2. Automatically turn 2D into 3D art. Interview with Konstantina Psoma from Kaedim.

TRANSCRIPT

[00:00:00] Konstantina Psoma (Kaedim): Everything is gonna start moving from 2D to 3d. Everything is becoming more and more immersive.

I think there is a, a, a certain phobia against AI in a lot of sectors. There's no intention to kinda like replace artists.

The goal is to augment and to accelerate them.

If these tools are used correctly and ethically, the speed up and acceleration they can provide to artists is huge.

Because the true creation is like in the detail. It's where human nature is like thrives.

[00:00:55] Mizter Rad: Welcome Konstantina. Thank you so much for being here. You're joining us today from London. You're the founder of Kaedim. I hope I'm pronouncing that correctly. And Kaedim is building an AI machine to transform 2D into 3D assets. 2D Objects into 3D assets faster than usual. 10 times to 20 times faster compared to manual work, which is the usual way of solving this problem at the moment.

Is that correct?

[00:01:25] Konstantina Psoma (Kaedim): Yes. Thanks. Thanks for the intro. yeah, that's me. That's what we do. And hi everyone. Thanks a lot for really listening.

[00:01:32] Mizter Rad: So how do I pronounce correctly? I just wanna make sure I have that right.

[00:01:37] Konstantina Psoma (Kaedim): It's Kaedim

[00:01:38] Mizter Rad: Kaedim

[00:01:39] Konstantina Psoma (Kaedim): K dimensions.

[00:01:40] Mizter Rad: Okay. All right, cool. So maybe you tell us a bit more about yourself to start with that. Konstantina you study computer science and then you did a masters in deep learning as far as I know.

[00:01:52] Konstantina Psoma (Kaedim): Yeah. Yeah. So yeah, I come from a computer science background. When I was at Uni Bristol, I was doing computer science with innovation and one thing led to another. I started doing a lot of 3D units.

So I started so I took up like 3D modeling. Character and set design. 3D animation and learned the hard way like firsthand how hard that is. And so there was one project where we had to take a physical place from Bristol and turn it into digital 3d. And I wanted to do a cathedral, one of the cathedrals in Bristol.

And I still regret that decision. It took like months to be able to finish that project and I didn't even get a good grade. I was like super, super pissed off. Anyway, so that led kinda like starting some discussions with like game devs and like animators in the space. Like locally in Bristol as well to kind like start asking them.

Hey guys, how do you do this in scale? It's super hard. Maybe I'm just not good at it, but if you have a trick, please let me know. So they say there's no trick. And they show me like a room full of 50 3D artists doing their thing on the same software I was using Autodesk.

And just making everything from scratch by hand. And this is when I was like: oh my God. There's a, there is a problem here, right? Because everything is gonna start moving from 2D to 3d inevitably. Everything is becoming more and more immersive. And and so this is when I was like, I really like this problem.

I felt it firsthand. I want to look into it. So then I did my masters in deep learning for 2d, two 3D construction. And started building some likes, prototypes testing them with some, uh, local partners. And that's how things got started.

[00:03:44] Mizter Rad: That's an interesting story. I mean, so for people that are not very used to the lingo that we're starting to use here. Basically, what Konstantina and her team are building could be compared to something like back in the days when a house was gonna be built, the bricks were made by hand, by people.

And that process took a long time and it was very expensive and tedious, until someone created a machine or a process that was more efficient. And now the bricks were made on scale and can be, could be reproduced in many, many numbers in a very short time. Also, decreasing the cost of production.

And you are doing this in the digital space. So people with this new internet that is coming up that is immersive, that is 3D. Designers and artists and architects have to do these 3D objects by hand. And that takes a long time and it's expensive. Right. So what you are trying to do here is to create this machine, this new process, this algorithm that does this, to put it simply, in simple steps and with a simple click, even though I believe you're not there yet, but that's the aspiration.

Is that correct?

[00:04:58] Konstantina Psoma (Kaedim): Yeah, very much so. Essentially we, there's no intention and we don't want to kinda like replace artists. What the goal of Kaedim is, is to augment and to accelerate them. And kind of like eliminate as much the crunch involved in the industry. While building very advanced algorithms to do that.

[00:05:19] Mizter Rad: It's interesting that you touched that topic because while I was inviting a lot of my artists friends and, and my network, some of the artists approached me and said, look, um, thank you for inviting me, but I'm not interested in this AI thing because that's gonna replace human. And it's gonna take away jobs from us, from humans. So what do you think about this in general ?

[00:05:41] Konstantina Psoma (Kaedim): Yeah, I mean, I think there is a, a, a certain phobia against AI in a lot of sectors. Like the same thing could have been said about Mid journey or Dali, right?

However, I think if these tools are used correctly and ethically, the speed up and acceleration they can provide to artists is huge.

And what we are all about at Kaedim is like, why would you start from scratch every single time, right? Instead of making 20 or even a hundred times, in like a year or two from now, the amount of creative material that you can today, right?

Because the true creation is like in the detail. It's where human nature is like thrives.

Having to make the basic shape every single time, will like almost exhaust ... exhaust creativity and so we need, we want to give just more space to people to be able to do more and to be able to focus on the things that they really enjoy doing.

[00:06:39] Mizter Rad: And I also believe that whether we like it or not, and I know sometimes this new internet or future technologies are controversial because people think we are gonna run out of jobs or, uh, the machine is gonna take over and so on.

But, uh, in a way it's, it is a stronger force than than us individually. So there's, whether we like it or not, not much to do about it. What we could do is embrace it in a way and, like you said, look at the positive sides of it. And, and, and we always transforming. I remember some years ago architects would be afraid of machines or the computer coming in and replacing them.

You know, back then they would like draw the sketches on paper. And now with all the softwares, you can do this pretty much in a very short time. So back then they were already afraid that the computer would, um, replace their jobs. But what happened is that they transformed and they became something else.

And maybe, yeah, they pay, they pay off better. And, uh, it's more efficient. Everything that, that, that comes around that.

[00:07:43] Konstantina Psoma (Kaedim): Yeah, absolutely. I think, you know, keeping an open mind and obviously the company's being, um, honest on like, I don't think AI can replace humans. And I don't think there is any alternative for like human creativity to be honest.

And, and yeah, I'm the founder of Kaedim and essentially like what we, what we want to do. Yeah, just repeating that again, is just to like accelerate and allow people to do more than they can today.

[00:08:12] Mizter Rad: I see what you mean. Maybe you can take us a bit around the pipeline that you guys built at Kaedim.

[00:08:18] Konstantina Psoma (Kaedim): Yeah, absolutely. So I'll take you a bit through the user flow of, uh, how people are using our tech and how it works for them. And essentially, like we have a web app where people can log in and, um, go in and submit images of what they want to convert into 3d. And this can be anything like a sketch, a piece of concept art or photos of, of real things that you have around you. And then we have developed a couple of options and settings that people can play with in order to get kind of like the, as ideal output as possible. Um, so essentially you can play with the poly count limit, because we work anywhere from like mobile to PC and AAA and essentially different people have different requirements as to what their outputs are. So there are a couple of settings that you can play with, but after that it's just a matter of pressing a button, and, let the software process. And so when, when your asset is ready, you can just look at it in our 3D viewer and then download it and continue your work flow.

So what we see people doing is they go online. They upload their images. They look at the output. They download it. And then they'll import it to their software of choice to continue, either adding more detail or, going to the next steps of the pipeline, like texturing, rigging, et cetera.

[00:09:41] Mizter Rad: While I was, looking at a document that you wrote recently, Konstantina, you were, talking about the pipeline, that that's what we just explained. But, at the moment you have, people in the team that do, that do the, uh, like a checkup. Is that correct? Like, you, like someone uploads the 2D image. Then there's an output that, most of the times is great, but sometimes it needs some retouches and those retouches, quality control sort to say, is done by a human at the moment.

Is that right?

[00:10:14] Konstantina Psoma (Kaedim): Yes, correct. Yes. We do have, manual quality control involved in our processes, which is vital for our algorithm development and like always progressing, uh, because essentially we create feedback loops. Based on what the algorithm comes up with and how we cleaned up to be ready for our partners.

So yeah, definitely that's part of the process. And the goal from our side is to build tools that, you know, both, do the cleanup faster and more efficiently, but also, on the other side of the coin, is making the algorithms more and more efficient.

[00:10:48] Mizter Rad: Yeah, and I guess that's super important for a, a high tech, deep tech company like yours, because you need that feedback loop that at first, has to be manually controlled and sort of the machine needs to be fed right, so that your machine actually learns how to react, in itself in the near future.

[00:11:10] Konstantina Psoma (Kaedim): Yes, yes. It's, uh, it's, it's not as simple as, it's, as you make it sound, but I guess that's a, a good explanation.

[00:11:19] Mizter Rad: Okay. No, of course. I, I, I'm sure it's very complex, but I'm just trying to make it simple here so everyone here in the space understands.

[00:11:27] Konstantina Psoma (Kaedim): Yes.

[00:11:28] Mizter Rad: All right. Right. That sounds good. Tell me something, because I understand that, from a startup owner, and entrepreneur perspective, I understand that building a company has a lot of, sides to it and a lot of faces to it. You come from a very technical background. You're very young. You don't have, you know, 10 20, 30 years of experience building companies. Maybe this is your first venture.

What has been the most complex part so far, in your case, when building, Kaedim?

[00:11:59] Konstantina Psoma (Kaedim): Oh God, where do I start? Um, , I mean, look, like, build, like, especially for first time founders the challenges are endless and the amount of learning that you go through is huge.

I guess, for me, my first challenge, which, I faced like fresh out of uni when I wanted to go straight into Kaedim was getting to the right mindset. Right? And what do I mean by that? I think you need to be in a certain mindset to be able to go through everything that you need to go through as a startup founder.

And that's because essentially startups can be compared to war. It's a very, very hard environment. And you need to give like 200% of yourself. And it's super hard work. And for me personally, I had very few examples back then. I had very few examples to look up to.

[00:12:54] Mizter Rad: Yeah.

[00:12:54] Konstantina Psoma (Kaedim): So getting into the right mindset of like, we need to build something fast. We need to build something... we need to build the right team. And all of that. And it was very, very hard and it took me time.

And then the second main challenge was building the team, right? Because building the team, um, has some requirements. And the most important of them are realizing who you are. So realizing what your values and culture is as a person and what you want to start, you know, cultivating into the people who come in.

And this is hard because for a young person, you know, people are still like exploring what they stand for. Like what is the way of work. What is their ambition and you know, it's a process.

[00:13:39] Mizter Rad: Right. Absolutely. I agree with you that finding the right team is probably one of the hardest, hardest task as an entrepreneur, as a founder, and not only finding them, but that also, you know, keeping them motivated and making sure they are happy, but they're also working as much as, as they can. So that's, uh, I, I feel you. I definitely feel you. And I also understand when you say that, you know, when you're young, a lot of times you don't know where you're standing or what you want. And, and so it's through experiences and through the years that you come to, uh, sort of molding that up, even though I think it's a, it's a task that never ends.

[00:14:17] Konstantina Psoma (Kaedim): yeah, yeah, yeah.

[00:14:18] Mizter Rad: Finding yourself. But, um,

[00:14:20] Konstantina Psoma (Kaedim): I think as long as you have a north start of, um, one: what is the problem you're solving?

I think for a startup, having the problem as the king of everything is the most important thing. Because you can experiment, uh, especially at the early stage, you can experience, experiment with a lot of solutions, a lot of push for concepts.

You just iterate, iterate, iterate. Until something sticks, until something works. But having the North star the problem that you're solving, Right. And then kind of like, what is your ambition? Like, where do you want to take this? That's the most important thing, and that's something that has to stay constant.

And that's why I am like skeptical of like pivots. Pivots I think it's something very, um, dangerous to do.

[00:15:07] Mizter Rad: Well, in my point of view, pivots are, um, it depends how you do them. And, how drastic the pivot is. But, yeah, it's a good point of view. You talked about personal motivation and, why you doing this?

Maybe you can tell us a bit more, where do you see Kaedim going as a company, but also why do you think what you're doing is important for not just the gaming industry, gamers, creators, but also artists in general. Artists that today maybe are sketching on paper or sketching on a computer, but on 2D or painting a canvas or painting a wall.

And they don't have the skills or the knowledge to transform those creations into, into 3d assets.

[00:15:51] Konstantina Psoma (Kaedim): So about my personal motivation, there's a couple of things. So the first one is I felt the pain firsthand. Obviously, not as a professional 3D artist, but as a person who was trying to learn and trying to get into it.

Um, and so I realized how like, time consuming it is. How steep learning care the software is. Really, respect to everyone who, who is, uh, a professional 3D artist and has mastered that art. I think it's really hard. So that's partly, my motivation. I want to allow people to do more. And then the second thing is being an example.

So, as I said, when I was a kid growing up in Greece. There was like, I can't think of any example that I had around me of a successful entrepreneur, that, a successful founder that started, on a journey to make something big. Right?

I would like that to change. I would like to manage to be an example for both like, Greek young people, but also for anyone who wants to make something of their own.

And be like, oh yeah, if she was able to do it, then I'm also able to do it. And that's the mindset I want to like create.

[00:17:02] Mizter Rad: That sounds interesting. Especially, uh, you don't see many women unfortunately, in the industry yet. I think this is changing, but it's interesting. When I saw you, as a young female founder, I have to say, it caught my attention as well. Especially cuz this is very techy and, uh, yeah. Congrats for that.

[00:17:21] Konstantina Psoma (Kaedim): Thanks.

[00:17:22] Mizter Rad: So tell me why do you think artists should care about this? Why do you think this is important? And how you think Kaedim will help them in the future?

Right now, Kaedim, as far as I know, caters only b2b, like meaning your clients are mainly businesses.

[00:17:39] Konstantina Psoma (Kaedim): Yes.

[00:17:39] Mizter Rad: Is there a potential future where you open up the service to individuals or individual artists that can use your product...

[00:17:48] Konstantina Psoma (Kaedim): Definitely, definitely. Our vision as a company is essentially to make, to make it as easy as it is today to create like 2D experiences. To be able to create 3D ones.

And so, essentially what we want to open up is the ability for people who have zero 3D modeling knowledge to still be able to create something in 3d, right?

Kinda like something that will allow them to express themselves in 3d. And the idea is that, you know, everything is moving towards 3D experiences, immersive experiences, and having the ability for everyone to be able to contribute is gonna be very, very important.

Definitely we want to create something that is commercial, like a consumer focused, when, you know, when the tech reaches that point where we can have like super, super low cost, super fast, like real time assets ready, and yeah, open it up for everyone to use.

[00:18:44] Mizter Rad: Absolutely sounds like it sounds like a great idea because, like you said, the barrier of entry for these individuals artists that don't have the skills to create 3D models, is quite high at the moment. And I also have to agree with you that those that have the skills today, are sort of very fortunate position cuz not only they get a lot of job offers, I believe, but they also get to be pioneers in the space and create beautiful stuff that maybe can be sold for a lot of money at the moment.

[00:19:16] Konstantina Psoma (Kaedim): Yeah. But, uh, I think, um, things like games like Minecraft have proved that people like everyone around us, given simple tools, Right to be able to create, they can unleash a huge amount of creativity. Right? So that's where I'm going where I'm saying, you know, we open this up to everyone and see how rich of 3D worlds we have then. And opening up this new skill for everyone who was outside until then.

And still even in that case, I don't think that in any way replaces the professional 3D artists because there is a huge difference between like people who will use Kaedim to do like 3D user generated content, and the professionals who will be building the AAA games and the movies and the VFX.

[00:20:08] Mizter Rad: You talked about Minecraft. I, I met a friend of mine the other day. He has a 8 year old son and he told me that he built this game within Roblox and he has already, um, and he has no computer like software development skills. He just built it within the game. And he has around 400 players, active users at the moment, which is crazy.

Wow. And he's only eight, so, so yeah, definitely. It's, uh, it's fantastic what's happening in the, in the space.

If you think that the internet and the websites, in general will be immersive, do you, do you bet for, the day or for the future that comes, where people will be able to create digital 3D content , as easily as today we take photos in 2D from our phones?

Do you think this is, this is definitely happening soon?

[00:21:00] Konstantina Psoma (Kaedim): Yeah, I think if the technology advances with the rate that it is today, like this is coming and I think Kaedim can be a leader in all of that.

[00:21:09] Mizter Rad: That sounds great. I know we have some artists in the audience, um, maybe have some questions. If not, where can people reach out to you, Konstantina?

[00:21:22] Konstantina Psoma (Kaedim): There is a lot of ways to reach out to us. So like on our website we have contact tab where you can, reach out to us. Like people, I'm really happy for people to reach out to me personally as well.

Either on Twitter or LinkedIn. And or like any, any mean, they, they want really.

[00:21:43] Mizter Rad: I have actually Carolina, she wants to ask a question probably. She's already a speaker. Caro, do you wanna go ahead?

[00:21:51] Guest 1: Hi everyone. Thank you very much. It's really inspiring, even though I don't understand like half of the things sometimes. But I'm just curious maybe to ask you, um, in terms of technical support, like, uh, what you need, like in order.

I mean, because as far as I, in the past I was doing a lot of maja and I was doing a lot of 3d, and then you require a lot the, the, the technical from the computer. So right now, how would it be, would be like, so difficult, so expensive, or there would be like another tools in order to enjoy and, and, and, and create this beautiful, uh, opportunity?

[00:22:33] Konstantina Psoma (Kaedim): Hey, Carolina. Yeah. Um, so in terms of like technical tools that are needed. Yes. I mean, uh, compute power is one of the most important things and we're using AWS for that. There is so many tools that we've built in house as well to kind of like, do things that we couldn't find, like ready made. Always we look at the ready made tools to try to like use them, but sometimes then I, what we figured out is in the industry, actually the ready made tools are just not sophisticated enough for even artists to do their job.

Um, like to, to help them as much as possible. So it's a constant, um, It's, it's like part of our operations to be looking for tools that can help us and building actually our own tools, both for the quality control and for the algorithmic, like your construction that we are using. I'm not sure if that answers your question.

[00:23:30] Guest 1: Yeah, I think, I think, yeah, I think it's a matter of trying and doing and learning through, through the process. I think it's like very best just to try out and go through the abyss.

[00:23:43] Konstantina Psoma (Kaedim): Very much so.

[00:23:43] Guest 1: Thank you very much.

[00:23:45] Konstantina Psoma (Kaedim): Thank you.

[00:23:46] Mizter Rad: I think the "Golden Robot" wants to ask a question as well Konstantina.

[00:23:50] Konstantina Psoma (Kaedim): Yes, of course.

[00:23:51] Guest 2: Yeah. Hello. Hello Konstantina. It was really, really interesting to to hear about it, what you're doing with the 3D modeling. I just wanted to ask, is it possible to make also textures with your technology?

[00:24:06] Konstantina Psoma (Kaedim): That's a really good question. Everyone is asking it. Um, no, currently we don't have a, a, an AI way to apply textures in production.

We do do have prototypes for that kind of stuff. But, they take kinda like the second seat, in the first, uh, like after geometry reconstruction. It's something that most of our partners are asking us and wanting us to get into, but as a startup and kind of like early stage, focusing on one thing at a time is really important.

So definitely in our minds and in our backlogs, but currently not in production.

[00:24:42] Guest 2: Okay. Thank you very much. That's, that's very interesting. Yeah. Because like we are building, some assets for a virtual world and we are building a little virtual world by ourselves.

[00:24:55] Konstantina Psoma (Kaedim): Amazing.

[00:24:56] Guest 2: And, for us it's really, really, really important to get a lot of assets very cheap.

Yes. And it's a, it's a long time, uh, to, it takes a long time to, produce basically, uh, all the assets by hand. You know?

[00:25:12] Konstantina Psoma (Kaedim): Tell me about it.

[00:25:14] Guest 2: Yeah. . .

[00:25:15] Konstantina Psoma (Kaedim): Feel free to reach out to me and we can have a chat.

[00:25:18] Guest 2: Yeah. Maybe I contact you later to exchange some, some ideas or something like this, if it's okay.

[00:25:24] Konstantina Psoma (Kaedim): Of course. Feel free.

[00:25:25] Guest 2: Cool. Cool.

[00:25:27] Mizter Rad: I think we have "Vito" also, on the line. You wanna ask us something "Vito"?

[00:25:33] Guest 3: Hello? sorry for my delay, but, some problems with internet. Hello for everyone. Yeah the question I have is for Konstantina, I think his name Konstantina? Yeah. Yes. Um, So, uh, about, I don't listen, uh, all the conversation.

Only, only now. So you have some software that, you develop that you can turn, the image to 3d. Exactly.

[00:26:01] Konstantina Psoma (Kaedim): Correct.

[00:26:02] Guest 3: But so my question is like this, I love toy art. I do some toy art for some companies in Japan. But not 3d because I'm, I'm a 2D illustrator. So if I grab my 2D illustration, you can turn my illustration in 3d.

[00:26:23] Konstantina Psoma (Kaedim): Yes. That's the idea. Obviously there are like specific input guidelines. So I, I would have to see your illustrations there, or you can give it a go yourself and see if, the software works well for you.

[00:26:36] Guest 3: Okay. So, but I have to do, uh, for example, a, for doing something for a toy, I have to, to put like front, back, left side, right side, up, down, something like that.

[00:26:49] Konstantina Psoma (Kaedim): So, how we currently work with inputs is you can upload anywhere from one up to six images. The less images you provide, obviously the less detail is captured than in some, uh, like common cases. Uh, we will, like, the algorithm takes an educated guess of what's, uh, what's not pictured. But the more views you have, the better.

[00:27:14] Guest 3: Okay, perfect. Because it, it's a really great thing because take too much time to, to do something in 3D. I want to, to do, I have some idea for a new project, NFT project, and I want to turn my NFTs in 3d toy.

[00:27:30] Konstantina Psoma (Kaedim): That's amazing.

[00:27:32] Guest 3: Yeah.

[00:27:32] Konstantina Psoma (Kaedim): Feel free to reach out to me and I can, uh, have you talk.

[00:27:35] Guest 3: Yeah, yeah, yeah. I want to, to talk you later. Okay. But congrats for the project.

[00:27:41] Konstantina Psoma (Kaedim): Thank you so much.

[00:27:42] Mizter Rad: Well, if anyone has any other question, you're welcome to raise your hands or, request permission as a speaker. Otherwise, Konstantina. "Efcharistó". Thank you so much for, uh, taking the time to talk to us and to share what you're doing. I'm sure a lot of people got inspired and, curious about your process and about your future.

I know it's not easy as an entrepreneur sometimes to deal with, uh, so many things and so many problems. Sometimes we're running, putting down fires here and there and, it's not easy and sometimes you get criticism, for the things you do, but I know, and I'm sure you are doing the best.

You seem to be very passionate about solving this problem, and I applaud you for having the courage to shoot at a problem that is very hard to solve. And with your solution seems to be very scalable, so that's, that's great. I think Caro, Carolina has another question. Let me, let me give her, the word. Caro, go ahead.

[00:28:42] Guest 1: Sorry. Yeah. Thank you. The last question. When, would it be available, the technology to start and how much would be the cost? Like to get it or how does it would work... it would work?

[00:28:55] Konstantina Psoma (Kaedim): Good question. Essentially for a consumer offering, we're aiming, for the next like one to two years, to have something that's, like accessible, for like individuals to have a go at.

And yeah, there's no way to predict the exact cost, but you know, if it's consumer friendly, it's consumer friendly.

[00:29:14] Guest 1: Super. That sounds amazing. Thank you very much, much.

[00:29:18] Konstantina Psoma (Kaedim): But yeah, Rad, as you said, the journey is hard. Putting out fires every day is. It's another day in life. And, the only thing that matters and my only message to like fellow founders and people who want to do their own thing, is just keep going.

There's nothing else that you, that you need to do. And focus on your users. Focus on what you're building.

[00:29:37] Mizter Rad: Absolutely. Well, "Efcharistó".

Have a great day. Enjoy London and talk to you soon.

Thank you so much, beautiful people for being here. Until the next time. Thank you.

[00:29:47] Konstantina Psoma (Kaedim): Thank you everyone.

[00:29:48] Mizter Rad: Byebye.