offs we've made here. But there are lots of nice situations to get wet. No, that sounds weird. Sorry. That was not part of the speech. But what if we made something really cool? And so that's the second thing that we're going to announce, which we're calling the Midjourney Spa.
So, the Midjourney Spa, the first, we've signed a lease and we already have the designs. It will be 25,000 square feet. About, well, I'll tell you where it is in a second. It's 5,000 square feet. It's four floors. There'll be hot tubs, saunas, cold plunges, weird, more things that I don't know that are very European but great. And a little gym even too. And nine or 10 full scanners.
It's going to look, come on, it's going to look really cool. It's going to look better than any spa I've ever seen in San Francisco. Maybe in America. It's going to be really nice. It's going to be really, really nice. It's going to be so nice. Again, you would go even if there were no scanners. It's going to be around Union Square about a four-minute walk from the Apple store. Hopefully I'm allowed to say that, but that's a landmark, right? So pretty easy to get to and we're going to try to make it as 24/7 as possible because like it's mostly just machines that run without people, but you know, maybe there'll be some downtime.
And yeah, that's, I don't have another really long well-produced video for this. I didn't have time. The first two videos I made, no one, you're not supposed to know this, but this is made by our architects who made like the Blue Lagoon. So like real big spas before and they're all really excited and they're really good. I think some of them are probably here. And it's a serious project where all the things are going and I guess I didn't say we hope it will be open by the end of 2027. That's our goal.
And so, you know, maybe a little more about our goals. You know, I do like a good hot tub, but one of the purposes of this project is to gain experience in running a lot of these machines and running a lot of people through these machines. And we're also trying to understand like, if we're going to have some kind of product, which is a place that you go to get scanned, what does that product ideally look like? Is it a large spa with 10 scanners and a hot tub and a cold plunge and all sorts of, like a matcha bar maybe, or is it turned out that people actually just want a gym and they want to get a scan and get out? We don't really have any way of knowing.
So, what we're doing is we're kind of building a maximally super spa. Super spa sounds cool. And we're gonna see what people use and what they like. And then the goal is, I mean, I think this thing will probably make some money. And then we'll have a template of like this is what people use. This is the sort of return on investment time. And then through some combination of our money and probably some other people's money, we're going to just build a bunch of these places around the world.
Again, the goal is 50,000 scanners. This is 10. So that means if this is what the spa looks like, we need 5,000 of these which is doable. Maybe some of them will be bigger like 50 scanners, some of them will be smaller like two. We got to learn about those models and try to figure that out. Midjourney itself is large enough to self-fund all of this. We don't have to ask anyone money to build a spa. But at some point if we want to build 5,000, you know, this is a big project. It's like $20 billion in upfront capital expenditure. But if these things might just pay themselves off in six months. In which case, like, hey, we can build a lot of these. Even if it takes more than six months, six months would be absurd. Looks like it'll be six months, but I don't know.
A lot of things look good, but so far a lot of things have been good. So, you know, all I know is I'm excited for this place. And yeah, what else we got? I think, oh yeah, let's see what else, so a little, what does this look like? So you're going to the spa, you might go to be doing the gym and you get these scans and you just have this just very regular health data about your body. The sort of sub-millimeter scans of everything. You know, every day, this thing is, it's not, there's no radiation, there's no X-rays, there's no crazy magnets. Like you could get a scan like a hundred times a day. It may be kind of weird to do that, but like there's no limit. It's as often as you want to go.
And so the idea of like there's a whole spectrum here of things you can do. You could get a scan just once a year. I mean, you're supposed to get an MRI scan once a year if you have a lot of money. I thought I would finally do it, but it sucks. Like you're going into a tube for 60 minutes, like gunshot sounds going off. It's not, I could afford to get more, but like it just sucks. Like my body doesn't even want to go. And so like, but even just like a once a year MRI is a huge thing to have in your life. Let alone if you're feeling an upset, you know, these are not all things you're going to be doing day one but like if you're feeling an upset stomach like you can literally see the food digesting through your body, like it's kind of crazy.
Like you can just like anything that's changing over any time period you can kind of get an awareness of from a scientific standpoint. So whether that's, you know, people, there's another question we don't know is that people can just go as much as they want to go. Are they getting a scan once a day, once a week, once a month, once a year? What's that distribution look like? All these things are really important to understand if we want to build a huge piece of infrastructure around this technology.
Yeah, obviously from day one we will be, we will probably be a little conservative. You know, we have this thing called the FDA and we'll be doing some things are really easy like body composition, really easy. Some things like we can't do surgery with, like there's obviously other things that are like really crazy that we're not allowed to do. But, you know, one of the goals is to do all the things that are easy, get a lot of data, and as we get data, show out all these other things we've been doing, and then sort of layer on approval after approval, functionality after functionality until this thing can do, you know, thousands of types of diagnoses or, you know, and eventually actually treatments.
There's some weird things you can do with ultrasound, whether it's like tendon and muscle healing, and I don't want to go into that because it's not day one, but over a 10-year period, these things are not just imaging devices. They're probably therapeutic devices as well. And there's a lot of really cool things you can do. This thing has way more ultrasound than it needs to just like, I mean you kind of want to have as much ultrasound as possible to just see the stuff fast but because of that it has a lot of ultrasound that can use to do other things as that becomes normal.
I mean there are a lot of things already that people use ultrasound to look at. I'm sure you all can probably think of a few and there are actually things we use ultrasound to treat too. So to some extent this system can do anything that any ultrasound system has ever been able to do and more. It's sort of like super over-specced, but I think it's just right.
And this is the, so this was the, this is the first generation scanner. One of the things we want to do is be able to start talking about that today and what we're working on and so the team, you know, can tell their families what they're doing, so we can hire more people on the team. And so, and we want to be able to give people updates. So, this is the first generation scanner. It's more of a prototype, if you will, more than it is like a push button every day, run it a thousand times a day kind of industrial system.
The next version we're building will be this Gen 2 system. Hopefully by the end of the year, hope to do another presentation for it and it'll be, I'm trying not to spoil it, but like significantly more advanced. Like this is the first one they got to make. And honestly, it was already a big jump.
Like the temptation was, 'Let's make a little one first, David.' And I remember looking out a window and I was like, 'No, build the big one.' And it felt so good to have the freedom to just do that. We went straight to building a big one from nothing. Literally not even a small one. Just went straight to this. And so we learned a lot in the process. And the team can't wait to tear the whole thing to pieces and build a better one.
But I wanted to be able to show people what we built before we tear the whole thing apart and start that process. But the second generation one will be really cool, too. And then beyond that, we've also commissioned the new chips even. So the third generation one will actually have custom silicon that's built for this application. None of the components here are built for this. This is like the MacGyver version of this, you know, and it's already really cool. So, when you start making custom chips and you start doing everything from scratch and you throw all the layers of all the AI, all the physics simulations and all the huge servers on top of it, this thing is going to be really cool.
I guess I'm just gonna—oh, they're not on my thing anymore. But, I was going to say I guess we might be ready to take it to a Q&A.
