Latest / Bad Decisions Podcast / #63 AI is changing ANIMATION Forever with Move AI CEO
Transcript
- 0:00One of our favorite TV series, Shogun Move AI, was used to
- 0:03actually animate the digital characters.
- 0:05I didn't actually know we were being used to the VFX
- 0:07supervisor. Remember he just messaged me and
- 0:08said, hey, by the way, I used you for this scene and I used
- 0:11you for a bunch more scenes. When I looked at it, I was, I
- 0:13was even trying to work out who's AI, who's a 3D avatar
- 0:19Tino. He's the Co founder and CEO of
- 0:22Move AI, the company using artificial intelligence to
- 0:25capture human motion from video. How did you come up with the
- 0:29idea of move AI when you? Have your first start.
- 0:31It's kind of crazy. You're not sleeping much, can't
- 0:33go to the gym, so let's work out from home.
- 0:34I thought, hey, let's make this more fun.
- 0:36Let's try to create a personal coach in my living room.
- 0:38So I started setting up cameras to track my motion in 3D.
- 0:41So that's kind of the seed of what became Move AI.
- 0:46We started working with games companies, Electronic Arts.
- 0:48They said they're doing a test and it was like a David versus
- 0:50Goliath moment because they had a big Studio 100 plus cameras,
- 0:53all synchronized. I remember being terrified at
- 0:56this. And it's easy to give up.
- 0:57And sometimes you should give up.
- 1:00How did you know it's time not to give up?
- 1:02It wouldn't be true to say I definitely knew, but we just had
- 1:05a vision, had a belief that it's possible.
- 1:07You were solving a completely different problem and it looks
- 1:10like there's this whole industry that could have used your
- 1:12solution. How do you see moves role in the
- 1:15future of digital animation? Do you know?
- 1:19Welcome to the Bad Decisions podcast.
- 1:21I'm going to reach out and get. It it's great to have you here.
- 1:25Yeah, yeah, yeah. Good to be here.
- 1:27You know what triggered this conversation was the post that
- 1:31you had on LinkedIn. You mentioned that in one of our
- 1:33favorite TV series, Shogun, in one of the scenes, movie AI was
- 1:37used to actually animate the digital characters.
- 1:40And we looked at that scene and when we re watched it again, you
- 1:44couldn't even think that first of all, they were digital
- 1:46humans. Second, they are all animated
- 1:48using. Yeah.
- 1:49So the the first thing is how was the entire experience being
- 1:53in ATV series that is watched by millions of people?
- 1:57Yeah, so like, I didn't actually know we're being used.
- 2:00So, you know, the VFX supervisor, we know we've done a
- 2:02lot of stories with him recently, but I remember he just
- 2:04messaged me and said, hey, by the way, I used you for this
- 2:06scene and I used you used you for a bunch more scenes in
- 2:09Shogun, which is which is, which is really good, I think because
- 2:13like we, we try to create a tool that's so easy for people to
- 2:15use. So it doesn't really need our
- 2:16support. And then we obviously do work
- 2:18closely with certain customers who really trying to push the
- 2:20limits of what's possible with markless motion capture.
- 2:24And yeah, and that's like that show won a lot of awards.
- 2:28It's just really visually compelling, like really brings
- 2:31out, you know, authentic feudal Japan.
- 2:33And yeah, when I looked at it, I was, I was even trying to work
- 2:36out like, OK, who are the who's like, like who's who's, who's
- 2:40AI, Who's a 3D, Who's a 3D avatar?
- 2:42That's when you know you have good fucking VFX.
- 2:44Yeah, that's crazy. I know 'cause if you can notice
- 2:46it, that's when it's good VFX. 'Cause you get certain shows
- 2:49where it's just like VFX is just everywhere, right?
- 2:52And some people like that and then you get these shows where
- 2:54they're trying to bring out a very compelling story, very
- 2:57authentic story. And then, you know, you want to
- 2:59use VFX to bring to life, you know, that story without being
- 3:02imposing on the user. So yeah, it.
- 3:04Was a good experience. You know, we were watching it
- 3:05with our parents and we had to stop the show multiple times and
- 3:08go back and they were like, why are you guys doing it?
- 3:10We were like, OK, we have to find out exactly who's the
- 3:13digital character, what tools they use.
- 3:15Because there was the post that you made on LinkedIn that said
- 3:18around 30% of the characters you see in this shot are not real
- 3:22and they're using movie AI to animate them.
- 3:25And then what we did was we went ahead and started doing more
- 3:27research and then found some of the behind the scenes, most of
- 3:31the characters are not even there.
- 3:32Like when they're actually shooting it you, they just have
- 3:34the four gown characters and everything else is CG, which is
- 3:38what's so fascinating because the show's done it so well.
- 3:41The production quality is just so high that you are so immersed
- 3:44in the storytelling and the cinematography and the acting is
- 3:48as well as marvelous that you like.
- 3:50This is beautiful. So thank you for making that
- 3:52post because if it wasn't for the post, we wouldn't be here.
- 3:55Right now it's all good. I think what's also interesting
- 3:57is so I'm not from AVFX background, but what's
- 3:59interesting is that, you know, they, they shot the scenes
- 4:02initially and then a lot was added post And so, so maybe you
- 4:05have an idea of what you want to do for the scene.
- 4:07You know, you, you, you go on set, you go to location, you
- 4:09film it and then the director or you know, whoever's like, hey, I
- 4:13want to, I want to bring to life there.
- 4:15So I want to make the army bigger.
- 4:16And then the cool thing is you can start to add these things
- 4:19afterwards. And so you can, you can, you
- 4:20can. This iteration cycle is just
- 4:24like much more flexible. Yeah, but there's another
- 4:26director, Gareth Edwards, who does it this way.
- 4:28He likes to shoot everything. That's what he did for the
- 4:30creator. And then after the fact, they go
- 4:33ahead and add to the story. So they go ahead and turn.
- 4:35For instance, this shot, they shot in the village of a lady
- 4:39and a child playing with grains. They didn't tell them to act in
- 4:43a robotic way, so they filmed them as they were without any
- 4:47markers. And then after the fact they
- 4:49turned them into robots. So now you have this human
- 4:51acting but with a robot body, which made the entire scene a
- 4:57lot more authentic in a way. And if you went ahead and
- 5:00started putting markers on them and started adding clothes to
- 5:02them, then the acting would have been slightly different.
- 5:05It wouldn't be so authentic. So this way of iterative film
- 5:09making is very cool. I actually love it.
- 5:12I didn't know Shogun was doing something similar to that.
- 5:14Yeah, yeah. I think also different people
- 5:16would like to create content differently.
- 5:18They have different styles or, you know, a particular style is
- 5:20better to bring out a different story.
- 5:23And so I think the more flexible you can create these tools just
- 5:26gives more power to the creator to create whatever they're
- 5:28doing. I mean, have you watched Shogun
- 5:31yourself? I've seen, I've seen some
- 5:32scenes. Yeah, OK.
- 5:33We literally started binge watching it and we cannot stop,
- 5:36man. It is.
- 5:37I mean, I'm fascinated. I think Farad is too with the
- 5:40Japanese culture and we have never been to Japan, but we have
- 5:44to freaking go because there's so many things that we love
- 5:46about. The food is amazing.
- 5:47Of course the. Culture.
- 5:49The culture. And the history and the samurai
- 5:52and the way they, the way they fought is just so different than
- 5:57what you've seen Western movies. You know, it's a completely
- 5:59different world. And so that's what I'm
- 6:01fascinated by. And we haven't finished it,
- 6:03though we've had. Two more, two more, two more.
- 6:06Yeah. Speaking of your background, you
- 6:08mentioned that you are not from the VFX background.
- 6:10Let's take it a little bit back before even you were starting
- 6:13Move AI. What were you studying?
- 6:15How did you come up with the idea of Move AI?
- 6:17Yeah, so, so my background is I guess technologists like
- 6:20engineer, like as a kid, as a teen, I've just always been into
- 6:25tech engineering, building things also super into sports.
- 6:28So then the the kind of precursor to move was I was
- 6:30doing my PhD at Imperial College London in engineering.
- 6:34So I was actually doing material science.
- 6:35So I was modelling materials and how they fracture and how they
- 6:38how they deform. So that was that was like my
- 6:41core, my core thesis was analyzing particularly from in
- 6:46this phenomenon of a fracture from blunt cracks.
- 6:49That's that's what. You were studying that for
- 6:52years. What is the purpose of this
- 6:53because it's very specific. Yeah.
- 6:55Well, the the purpose is you want to understand material, the
- 6:58better you can understand materials, particularly, you
- 7:01know when they break or when they fail, the the better you
- 7:04can push materials to the limit. So take an aircraft, right.
- 7:06So like, like when you design something, often you have to
- 7:09build in redundancy just in case.
- 7:11And particularly with an aircraft or something like that
- 7:13or building, you know, you want to make sure that if it does get
- 7:16pushed to its limits, it doesn't just collapse and paste.
- 7:19And, and one of the main failure modes of materials is fracture.
- 7:23So, so materials can fail by just deforming too much, but
- 7:28the, one of the, the most catastrophic ways it can fail is
- 7:30where it fractures, where you get a new surface created.
- 7:32So the better you can understand when will a material fracture,
- 7:37the better you can understand it, then the better you can push
- 7:39materials to the limit and then maybe reduce some safety factors
- 7:42for particular applications, whether it's like a turbine
- 7:44blade in an engine or the OR the hull of an aircraft or a
- 7:47building or whatever. Wow, so your your degree would
- 7:51be used for construction aircraft building?
- 7:54And so I was like at the very, very low level.
- 7:56So like how it would be used, like don't know.
- 7:59So. So I was actually studying like
- 8:01the precise area around a crack tip.
- 8:04And so I was modelling it in 3D. So I was using software like
- 8:07Abacus, like the finite element method, so numerical methods to
- 8:10try to model the material and when it will fracture or when it
- 8:14will break. Also doing some like
- 8:16mathematical modelling and geometric modelling too.
- 8:19And then I was using computer vision to actually do
- 8:21experiments to analyse what's happening around the crack tips.
- 8:25So that's when I first got into, I guess AI computer vision kind
- 8:30of image processing. I failed to see the connection
- 8:33of that to move. I was like, OK, we're not
- 8:36fracturing anything when we're using our motion.
- 8:39The only part is computer vision probably, but let's see where
- 8:42it. Goes.
- 8:42I want to see where this goes. It's coming.
- 8:43It's coming. Yeah, maybe there's a few twists
- 8:45and turns. OK, so I mentioned that I I
- 8:48like, I love playing sports. I love, I love working out.
- 8:51So around towards the end of my PhD, my first child was born,
- 8:55who's now 7. So it's about around seven years
- 8:57ago. And even if you have kids.
- 9:01No, no. When you have your first child,
- 9:02it's kind of crazy. You know, they're not sleeping
- 9:04much. You know, your, your routine is
- 9:06messed up. And so I didn't have time to go
- 9:08to the gym and but I, I like to work out.
- 9:11And at this time you're studying still.
- 9:13Yeah, I'm still doing my PhD, right.
- 9:16So then I'm like, OK, I need to go to the gym, can't go to the
- 9:19gym, so let's work out from home.
- 9:21So. So I start to download these
- 9:22apps, they just say like they just give you a routine of like
- 9:26do press ups and like do squats, do sit ups.
- 9:28I just find it really boring. So I thought, hey, let's make
- 9:31this more fun. Let's try to create a personal
- 9:32coach in my living room. So I started setting up cameras
- 9:36in my living room to track my motion in 3D.
- 9:39So that's kind of the seed of what became, I guess move AI
- 9:42like a few years later. So I was just trying to
- 9:44understand my own motion in 3D. Trying to create a coach that
- 9:48can automatically count your sets, your reps.
- 9:51So you would squat and you would measure the distance and every
- 9:55time, for example, you you hit the distance, you would say, OK,
- 9:58first Rep, second Rep, third Rep.
- 10:00Is that what you were doing? Yeah, yeah.
- 10:01And I was trying to give like real time feedback too.
- 10:04So I was just creating this AI or this model I can give you.
- 10:07Like imagine there's actually a coach in your living room with
- 10:10you and who and who'll be there every like every day, like in
- 10:13the morning, evening. So I'm trying to create this
- 10:15coach that could be with you every time you work out from
- 10:17home. OK, so where is this application
- 10:18now? I actually need it, yeah.
- 10:21Yeah, it didn't happen. So it was a bit too early, I
- 10:25think because it's expensive to the processing, the quality
- 10:29wasn't quite there to actually give you detailed insights.
- 10:32Right. OK.
- 10:32How many cameras you were using at that time?
- 10:35I I was experimenting with one to like 8 probably.
- 10:38Is it this is in your living room?
- 10:40Yeah. What did your wife think about
- 10:42it? She was also working out.
- 10:45She was a test subject, OK. Test subject number one.
- 10:49Yeah. And then your kid was not big
- 10:51enough at the time to do any any tests for you.
- 10:53Now he's big enough and he's a test subject, too.
- 10:55Oh yeah, test subject. OK, so you've got your own set
- 10:57of test subjects. I can't imagine Tinos has like
- 10:59cameras everywhere. In the, however, moves emotions.
- 11:02Are captured for training and quality purposes.
- 11:05Yeah, yes. OK.
- 11:06So then so I actually created a business at the time to help use
- 11:10that early technology I built to help sports teams.
- 11:13So I started working with like the UK boxing team, UK tennis
- 11:15team and other and other football teams, ice hockey
- 11:18teams. How?
- 11:19How did you help them? So there's kind of three things
- 11:22in particular. So one is they'll give me videos
- 11:24of opponents and I would use that the AI and the computer
- 11:27vision to to get physical performance characteristics and
- 11:31then automatically create highlights I see so they can
- 11:33analyse and create game plans against opponents.
- 11:36Secondly, helping using the technology to help them get
- 11:39performance characteristics of youth players to bring them,
- 11:43help them decide who will be in the professional program.
- 11:46Oh, selection you mean? Yeah, selection, yeah and and
- 11:49then also just tracking train. The third thing is tracking
- 11:51training to because like when you perform in a match like you
- 11:56can't like you can't put sensors yes, of course the athletes.
- 12:00So the more information you can get on on when they're
- 12:02performing and competing and then pairing that with training.
- 12:06So I've been training to to then benefit them for for the
- 12:09matches. That is really cool.
- 12:10You know why? Because I used to play League of
- 12:12Legends and I used to hold. That related now, yeah.
- 12:16I'll tell you why. So I used to also host E sports
- 12:19tournaments and then I would see after every game the players
- 12:21would go in with the coach and the coach would have all the
- 12:23stats from the game. Because the game does give you
- 12:25all the stats of the enemies of the opponents.
- 12:28So you have their attack, their defence, everything in numbers.
- 12:31But when you're playing boxing, you don't have any of those
- 12:34stats. Were you sort of giving these
- 12:37kind of stats? I mean, definitely.
- 12:38Attack and defence, me too. So when you play games, there's
- 12:41you know, it's a 3D experience and all the data is available,
- 12:44right. But when you play sports, it's
- 12:46like real, the real life, real world, you don't have this data,
- 12:49right. And so the challenge is if you
- 12:51can get that data, then you can start to create game plans and
- 12:53get, you know, do strategies, tactics more effectively, right,
- 12:57and then improve training as well.
- 12:58So it's essentially you could say I was trying to digitize the
- 13:01real world, just like a 3D World is digital in a game.
- 13:05Yes, wow. And and, and you know, the world
- 13:07is just going more toward that way when you start to digitize
- 13:11voices motion, you know your your body as well.
- 13:15Yeah, no, that's very cool. So this is in the living room.
- 13:19You're doing these these tests and then you're working with
- 13:21some companies now and experimenting what happened
- 13:24after that. Yeah, So it was a lifestyle
- 13:26business. It was super fun.
- 13:29I just found that the sports industry is quite different to
- 13:33the games industry, for example. So they tend to be later
- 13:35adopters of technology. Also culturally, they just tend
- 13:38to invest less money into sports and tech.
- 13:43So then so then I raised some funding to then start Move AI to
- 13:46essentially take this core idea and start to look for other
- 13:49opportunities outside of sports. And that's how how Move AI was
- 13:54born. So we raised some funding, spent
- 13:57about 3-4 years improving the technology so we could get to a
- 14:02higher standard. So as soon as we started working
- 14:06with games companies, actually the standard of quality needed.
- 14:09And, and you guys are from this background like VFX games, like
- 14:12they need a high quality of data so that they have low clean up
- 14:15afterwards and so they can ideally go directly in engine.
- 14:18So we, we found that we had to keep pushing the technology and
- 14:21it took about 3-4 years before we reached this threshold where
- 14:25Electronic Arts, they so EA they, they did a, a test.
- 14:30I remember, I remember being terrified at this because,
- 14:32because they just kind of did it behind the scenes.
- 14:34They said they're doing a test. And I was like, OK, well, let's
- 14:37see how this comes out. And it was like a David versus
- 14:39Goliath moment because they had, they had these obstacle tracking
- 14:42cameras. You know, you put on the suit
- 14:44and you've got the dots. Was it Vicon?
- 14:46Yeah, I think it was Vicon, Yeah.
- 14:48And so they had a big studio 100 plus cameras, all synchronized.
- 14:53You know, typically you need like 50 plus markers on the
- 14:55human body like like particular lighting that actually dressed
- 15:00up a guy. He he kind of looked like a like
- 15:02a Christmas tree I guess because he was in like a multi coloured
- 15:05suit. He had an inertial base system.
- 15:08He had the markers for the optical system.
- 15:10So he's got both now, so. He's wearing both and they're in
- 15:12that kind of studio designed to capture optical data.
- 15:16So, you know, with the with the suit and the markers as
- 15:19effectively as possible. And then they used to set off
- 15:20some Go pros unsynchronized and and then yeah, the results came
- 15:25out and they were comparable to the to the obstacle.
- 15:28So that's that's a huge moment for us.
- 15:29So that's after four years working on the end.
- 15:31Yeah, yeah, like grinding for for a lot of time.
- 15:34And with this company and other companies, you know, we'll give,
- 15:37you know, we'll maybe show them results like after three, six
- 15:40months. And they said, yeah, it's
- 15:42looking good, but like it's not quite there yet.
- 15:43So you kind of get punched in the face and then you just have
- 15:46to come back later and then show them more results.
- 15:48So we just kept working and we got an excellent research and
- 15:51engineering team who just kept kept working on this using
- 15:54techniques and computer vision, bio mechanics, physics,
- 15:57statistics. So we got a multidisciplinary
- 15:59team to just try to solve this problem of capturing and
- 16:02understanding human emotion. And yeah, initially we were at
- 16:08least the area where we use the most right now is games, video
- 16:11games, film and TVVFX. And then increasingly we'll see
- 16:15other other industries as well, right?
- 16:17I think we need to focus on this three to four years because I
- 16:20feel like these three to four years might have already gone by
- 16:23and you guys are definitely a lot more successful now.
- 16:26But these three to four years is what a lot of people have, and
- 16:30they give up within that three to four years because it's the
- 16:32critical time where you're building.
- 16:34No one believes in you. No one understands what you're
- 16:36building. You have to prove to everybody
- 16:38that what you're building is actually going to be successful.
- 16:42And I want to go back to the beginning because you said you
- 16:45raised some money. Was this done publicly?
- 16:48Is this something that you've disclosed, like the amounts and
- 16:50how it was done? Because I feel like a lot of
- 16:52people might have the idea they need the funding.
- 16:54They don't know how to do it. Yeah.
- 16:56Yeah. So we raised around 10 million
- 16:58in total, but I I think at the time we'd raised maybe 1 to
- 17:01$2,000,000. OK, so not a crazy amount.
- 17:04Yeah. And we just had a lean like
- 17:07research focus team. I tried to hire the best people
- 17:10that we could find. We got somebody who was in my
- 17:13research group at Imperial. So he joined and we hired a few
- 17:16other people. But you're right, it is that
- 17:19kind of pre product market fit is is really tough and it's easy
- 17:25to give up and sometimes you should give up.
- 17:27How did you know it's time not to give up at that time?
- 17:30Because I'm sure you, you brought your product, you went
- 17:33to EA or whatever company and they said not good enough.
- 17:36How did you know that you could go back and actually improve it
- 17:39and come back another time? Yeah, I guess you don't always
- 17:41know, right? Yeah, it wouldn't be true to say
- 17:46I definitely knew like with 100% certainty you should keep going.
- 17:50But we just had a vision, had a belief that I like, I think like
- 17:54I'm, I'm confident or I'm like, I'm let's say 9070 plus percent
- 17:58sure that it's possible. And we just, yeah, we just had
- 18:02the vision, had a passion for the industry, had some
- 18:05resilience from other past experiences to just where I'd
- 18:08failed and stopped, I think for the right reason.
- 18:12And then this time I just said like, let's, let's just keep
- 18:14going. Let's do it.
- 18:15I think in our last episode we had the founder of Epidemic
- 18:19Sounds and he said that we as humans tend to make a lot of
- 18:22logical decisions for small decisions.
- 18:25But when it comes to the big life decisions, we're making it
- 18:28emotional, something deep inside us and our gut tells us to keep
- 18:32going. Did you have that in you?
- 18:34Like you, Even if people were telling you this is not, this is
- 18:37not going to work. Like this is not good.
- 18:38A big company like EA won't accept it.
- 18:40They have their set up, why would they come and use another
- 18:43product which nobody heard about?
- 18:45Yeah, I definitely believe that it will be possible.
- 18:48I definitely believe that it would be valuable to people.
- 18:52And then what's crazy is there's other bigger opportunities even
- 18:56that have now come up. And so that's kind of grown.
- 18:58Like I definitely didn't have that idea in the 1st place.
- 19:01But I guess like a hobby to me, like I work on the business,
- 19:04like as a hobby, you know, I could spend way more time
- 19:07working on it if I had the time. And and even before it was a
- 19:11company, like I was working in the space.
- 19:13It's just a passion. So I think you have to have that
- 19:14passion, yes, because it's a lot of people talk about this like
- 19:18unless you have that that core passion, you come across these
- 19:21roadblocks and it's very easy to give up.
- 19:24What's your working schedule like in those four years?
- 19:28On those four years, like you have to put a lot of hours in.
- 19:32Yeah, I've got like 4 kids and so you got to manage family as
- 19:36well. So just making sure that, you
- 19:38know, I, I, I place a high priority on family.
- 19:41So I'd make sure that that's taken care of.
- 19:44And then also just a lot of hours, even before move, you
- 19:47know, I'd easily spend, yeah, a lot of hours into the late
- 19:51nights. Yeah, just coding and doing all
- 19:54the support and and everything myself.
- 19:56Wow. Let's talk about the technology
- 19:58a little bit because when you guys go back and keep improving
- 20:01it. Which side of the technology you
- 20:03guys think you had a breakthrough that you were able
- 20:06to reach the threshold that these big companies are asking?
- 20:10Yeah. I think it's a lot of
- 20:11incremental things that kind of get you there.
- 20:14We definitely had some kind of, I guess like scientific kind of
- 20:17breakthroughs, engineering breakthroughs.
- 20:18But it's just these incremental things add up over time.
- 20:22And particularly for the industry, VFX video games, there
- 20:25is a certain threshold you do have to get above.
- 20:27So we got above that. And then as you want to, let's
- 20:30say we get to 95% accuracy. To get those incremental ones,
- 20:33you have to do a lot of work. And so in the early days,
- 20:37sometimes you can get rapid progress, but then often to, you
- 20:41know, to keep pushing, you have to put in extra work to really
- 20:43push that push that accuracy level.
- 20:46And so I'd say it's just, it's been like very incremental and
- 20:48we've explored a lot of areas. So like we tried to model the
- 20:51human body a particular way. We tried to model the physics to
- 20:56make sure that the motion obeys physics.
- 20:58Because if you're capturing motion of somebody in real life,
- 20:59it's obeying physics. So you should also obey that.
- 21:02When you translate it to a 3D model, you'll be using
- 21:06techniques from AI to like train data set to learn from how
- 21:10people move. So capturing motion of people
- 21:12and then training models. You do prediction, for example,
- 21:16on the actual motion capture that you guys have, it's.
- 21:18Kind of prediction and clean up. So for example, you want to
- 21:22capture someone's motion in real life, you know, you could in
- 21:26some cases you can like load somebody up with either physical
- 21:28sensors or like hundreds of cameras.
- 21:31But in reality, when you put it into production, you don't have
- 21:34that. You may, you might just have a
- 21:35single camera, you might have 345 whatever cameras, you might
- 21:38be trying to do an entire football pitch, soccer pitch and
- 21:42you only have, you know, it's just expensive.
- 21:45Therefore, you know, we'll create models that try to help
- 21:48you get to the high level quality with less hardware as
- 21:50well. And so that's where we'll use a
- 21:52lot of a lot of AI to essentially create a fundamental
- 21:55model of how people move and then try to help people get that
- 21:59same level of quality with less hardware when you put into
- 22:01production. I find a product that you guys
- 22:04have created fascinating. So just a little bit of a story
- 22:07on our side. We come from the artist world,
- 22:10so we use it to create content, but also for services, for 3D
- 22:14animation, let's say. And we started with inertia
- 22:17suits. So for anybody that doesn't
- 22:19know, those are the suits that you have to put on and they cost
- 22:24seven, $8000 that that's usually the price it starts at.
- 22:27Starts at yeah. Starts as, which is pretty steep
- 22:30if you are just beginning. If you're an indie freelancer,
- 22:33it's very hard to get a suit like that.
- 22:35And then you can go all the way up to optical systems, which you
- 22:40have to set up multiple cameras for.
- 22:41They go way above that 7020 thousand, $50,000 price pool
- 22:46price, pool price range. And that's when you go to the
- 22:50studios, right? That's the studios for AAA title
- 22:53games or video for movies that are using that.
- 22:58When we started, we had the suits, and honestly, when you've
- 23:02wear the suit for the first time, it's a great experience
- 23:04because you feel like you're getting into the mood of acting
- 23:08and it's bad ass. But then when you start actually
- 23:10using the suits, you realize first of all, it's limiting my
- 23:12emotion. I don't really feel like I can
- 23:15act in an authentic way. It feels kind of weird.
- 23:18Then you have to calibrate every single time you want to act, and
- 23:22that makes it even more restricted for you to want to
- 23:25express yourself every single time you want to create an
- 23:27animation. And then there's the price
- 23:30point, as we discussed. And also, sometimes your hand
- 23:32stops working, yeah. So you'll you'll have to
- 23:35recalibrate that again. Exactly.
- 23:37So for us, when we first heard about Move, just like you said,
- 23:41actually before we started the podcast, we didn't actually
- 23:43believe this was a thing. We thought, OK, it's just some
- 23:45gimmicky thing. Then we tested it.
- 23:48You guys didn't have Move one at a time.
- 23:49You guys had Move with multiple cameras.
- 23:52We were blown away because I remember our first animation, we
- 23:55laid on the ground and then we got back up and that was
- 23:58something the suits couldn't do. Like the legs would all start
- 24:01flipping around and you guys did that perfectly.
- 24:03It would lay on the ground and get back up.
- 24:06And that's when we realized there's true potential in this.
- 24:09And now Fast forward all the way to today, when you guys have
- 24:12moved one, you can set up our iPhone, just one iPhone and
- 24:15shoot yourself within a few seconds.
- 24:17And then you go from your idea to the motion being create it,
- 24:23and then boom, you can take that into Unreal Engine.
- 24:25And after about 30 minutes, you've got the animation that
- 24:28you performed in your room. That is revolutionary for me and
- 24:33for Farrah and whoever's been working in bad decisions.
- 24:36So for us, we see exactly the potential.
- 24:40For you yourself, how do you see the market of motion capture
- 24:44changing? Because of course, we had the
- 24:45initial suits. We have the the Vicon cameras
- 24:49and now we have Move AI. There's many different markets
- 24:52this can go into. We've got Avatar level movies
- 24:54where they have the the craziest technologies.
- 24:57Underwater motion. Capture, underwater motion
- 24:59capture, and then we have studios, smaller studios which
- 25:03create animation, and then we have indie content creators just
- 25:06on their own alone. How do you see the market
- 25:09changing in the next few years? Yeah, I appreciate the kind
- 25:11words about the product. Yeah, I've got some ideas, but
- 25:15also my mind gets blown every every kind of few months.
- 25:19It used to be like every like quite regularly.
- 25:22And now I've I've seen a lot, but there's still a few things
- 25:24that I'm really excited by. But just for, for I guess
- 25:28content creation. So, you know, we've got we've
- 25:30got like these content creators on YouTube, TikTok, where
- 25:34there's somebody who's a big fan of Pokémon and they create like
- 25:36this Pokémon fighting scenes, let's say, or there's somebody
- 25:38who wants to be a Motown kind of singer.
- 25:40So they create a virtual version of themselves.
- 25:42They emotionally capture themselves.
- 25:44They either use a digital voice or real voice.
- 25:46But you know, these people have thousands, millions of
- 25:49followers. And so I think what's really
- 25:52interesting is when you create a tool for people, you know, they
- 25:55naturally will use it in lots of different ways.
- 25:57And then, you know, you see the output and then some of your
- 26:01mind is blowing with the type of content when they, especially
- 26:03when they're combining different tools together.
- 26:05And so I definitely see content creation being democratized in,
- 26:09you know, in 3D, as you guys know, 3D content creation is is
- 26:13difficult, expensive, it's time consuming.
- 26:16Yes, it's a high learning curve. But I think even with 2D content
- 26:21creation, that was the case. You know, there's a time when we
- 26:24had, you know, there were not that many photographers and, and
- 26:28image editors, right? Let's say from the 90s, two
- 26:30thousand, you have to have this huge Canon camera, you have to
- 26:33edit it. Whereas now you've got billions
- 26:36of, you could say 2D2D photographers and 2D image
- 26:40editors, right? You know, you've got billion
- 26:42plus people. You can just load up your phone
- 26:44and then edit it, right? Yeah.
- 26:46And and then that's because the technology became more
- 26:48accessible. It became more powerful, became,
- 26:50you know, it's easy to use and it's cheaper right now it's in
- 26:54everyone's pocket. And content creation I see going
- 26:57that way. And so the more we're creating
- 26:58these tools, whether it's move or, or creating a clone of
- 27:02yourself, your voice or, or what you look like, and then altering
- 27:07that, you know, as, as, as this tooling becomes easier to use, I
- 27:10see content creation becoming easier.
- 27:13And then we'll start to see new use cases coming up.
- 27:15And, and that's, and this is just within, you know, content
- 27:18creation for, for move also, like I see, I see use cases
- 27:21outside of content creation. So we've been used increasingly
- 27:23in industrial use cases, training, training robots, life
- 27:29sciences for rehab and diagnosis of conditions.
- 27:33Again, they're quite nascent areas, but it's it comes from,
- 27:36it comes from interesting, you know, people who are curious and
- 27:39at the leading edge of tech by trying things out in their
- 27:41industry as they see these new tools becoming more available
- 27:46and then starting to to piece piece things together.
- 27:49Well, there is also the event side of things that me and Farad
- 27:53are very much interested in. So one of them would be the
- 27:56Fortnite events and you guys did a recent one and this one was
- 27:59with Miles Smith for Fortnite. And to my knowledge, I believe
- 28:05up until this moment they've been using really high tech
- 28:08technology in terms of hardware to create these motion captures,
- 28:12but now they're using Move. Is that right?
- 28:14Yeah, that's right. When was that Fortnite concert?
- 28:18I think it was around October it launched.
- 28:20This October. Yeah, like 2024.
- 28:22Then we used moved in fortnight much earlier.
- 28:25We used it last April, yes. So if you look at the avatars up
- 28:28there, so we we use that for Abu Dhabi Autonomous Racing Lake and
- 28:32we made a fortnight map for them.
- 28:33OK, so the entire trailer was animated using move AI.
- 28:36Awesome. Yeah.
- 28:37So we had all these avatars that you want to talk to you.
- 28:40Yeah. So we did it in April.
- 28:42Yes, in October. And.
- 28:43Then we did a recent one for Dubai Police.
- 28:45We created Downtown Dubai and Dubai Marina, and then we took
- 28:49that into Unreal Engine. Fortnite Editor.
- 28:51Yeah. And then we had the SWOT
- 28:52character who was doing the intro.
- 28:54So we like to do cool cinematic intros like video games.
- 28:57You've been. You've been.
- 28:58Sitting on this information. Yeah, we wanted to kill.
- 29:01Yeah, yes. So we had the SWAT character do
- 29:05an animation, and I'm not joking.
- 29:07It took us 30 minutes, literally from the idea.
- 29:10We had it on Figma and then we wrote down the script and then
- 29:12we I I found. Out the camera.
- 29:14So the camera was set up right there.
- 29:16In this room. In this room.
- 29:18In this room. I'm working on that chair.
- 29:20Faras is walking in the room acting like the police squad
- 29:24head, and then he's talking in front of the camera.
- 29:26In 30 minutes I can see the motions in Unreal Engine.
- 29:30Five yeah, I I'm, I'm going to be honest, you guys are doing a
- 29:32fucking fantastic job because we have been there when we use the
- 29:36suits. And I do want to give a shout
- 29:38out to the guys who worked on the suits, because the suits in
- 29:40their own time, they worked great, but they were limiting
- 29:44for artists. They were limiting.
- 29:46And I can see that a lot of new and up and coming artists will
- 29:49always be limited by the fact that they need to pay a certain
- 29:53amount of money to get into it. So there's that upfront cost to
- 29:55get the suit. And then there's a problem that
- 29:58you cannot really express yourself freely if you can't
- 30:02afford the suit. You have to go to Miximo and use
- 30:04somebody else's library of animations to create your own
- 30:08story. And that will never be your own
- 30:09story. So, for example, for Fortnite,
- 30:12we managed to create our own story because it was my own
- 30:14motion. And the best thing is you don't
- 30:16wear anything. That's the best fucking part.
- 30:18So you guys literally just went in and you.
- 30:19Said you didn't wear anything. You know all your videos
- 30:23actually, you guys have fucking access.
- 30:28I was wearing something. OK, Yeah.
- 30:30Just for everybody, no markers, no markers, no markers, no
- 30:34markers, no markers, no suits, which made it so much easier.
- 30:37And then the setup is also the thing.
- 30:39So whenever we were using the mocap suits, one of us had to
- 30:43take 10 minutes to get into the suit.
- 30:45Put the cable, put the battery behind the yeah and then you
- 30:47have to person. Yeah, you cannot.
- 30:49Actually, it looked kind of weird and cyberpunky because
- 30:51then the battery pack was behind Farhad and then I had to hook
- 30:54him to the computer and he couldn't move.
- 30:56It was like the weirdest thing. And now we don't have to do any
- 30:59of that. So I I don't see a future where
- 31:02Muvaya is not part of the picture.
- 31:04I I see that future where Muvaya is just taking over in terms of
- 31:08motion capture technology. Of course, for us, it has to do
- 31:11a lot with the animation world. But as you mentioned, there's so
- 31:15many other use cases that you yourself are surprised by when
- 31:19you hear about it. Because I'm sure people are
- 31:21reaching out day by day, either on LinkedIn or other platforms
- 31:24just showing, hey, we use Move AI for this.
- 31:26Which one was the most surprising to you?
- 31:29Most surprising? Most surprising use case of Move
- 31:31AI please. Not only fans.
- 31:34Or or maybe only fans. Actually, could be, never know.
- 31:38I'll have to think about it like 1.
- 31:40Interesting. 1 was like digital invitations, so.
- 31:44What do you mean by that? So, you know, you want to invite
- 31:46somebody to your wedding, you want to invite someone to your
- 31:48birthday party. You get these websites where you
- 31:50can, you know, you put in that e-mail address.
- 31:53So they get like an animated card, right?
- 31:55But then they're all looking at having like a 3D avatar kind of
- 31:59invite you. So so imagine it's like Elsa
- 32:01from Frozen inviting your friends to like your 6th
- 32:04birthday party or something like that.
- 32:06I see. So that like I found that like
- 32:08mind blowing because it's, you know, these these products have
- 32:12hundreds of millions of users, like active users.
- 32:14You know, there's events happening, I think like bridal
- 32:17showers, kids birthday parties were like the the really popular
- 32:21kind of use cases. And yeah, I wanted to bring 3
- 32:25DA, more personalized invite, a more creative creative invite
- 32:30like that, that I found that really interesting because of
- 32:32the scale. And then also just the use case
- 32:34that I hadn't thought before. Yes.
- 32:36And what about the the future use cases that you think movie I
- 32:40would be extremely suitable for the application of it and you
- 32:44think it has not been used yet, Is there any that you have in
- 32:47mind so? Just before I answer that, So
- 32:48you're talking about the suits? I don't see suits becoming
- 32:51obsolete. I definitely.
- 32:52You don't. Yeah, well, I'm excited by the
- 32:57tooling getting better for everyone, whether it's suit suit
- 32:59based or, or or non suit based and motion capture and creating
- 33:043D animation like that. Market expanding.
- 33:06I think it's gonna expand significantly as it's cheaper
- 33:10and easier to create this 3D content.
- 33:13Of course, you know if on the consumption side like with
- 33:16headsets, you know that can also increase the market further
- 33:19because that natively doesn't need 3D.
- 33:21But even without that, there's, there's a lot of growth as the
- 33:23tooling becomes more effective. I mean, you know, the suits are
- 33:26useful for like situations where you can't put a camera.
- 33:30And so like that, I don't see the suits going away forever,
- 33:33but it might even increase, you know, in the long run because
- 33:36the market expands as 3D content creation's easier, as motion
- 33:40capture's easier. But yeah, I definitely see move
- 33:43being used in in more and more more work flows.
- 33:47I see. And then the use cases, yeah,
- 33:50that you think Move AI would be perfect for like a market or an
- 33:54application that you haven't seen yet, but do you think Move
- 33:57AI will go in that? Market yeah, so so one
- 33:59interesting area is so we launched this real time system
- 34:01so on move live. So this gives you a good quality
- 34:05motion. It's optimized for speed and low
- 34:07latency. So the quality is still good,
- 34:09but it's not at the, you know, post processing multi camera
- 34:13level quality and so that will give you data back within around
- 34:16100 milliseconds. Wow, that's.
- 34:18Insane into and you can put into Unreal or Unity.
- 34:22How many cameras? Do you?
- 34:23Do you need? For that, like we recommend 4 at
- 34:26a minimum, OK, but it can be. They're not like expensive
- 34:29cameras. They're like a few 100 bucks a
- 34:31camera. But these are specific cameras.
- 34:33Yeah, like a Fluor camera, you can just order it and it comes
- 34:36within within the week. This would be good for live
- 34:39entertainment. Yeah, imagine we had Yogis in I
- 34:43think the last three episodes. They are handling all the
- 34:45production for Tomorrowland, which is the music festival and
- 34:48they use Unreal Engine for all the LEDs.
- 34:50I can't imagine they can put 4 cameras around the DJ and
- 34:53capture their motion because what they have done with
- 34:56Tomorrowland is that everything that is in the physical
- 34:59environments reflects in the Unreal Engine board time of the
- 35:02day. If it's real time, it's raining,
- 35:04it's all real time. So I can imagine if the DJ turns
- 35:07into a character and it's within the LED's they can use that.
- 35:10So this has it been used for the music festivals or anywhere?
- 35:14Yet, yeah, I mean, it's a relatively new product so that
- 35:16people are still building the integrations.
- 35:18But we've been used by like, like last year, three Nike
- 35:22stores as part of an activation for the EUR, the, the football
- 35:27competition. And they, they set up these kind
- 35:31of units near the entrance of Nike store.
- 35:35So it was London, Berlin, Paris. And you know, you can, a
- 35:39consumer or like somebody walks in and can just jump in the
- 35:42space and then your motion was captured and then it was
- 35:45transferred in real time to an avatar.
- 35:46And you can see yourself move and and they did a particular
- 35:49activation around the the fabric going from hot to cold.
- 35:54Oh, I see. Yeah.
- 35:55But the, the, the interesting thing was what in general is
- 35:58that when, when you, when you're being captured in real time and
- 36:03then either you see yourself how you look or like as a different
- 36:07type of avatar, it's really, it's really engaging.
- 36:09So for example, in the store, you have different music coming
- 36:11on. So they'll be like like
- 36:13Bollywood music or something. And then people like dance like
- 36:16that style. And then different type of music
- 36:17comes on. Then people dance that and, and
- 36:18they'll just spend time in the, you know, in the, in the
- 36:22section. We've even gone to a few parties
- 36:24like or a few events and we just set up the system just for like
- 36:28networking and like somebody requested it and then people
- 36:31would hog the system like people would just.
- 36:33Like spend like 20 minutes a second there's.
- 36:36Like a, there's like a line of people and there's just a one
- 36:38guy and I think 1 case there's like a mocap actor and he could
- 36:41do like flips and, and like all these gymnastics and he was just
- 36:44like doing all this crazy stuff. The system was working and then,
- 36:47you know, people were just watching it on the screen and
- 36:48then everyone was like, hey, I want to jump in now.
- 36:51And then they don't want to. They don't want to let it go.
- 36:53Yeah, we, we've had some, I've seen some use cases where people
- 36:55have set up the system and they're live streaming like a DJ
- 36:58performance on Twitch. And so as yeah, like as people
- 37:02start integrating into their workflows, I definitely will
- 37:04definitely see more and more interesting use cases coming up.
- 37:07This is like a human psychology. Humans like mirrors and this is
- 37:10mirror on steroids. You can, you can even change
- 37:13your output, change your size, change everything.
- 37:14And I, I believe it definitely could increase the time spent in
- 37:19the shop, in the store probably they bought more stuff as well.
- 37:22So it's a great marketing campaign.
- 37:24So when you guys are looking at. Different So just on this point
- 37:27though, so yeah, people love mirrors.
- 37:29And then so I mentioned like how I test out the product on me, my
- 37:33wife and the kids, so I remember.
- 37:35One test subject, yeah. So in that three to four year, I
- 37:38guess, like willingness period. What do you want to call it?
- 37:40What you said? So I remember we were like, OK,
- 37:43it's now working pretty well on like adults, like let's try it
- 37:46on kids. So then I went to one of our
- 37:48investors. He's got a tennis court.
- 37:49So we went to his tennis court outdoors and we just set up some
- 37:52cameras. I think it's, I think it's go
- 37:53pros. And then and then we just got
- 37:55my, my son to just like run around.
- 37:57And then we so we motion captured it.
- 37:59The results came out really good and we put him on a like a
- 38:01bipedal robot, like just to retarget.
- 38:05And then we were super impressed because like, now he's working
- 38:07on kids. And I was really excited to show
- 38:09it to Jacob, who I think was five years old at the time.
- 38:11So like, pretty, I'm not that tall.
- 38:13And I showed it to him and he was like, OK, like.
- 38:18Not in the amount of the amount of passion.
- 38:21But then at the time, he's really into dinosaurs.
- 38:23He's like, he's like, I want to be a dinosaur.
- 38:25And then we retargeted it and we created a pipeline to retarget
- 38:28to, I think it was like a mini T Rex or something.
- 38:31And he looked at that and that blew his mind because now he
- 38:34could, like, mirror himself and see himself as a dinosaur.
- 38:38Like he's really into dinosaurs. He wishes he could.
- 38:41Like he probably imagines he could be a dinosaur like in his
- 38:43dreams or during the day and now and now you can kind of bring
- 38:47that to life and and have fun. Yeah, I feel like for kids, that
- 38:50would be very, very interesting if you can turn them into what
- 38:53they want to be turned into. Yeah.
- 38:54So the robot was not impressive enough for.
- 38:57I don't want to be a robot, like, like dinosaurs are cool,
- 39:00yeah. That's insane.
- 39:01So you had an idea that it would work on kids at the time or no?
- 39:05You, we just never tested it and we thought, OK, we've optimized
- 39:08it for adults because that's one of the main use case.
- 39:11And then it it works for kids. And right now we discuss this
- 39:15briefly before the podcast. How's the user base growing for
- 39:18Move? Yeah, it's growing
- 39:20significantly. So we've been using more and
- 39:21more high end project as well as on a YouTube creators you get
- 39:24like 10s of millions of views per day with just kind of simple
- 39:273D explainers for example. And so, yeah, we've been used
- 39:31in, you know, high end on a AAA projects for film and games.
- 39:35So for example EA use US for games, Ubisoft for Just Dance.
- 39:39Which games in EA is it? FIFA, yeah, I can't mention.
- 39:43Oh, exactly Which ones? Yeah.
- 39:44It should be FIFA. Well, it's not called FIFA
- 39:47anymore. It's called FC, yeah.
- 39:48They change it, yeah. But yeah, I, I, I kind of deep
- 39:51down still call it FIFA, yeah. It should be FIFA.
- 39:54For so many years. For so many.
- 39:56Years, yeah. Yeah.
- 39:57So yeah, Just Dance, EA games. What else is there?
- 40:02So like music videos like MGMT music video, Grimes?
- 40:06A Coachella concert, the boys Shogun.
- 40:11So like a lot of, yeah, a lot of a lot of kind of high profile
- 40:14projects you could say. So for this project, for
- 40:17example, Grimes for Coachella, do does the team come to you
- 40:20guys and get help or get support or they're using a movie on
- 40:24their own and you just find out later on that they use move?
- 40:27In a lot of cases, we try to design the tool to be
- 40:30self-service possible, but we of course provide support for
- 40:33anyone who needs support. I think in that case we did
- 40:36provide some support because they were using our newly newly
- 40:39developed real time system. But then there was also another
- 40:42partners partner agency or studio in LA who helped kind of
- 40:47with the logistics of a studio and everything.
- 40:49So yeah, we'll partner with different people to to help kind
- 40:52of bring these use cases to life.
- 40:54One of the coolest things I saw was a post from you with the Ray
- 40:58Ban Meta glasses and you're recording somebody else and then
- 41:02you took that footage, managed to get the digital motion
- 41:06essentially out of the person. I didn't know.
- 41:08I mean, it makes a lot of sense because there's a camera on the
- 41:11rabbit Meta glasses and naturally you should be able to
- 41:15get it. But when you did it and I saw
- 41:16the video with, I was like, holy shit, that's insane.
- 41:18So I can just be walking without my phone and I can see a cool
- 41:22motion. Of course, get permission from
- 41:24the person and then record them and then be able to get any
- 41:27motion that I want. That is for me the accessibility
- 41:31that you guys offer. How do you see the future of
- 41:34this though? Not the future move on its own,
- 41:36but the future of spatial computing.
- 41:38You had a post about this on LinkedIn.
- 41:40We now do see Apple moving into the VR world.
- 41:44Of course the form factor of these VR headsets are going to
- 41:47get smaller and smaller. How do you see moves role in the
- 41:51future of digital animation and these glasses spatial computing?
- 41:57Yeah. So as you said, the, the, the
- 41:58rebounds, they've got a camera on them, right?
- 42:00So we're getting a proliferation of cameras in the world.
- 42:03Like, you know, you know, a Tesla has a lot of cameras.
- 42:06Like there's a time where cars didn't have that many cameras on
- 42:09them. I think vacuum cleaners have
- 42:12cameras on them. Like I'm sure lawn mowers, if
- 42:14they don't already, will have cameras on them.
- 42:16You start to get more because they need to perceive what's
- 42:19happening in the real world to make the product better.
- 42:22And so we're fundamentally working on trying to understand
- 42:25the digitize how living systems in general work.
- 42:27We're right now really focused on how humans work because
- 42:30that's that's a pretty chunky and a challenge in itself.
- 42:33And so the more kind of analytics understanding we can
- 42:37bring to like we can create, whether it's running for 3D
- 42:42experience or whether it's helping to perceive the real
- 42:44world, that's what we're focused on.
- 42:46And as as cameras proliferate, it just, it's just you get more
- 42:49sensors available in the real world.
- 42:52I mean, maybe in 5-10 years time or, or next year, like everyone
- 42:56will be walking around with a camera, like facing forward
- 43:00every, you know, machine on most machines, a lot of machines in
- 43:04warehouses already have cameras on them.
- 43:06So, so, so imagine every car has a has multiple cameras, you
- 43:10know, you have multiple cameras that are online or not online.
- 43:15So there are more sensors available.
- 43:17And the more you can perceive the real world, the better you
- 43:20can make a lot of products. And so I definitely see Move
- 43:22helping, yeah, helping with that perception, helping to
- 43:25understand what's going on in terms of kind of the 3D motion
- 43:29animation. So, yeah, I don't know like
- 43:33which areas will take off the most, but you absolutely can see
- 43:36in a few years to five to 10 years how kind of everything
- 43:40around the world will be perceived.
- 43:43You know, satellites are being put up at a faster rate than
- 43:45ever before. Yeah.
- 43:46So there are cameras looking down, I guess not at other Yeah,
- 43:50because yeah, the the curtains are drawn I guess, but yeah.
- 43:53We're good for now. We're.
- 43:53Safe. Yeah, but you step outside and
- 43:55there's there's so many satellites who that can kind of
- 43:58see you, whether it's an RGB image, it's a it's a kind of a
- 44:03infrared image. So yeah, like this is this.
- 44:07The world just seems to be going this way and, you know, we just
- 44:09want to help to perceive what's happening in the real world, to
- 44:11build, you know, important applications.
- 44:13Sounds scary man. We're being watched everywhere
- 44:16but. Your guys focus on human motion
- 44:19capture. I correct me if I'm wrong,
- 44:20initially you guys didn't have the finger tracking when you
- 44:23guys started and then I think you added the finger and hand
- 44:26tracking. Is there any plan to add facial
- 44:28motion capture as well? Yeah, potentially.
- 44:31I think there's a lot of other good solutions already out there
- 44:33for for face. So there's no point reinventing
- 44:36the wheel. So we may integrate or we may
- 44:38build in our own solution. But yeah, it's an important part
- 44:42of yeah, of like perceiving the real world for, for, for, for
- 44:46performance capture. What about object tracking?
- 44:49Yeah, we already do some of that.
- 44:50So one of the most common objects we see that our
- 44:53customers have is tracking. Tracking balls.
- 44:55So like soccer balls and and. Guns maybe?
- 44:58Swords. Yeah, that's pretty.
- 45:00That's pretty. Fun.
- 45:01Yeah. We gotta do Shogun.
- 45:01Yeah. Star Wars Shogun.
- 45:03Yeah, I'm, I'm very curious. You guys are already doing that
- 45:07internally like doing these tests, is it?
- 45:08Part of the focus. Yeah, yeah.
- 45:10So, so, yeah, we're already, we do a lot of, we have a lot of
- 45:13sports use cases. So we do track like balls and
- 45:16and they're all kind of the same shape.
- 45:19Yeah. And they're quite, yeah, like
- 45:20this, the same kind of degrees of freedom.
- 45:23Yeah. In the road map, we do have the
- 45:24ability to track more complex objects.
- 45:27So yeah, that is coming. But yeah, it's it's not
- 45:30something we have. Swords are not considered
- 45:31complex objects. I'm just curious about the road
- 45:34map. Yeah, Yeah, it's good questions.
- 45:35So swords, I guess they're not that complex, but you know,
- 45:39swords have different lengths, different.
- 45:42Yeah, it's just like different attributes.
- 45:44And so because in many cases, if we used in like a in a film
- 45:50shoot or video game shoot, somebody might just decide to
- 45:53bring on a, a totally new object.
- 45:56Yes, yeah. And so we want to make sure that
- 45:58the product can handle those situations.
- 46:00So whether it's like putting a sticker on the object so that
- 46:03you can just, it can just run like immediately that you can
- 46:05start tracking just like a few a few objects, or it's creating an
- 46:09AI model that can do it totally Mark less like we're
- 46:11experimenting with the best kind of user experience.
- 46:14Well, that's definitely a very tough task for your engineering
- 46:19team to handle. I can see because there's so
- 46:21many possibilities and you have to choose and create a road map
- 46:26that will be best suitable for the mass market.
- 46:28So of course swords is interesting to us, but maybe the
- 46:32the majority of your market doesn't actually want to use
- 46:34swords like we know we want to use it.
- 46:36What about API integrations? I know you guys are working on
- 46:39that as well. What is the most famous ones
- 46:41that you guys are doing and is there anything that you can
- 46:43share about the road map of? We got a few really interesting.
- 46:47So we launched our API last year and so we got a few interesting
- 46:50people building on the API anywhere from consumer apps to
- 46:53industrial applications. So I guess 1 interesting use
- 46:57case is there's a company called Winning Edge.
- 47:00Winning Edge. So it's a really interesting
- 47:02company working on a really interesting use case.
- 47:04So it's a it's a sports use case.
- 47:07So it's it's former Olympians who really want to help their,
- 47:11you know, people like them achieve elite performance and
- 47:15being able to have an easy to use tool that can capture motion
- 47:17to help with the analytics, help with the training for the
- 47:21athletes. So that's a really interesting
- 47:22use case whereby elites, yeah, elite performers can now start
- 47:26to have analytics to improve their performance.
- 47:28We've got people also integrating it for consumer kind
- 47:31of fitness as well. So your.
- 47:33Original idea. Yeah, the original idea.
- 47:35That's right. Yeah.
- 47:36So they're using it on an app now, like a mobile app or?
- 47:38Yeah, so we've got a Swift SDK, we've got a Python SDK, right?
- 47:42So it's really easy to integrate.
- 47:43All you need to do is just give it video.
- 47:45And you can also, if you also give it other information like
- 47:47the orientation of the camera, then our, our core model can
- 47:50then use that to get even high quality data out there from it.
- 47:54But then also we've got industrial applications where
- 47:56people need to understand how people are moving in warehouses
- 47:59for safety reasons. There's a lot of warehouses that
- 48:02have people doing very repetitive tasks and, you know,
- 48:06lifting heavy things, awkward, awkward motions.
- 48:08You can get injuries, like very serious injuries.
- 48:11So being able to assess like or predict when somebody a
- 48:15particular motion like could injure somebody or then be able
- 48:19to monitor if it's like they're already injured or hurt or
- 48:21something. So so putting the sensing into
- 48:24warehouses is also an important. Place this is very interesting.
- 48:27For example, for the safety of a factory, when they upload the
- 48:29video, is it a real time video that's kept being fed to your
- 48:33server? And then this signal will come
- 48:35out that for example person A or this person is doing the wrong
- 48:38move? Is it a real time interface?
- 48:40Yeah. So experimenting with there's a
- 48:41lot of different warehouses are quite complex and so in some
- 48:45cases maybe it's a moving rig, in some cases it's a real time
- 48:49feed. So we're experimenting with the
- 48:51best ways to to deploy it. OK, yeah.
- 48:55So like it could be real time, but then also compute is
- 48:58relatively expensive. I say relatively, but then if
- 49:01you if you process it on edge with local hardware, then it's,
- 49:04you know, it's far less expensive.
- 49:06So we're different customers experimenting with different use
- 49:09cases. So those are still emerging and
- 49:10hopefully next time we catch up, I can tell you more about those,
- 49:13OK. How big is the company at the
- 49:15moment in terms of number of employees?
- 49:17About 2025 people. 2025. People based in the UK.
- 49:21UK in Europe, OK. Mostly.
- 49:23So when you started this in your living room just trying to
- 49:27create a sports utility application for yourself, you
- 49:31were not thinking of becoming an entrepreneur.
- 49:34Because when you become an entrepreneur, now you have to
- 49:36lead people. You have to unite people around
- 49:39a mission. Did you have any idea when you
- 49:42were in university that you want to actually create teams and
- 49:44guide people towards a mission? Or were you like a lone wolf?
- 49:48When I was at university, I I experimented with different
- 49:51entrepreneurial things like I like I was consulting in
- 49:54finance, which is totally, yeah. Yeah, that's why it's it's quite
- 49:58interesting. Why were you doing that at the
- 50:00side? This is just a passion for you.
- 50:03Fractures. And finance?
- 50:05I thought it might be a passion and I was definitely interested.
- 50:08So for example, yeah, during university I was consulting in,
- 50:12yeah, in finance. So I was part of a small team.
- 50:14We managed liquid asset portfolio for a big UK retail
- 50:18bank, like a £50 billion liquid asset portfolio.
- 50:21So like bonds, mostly government bonds.
- 50:24And then we did all the all the interest rate risk hedging you
- 50:29guys are looking at me. Like where did you?
- 50:31No, because we starve in material science fractures.
- 50:35Then we went to motion capturing video games and now we're going
- 50:38in the middle. And then wow, we're.
- 50:40Discovering I'm glad I can talk about something you don't know
- 50:42because you. Guys know way more.
- 50:44About me than VFX, but yeah, yeah, it's finance, liquid asset
- 50:48portfolio, right, interest rate risk swaps, FX risk hedging and
- 50:54then restructuring. So I, I was just really
- 50:55interested in that area. I was doing my PhD.
- 50:58So actually I went through this period of my life where it's for
- 51:01about 2-3 years where I do my PhD full time.
- 51:05I was also spending like 4 hours a day like at this trading desk,
- 51:11like doing this hedging, hedging stuff.
- 51:14And then I also did my first start up like in Parallels.
- 51:16I I like this two years of I call it the two years of no
- 51:19sleep. Right.
- 51:20Of course, yes, the start up was related to what we're talking
- 51:23about now or that was no, no sort.
- 51:25Of that was like events, I was like creating events for
- 51:28students. Yeah.
- 51:30So trying different things, but one thing that I think really
- 51:34probably helps you in the beginning is coming to the VFX
- 51:37or cinema industry without having the pre assumption or pre
- 51:41knowledge as you may because you came from a totally different
- 51:44industry. Do you think it was a helpful
- 51:47gift that you came into this not knowing the assumption, not
- 51:50knowing about the motion capture suits and then come up with a
- 51:53totally different idea, not going to improve the suit, but
- 51:56you know. Yeah, I don't know.
- 51:58So I definitely came into it naively and not knowing it.
- 52:02And I think other other entrepreneurs and investors,
- 52:07they say it's good to come in naive like this in many cases,
- 52:10but also the learning curve is super high.
- 52:12Yeah. OK.
- 52:13So, so I think in some cases, like if you know the industry
- 52:15and you can have a fresh kind of fresh mind, then then I think
- 52:19you can kind of not have these underlying premises and
- 52:22assumptions and kind of break them and create a new model.
- 52:25So, so it's hard to kind of make the case for the general, but I
- 52:29think for me, I guess in hindsight, it has helped.
- 52:33Like I knew nothing about mocap. Like I didn't even know the word
- 52:37when we first got into it. Like, I didn't know anything
- 52:39about VFX. I'd barely created anything in
- 52:423D. So yeah, like all I'd done in 3D
- 52:46was like research and then had to learn like about this VFX
- 52:49industry and and, you know, 3D content creation.
- 52:52That's fascinating because you have it from the outside, it
- 52:56looks like you know a lot in the beginning.
- 52:58It looks like you've known about this industry for the longest
- 53:01time and you're coming in to solve a problem, but you were
- 53:04solving a completely different problem and it looks like
- 53:06there's this whole industry that could have used your solution
- 53:10and now you have to learn about it.
- 53:11I was. Just trying to work out, you
- 53:12know? That's what I'm saying.
- 53:14That's what I'm still trying to do.
- 53:16Do you have time now to work out less I'm?
- 53:19Trying to, yeah, yeah. You completely drop that sports
- 53:22application idea by the way, right?
- 53:24I, I just, I hack away on it like every now and then, but
- 53:26yeah. So the reason why I also asked
- 53:28you about the number of employees because I want to know
- 53:31how tough it is for you to manage 2526 people and are you
- 53:36growing this fast? When you're growing this fast,
- 53:39how do you learn to deal with these different teams that are
- 53:42now being set up 'cause you have the engineering team, but now
- 53:44you have to have a support team because you have this
- 53:46application. Well, technically you have two
- 53:48applications now you have move one and then you have the we
- 53:50have. Move we have move one which is a
- 53:52single camera. We have Move Pro which is the
- 53:55multi camera post processing solution and we have Move Live
- 53:58which is the real time solution. Yes.
- 54:00So now you have support for those applications.
- 54:02So all of these different teams that are being set up, this is
- 54:05an area where you didn't expect to get into, but now you have to
- 54:08deal with. You guys are doing a great job
- 54:10at it. I just want to know where did
- 54:12you learn about all of this? Did it come through a lot of
- 54:14trial and failure? Did it come through mentors?
- 54:17Did it come through a lot of book reading or an experience
- 54:19from your early start up? Yes, it's definitely a lot of
- 54:22trial and error, right? I'm not like I didn't come from
- 54:25management or like Business School.
- 54:28So definitely a lot of trial and error, a lot of reading, a lot
- 54:31of mentors. I think the more you can have
- 54:34those mentors, like having a coach who can help you with all
- 54:38these things, it definitely can help you learn faster.
- 54:41I think always people are learning and failing in the
- 54:43early days. You learn and fail like much
- 54:45faster. And you want, you want to, but I
- 54:46think it's a good thing you want to, you want to keep up that
- 54:48quick iteration process to keep learning.
- 54:50But yeah, mentors have really helped a lot of reading and
- 54:53research. And then you just have to put
- 54:55into practice and then just kind of review and and learn.
- 54:58I love that. Do you have any advice for any
- 55:02aspiring entrepreneurs who are watching?
- 55:05Perhaps from one of your mistakes that you made, one of
- 55:07the the trials and errors and fixes that you came up with?
- 55:10Do you have any advices that you would give?
- 55:13Yeah, I guess like as we've been talking, I think the the passion
- 55:16side of it is really important for an entrepreneur because the
- 55:18journey is, is really tough. Like it's really difficult.
- 55:21You have to make a lot of sacrifices.
- 55:24And you know, it's very easy to give up and in some cases you
- 55:27should give up. So I'm also not not saying like
- 55:29absolutely be an entrepreneur, like absolutely keep going, even
- 55:33though it looks like it's not going to work out, like you need
- 55:36a lot of wisdom, a lot of judgement and you have to work
- 55:39out. It's a sacrifice worth it, like
- 55:41sacrificing, you know, social and family and things.
- 55:44So you definitely want to, you want to go into it carefully,
- 55:46but I think being driven by the passion is extremely important,
- 55:50just like you guys with, with what you're doing, because you
- 55:53come across obstacles and, and in many cases like no one,
- 55:57sometimes people even even turn their back on you and, and you
- 55:59just have to keep going yourself.
- 56:00And if you're driven by the passion and the vision that is,
- 56:03that's what helps you keep you going.
- 56:05So another tip I guess is to work out because.
- 56:09You might come up with a business idea.
- 56:11If you work out, you might just come up with the right business
- 56:13side. Exactly like a moral of the
- 56:15story is like exercise, Yeah. I mean, it's, it's not only
- 56:18helpful for your physical health, but it helps you create
- 56:20businesses, yes. Yeah, in this case, but not it.
- 56:25Didn't help us. Make any businesses.
- 56:26Not yet, Not yet. Maybe we can come up with that
- 56:29idea for business while we're working out.
- 56:31Do you do you work out now or? Yeah, yeah, yeah.
- 56:34Gym or at home living room still.
- 56:36Mostly at home just because I have very little time available.
- 56:40Like I have a very small home gym, but it's like not that well
- 56:42equipped and so I just try to get as much as I can in a
- 56:46shorter space. So when you have now move AI
- 56:50that has 25 people, what is your core focus now in the company
- 56:53because as the CEO is it leading and management finances,
- 56:57marketing research? You always have to do a bit of
- 57:01everything, especially the start up, but you know, there's
- 57:03different bottlenecks in companies and so sometimes you
- 57:06have to dive into kind of sales and marketing, sometimes
- 57:10engineering. So the early days was just
- 57:12research and engineering. Unfortunately that's kind of one
- 57:16of my strengths and then, but now it's kind of really growing
- 57:20the business, you know, increasing the support material,
- 57:23showing that these use cases just to help bring to life for
- 57:26different industries like how move can be used.
- 57:28And so it really does vary. So sometimes it varies week by
- 57:31week, sometimes month, month, month, month by month.
- 57:34Which one is your favorite if you could pick and if you could
- 57:36hire other people to do the other jobs?
- 57:39If you wanted to focus, which task would be your favorite?
- 57:42Yeah, I think at heart I'm an engineer, OK.
- 57:45So I love building things, especially like with my own
- 57:48hands, with my own code, like keyboard, I guess master
- 57:51keyboard, maybe headset one day, who knows?
- 57:55But but also I just love the vision of where it's going.
- 57:58So I think if you're led by the vision, then you're willing to
- 58:00do whatever stuff it takes, even if it's not your core
- 58:04competency, even if it's not your interests, like you would
- 58:06just do it because you're you're led by the vision.
- 58:09You know, I just want to say first of all, thank you so much
- 58:13for your time. It's been an honor just
- 58:15listening to the story because one, I find this story truly
- 58:18fascinating. Again, I did make fun of the
- 58:20fact that you should work out if you want to come up with ideas.
- 58:23But generally, we understand the pain of going through building a
- 58:27new idea, especially when you're trying to disrupt the market in
- 58:30a sense that you're coming up with an all new solution in a
- 58:33market that already has solutions, but people don't know
- 58:36that they're outdated or they can be much better.
- 58:39And what you guys have created has been truly useful to the Bad
- 58:42Decisions team. We've used it in our work and
- 58:44we're continue, we're continuing to use it as well.
- 58:48And so I'm looking forward to see the new technologies, the
- 58:52new applications that you guys will have over the next few
- 58:55months and years. And by the time we have this
- 58:57podcast again, there's one thing I want to try is movie eye real
- 59:01time. That's the one thing we have not
- 59:03tried yet. Yeah.
- 59:05And that's why I want to ask you, is that available for
- 59:07public? How can people get access to
- 59:09that? Yeah, it's available for the
- 59:10public. Just go on the website, book a
- 59:12call with the sales team and it will help you get set up with
- 59:14getting the right camera equipment and then just making
- 59:17sure you've got a 3D workflow. So maybe you should do that.
- 59:20We should do the podcast in 3D. Oh, and then next time, Oh yes,
- 59:24actually, then we can put 4 cameras.
- 59:25That would be insane. Yes, that would be insane.
- 59:28I would love to set that up. That would be crazy.
- 59:30Would that be the world's first? I don't know.
- 59:31I'm sure other people have done something.
- 59:33Like they've done, they've done mocap, but they're not real time
- 59:36without any suits. So we had a friend, James Afina,
- 59:40so he did a podcast, but he did it with Vicon system and meta
- 59:45humans. But then they had to have, you
- 59:47know, suits and the the markers on, but then without anything
- 59:51on. But let me be clear, with
- 59:57clothes on, no markers, No, be so careful.
- 1:00:01I don't know. What's on your mind?
- 1:00:04Yeah, please don't look at your training data from my account,
- 1:00:06the bad decisions account. Let me know when you want to do
- 1:00:10that and I won't be here. I'm not going to talk anymore,
- 1:00:14Farah. You do that outro.
- 1:00:16No, but we really want to thank you for your time.
- 1:00:18It's been great talking to you. Thank you.
- 1:00:20Thanks guys. Appreciate it.
- 1:00:21Thank you.