Building the Future of AI Content

Speaker A: Hello world. This is Better Tech, a podcast where we chat with some of the most successful leaders about the latest industry developments. So join us as we explore the world reliance on tech.

Speaker B: Hello and welcome to Better Tech. This is Jocelyn Houle, your host on data and AI topics here at Better Tech. And I'm really excited to welcome Pravag, uh, Patel, who's the uh, head of engineering and AI at Invisible Universe. And I'm really excited to talk with him a little bit about building content in the AI dominated future. Welcome Pravak. How are you?

Speaker C: Pretty good. I'm um, doing well. How about yourself?

Speaker B: We're doing pretty good. This is of course my favorite topic to talk about. I'm really excited to meet with you because you've had quite a varied background actually of traditional uh, computer science moving into machine learning, which was dominant in certainly an industry for many years. What, uh, tell us a little bit about your background and why you've turned from the machine learning world into the AI world.

Speaker C: You started really nice again. I'm the head of engineering and AI at Invisible Universe. Um, and one of the cool things about transitioning is it's not really transition, right? AI uh, is still a machine learning at the core and that's what we still do at Invisible Universe.

Speaker B: Oh good. I was hoping you were going to say that.

Speaker C: Uh, and that's what we're doing, kind of building an end to end platform at Invisible Universe that starts from ideation, uh, writing scripts and final cut video and all with AI.

Speaker B: Who are your customers at Invisible Universe?

Speaker C: The core customer for us is a creative person who just wants to create content. Uh, and we can go into the journey itself. But the way we started was we started building an internal tool for ourself, um, and a bit of a background about the company itself to help clarify what I'm about to say. So we started as an animation studio, um, and that too as a um, social media first animation studio. And our thesis was if Pixar were to start today, they would probably go social media first with short form content. And we have pretty good industry partnership with big celebrities. And that's how we started and thing that flipped for us was when AI came in we had a strong thesis, thesis around AI is just going to change how we create animation. And that's where I came in. It's like let's just build something that helps internal team leverage AI and how we can make it efficient.

Speaker B: Interesting. Let me back up for a second. So you really started out with a different idea which was a Social first media production company and one of the big assets there were your relationships with like, I assume, influencers, celebrities, that type of thing. Right. Um, educate me a little bit. I've been locked up in enterprise software for a long time. Um, is this how the world works really? That Pixar would start off by doing short form social? Tell me a little bit about sort of the macro trends there.

Speaker C: Right, so let me use an example of our own brand to help clarify an example. So the first ever brand the company launched that was based on a doll of Serena Williams kit, uh, and that brand itself we just started posting on social media, of course, initially with her backing and stuff that helps brand grow a bit. But then you're still on your own to grow the brand and talking to the audience and stuff. Uh, the cool thing about this is you're kind of constantly doing micro a B test with the audience themselves and audience help you drive the brand narrative. So you're not only looking at views per se, you are looking at how, what's the sentiment on the comments and help them define the evolution of the brand is.

Speaker B: So this sounds much more immediate, uh, if you can get enough people. Because my experience has been, of course I said big companies doing a B testing, you have to, uh, you know, work with many people to launch something and then a whole bunch of people get the raw numbers and then they process it and then, you know, even if you have good tools, it's hard to get that information. And so is this what I'm understanding you to say is now sort of industry standard would be to just use social media platforms to do quick tests?

Speaker C: Exactly. And not only that, I think let your audience drive your storytelling instead of you sitting in a silo for months, or maybe in most cases more than a year trying to define how that brand would look like and those kind of things. Now you're going directly to the audience day one, um, and going again with one of the cool things that I realized coming from the tech world into this world is like just one example. We have another brand called Amber. Uh, she's the devil daughter came to earth and she has a huge Instagram following. Right. And everybody on the social media were like, hey, when can we look at your dad? And that helps us build the storied arc around like, no, we're not showing you the dad yet, but let's just tease out the content that helps your audience engaged enough. Right. And that's your storyline right there.

Speaker B: I love that story because it's not, it's truly Collaborative, you're not just doing what the audience is telling you. You are already going to introduce that dad figure. But it gives you a sense of how to build the anticipation and what they're really going to enjoy. Uh, so that's a great example. Um, well, so then your pivot really was AI comes on the scene much faster. You have a comp sci background. I mean, AI has kind of been the up and coming neighborhood for so long and I stopped paying attention and then all of a sud kaboom, here it is. And so you decided that this is really something like a platform that content creators would want.

Speaker C: Right. And um, so the way we started on this journey is, um, so when I joined the company, I think it's a little bit over a year, we had a strong thesis around like a. It's going to uh, just change the whole industry. Right. But how was still not clear in a way. So back then I started shadowing all the veteran animators that we have within our company. Um, they are awesome people who still uses Maya and they were composing shots. Like if you think about a one minute or two minutes animation, most people don't think about how much effort it goes into creating that even two minute shot. To date, that entire animation is tuned hand by hand, shot by shot. Of course we are not drawing anymore, but it's still animated shot by shot. And it's incredibly complex process that goes to creating that particular two minutes animation. And I was watching them see, it's like, oh, this is where it's stuck. AI is going to just change how we do animation. You no longer need to do that hand by hand animation. You just type it and it'll do it for you. And that was the kind of a revelation. Like how do we build a software that lets you do that with a human intuition and artistic things in mind?

Speaker B: And so, um, let's actually this is a useful thought process. So um, why don't you tell me step by step what functions your product covers and then I want to talk a little bit about how we do that today in the manual world. So the new thing that you're building, the transformational thing, ah, let's say, ah, you and I are building a whatever one and a half minute, like some, some clips that we want animated clips. We want to go out to help support our brand. Um, walk me through the process. I'm your new, your new customer. I want my brand out there.

Speaker C: Right. So it really depends. So let's start about, let's say you don't have uh, you're trying to ideate a new brand versus your more established brand. Let's first focus on the established brand. That's where we shine in.

Speaker B: Yeah.

Speaker C: So in general, let's say I have a brand.

Speaker B: Let's say I'm like, I'm a successful celebrity and I want to a line of handbags or something.

Speaker C: Can um, we change?

Speaker B: Tell me a good example.

Speaker C: Huh? Yeah. So what we focus on is more of character building. We launch animation brands. Right. So it can be around going, uh, back to that example around Quakeway. You want to launch an animation series based on this character. How would you do that? Right. Uh, the biggest problem with AI tools today is you don't get consistency around whether it's images or even text prompts. Right. Uh, but when it comes to brand, even if it's a bag, every time you type in a prompt, you want that exit bag to show up or exit character to show up. And that's what we make it really easy.

Speaker B: So let's say I'm spinning out a character from uh, a successful animated series and I'm coming to you to spin out that character what you want. This is actually exactly what I wanted to ask about, which is, um, because it's a non deterministic kind of system, Right. You can't be sure what's going to come out. I guess you have to have like really hard and fast rules about what this character looks like, what types of things they do. Um, maybe the way they move or talk has to be at least reliable. Right?

Speaker C: Exactly. So the way we have what we call is a creative sandbox, um, and we have a mix of proprietary technology and the open source tools that help us define that. And I can go into more detail now. So when I say that, let's talk about the biggest fundamental problem, image generation. For now, you want this pack to look exactly like that. So what we do is we work with uh, a partner to say, hey, can you, let's source your data sets, give us your past scripts, uh, images that you might use, and then we fine tune, uh, image models. That helps us get the exact character that we want every single time.

Speaker B: That's nice. Um, then we, not to interrupt you, I want to hear more about that. But it is really nice because people don't know the answers. Often you're observing the material that's available and kind of building up organically based on the character's behavior, the branded output that's already worked. And so I think that is really helpful for customers. There's rarely one Person who can just write down everything.

Speaker C: Right. And that's where the creative sandbox, uh, helps really well because you're encoding all your past data, whether it's scripts, your brand bible, all into this sandbox. So now let me answer you that specific question. Let's say once you have that creative sandbox configured you with fine tuned models, prompt engineering and all those things, now a writer or a creative person can come in and say, hey, let's ideate on a new YouTube shot. And the sandbox would spit out this exit script that is tailored to your brand because we know about your brand based on all the data that you have provided us. Mhm. And then you go from that step to say, okay, now I have figured out this, how my storyboard looks like I can start creating images for each shot by shot.

Speaker B: So this is where I need you to educate me a little bit. When you talk about a storyboard, I'm imagining that this is in some sort of editor, right? You have an ide of some kind that will generate this.

Speaker C: Right. And this is a little bit, I can go deeper because it's a bit different than how we in tech operate versus this. You can think about a storyboard as uh, frame by frame description of what's going on in a uh, two minute animation, uh, a character starts with a dialogue, they enter a different room, um, and those actions are defined in a storyboard to execute it. From the traditional standpoint, uh, with a production pipeline. When I say production, I mean animation, production pipeline. You would create uh, 3D models and then hand by hand, start from entering a room to exiting a room, going somewhere. You'll do entire process end to end. While with this new world you're starting with a text. You know, you can use somewhat automation to create this well formatted storyboard for you. Where you have visuals, you have audio defined for you. Mhm. And then you can now click of a button, say okay, create me an images for this visual.

Speaker B: So you get the script, proposed script, you read through that, see if that's kind of what you want. I think this is such an interesting moment too that you can create these storyboards at a high level. So these are like the big moves in the story. Person enters the scene, person exits the scene. And so if that's meeting your requirements, then you can really automate. You know, here's the seven steps that it took to get from the chair to the door. You don't really need to somebody hand configuring every single frame precisely.

Speaker C: And uh, this is still not out there. One of the things that we have been thinking about is once you have storyboard defined and this is where you need human in the loop because they have a very nuanced understanding of the punchlines, the comic sense, like AI does not have it yet.

Speaker B: It doesn't have it. And this is. You're actually convincing me. Because I think on the data, like analytics side, people put a human in the loop and it's never really great. It doesn't work. I'm against that. I think just let software do as much as possible. But in your case, of course, it's, it's essential to the whole process. It really kind of won't work without that interaction.

Speaker C: Right.

Speaker B: That's cool.

Speaker C: Exactly. And uh, the cool thing is I sit next to creative people as a tech person and they tell me that hey, the script doesn't work and as a tech person, it produce text. It works. Right. But a creative person is a totally different viewpoint than I as a person who knows AI has.

Speaker B: I'm dating myself. I don't know if. But I remember user acceptance testing used to be kind of much more subjective and half of real human saying yes, it'll work or no it won't. So that's like the ultimate user acceptance testing. That's harsh to have to sit right next to the person.

Speaker C: Yeah, it helps. Right? Because now you know the limitations where the AI ah, falls out. And that's where this uh, copilot works way better for creative people than the tech people.

Speaker B: Mhm.

Speaker C: Because with copilot you are helping them where they need. And a human brings in their own creative instinct to augment that.

Speaker B: Especially for what you're doing, there's a part of it that's kind of undefinable. Even the creative person doesn't really know how they're doing it or why the joke works. But you know, that's what they do. And so I really think that's so that. Thank you for explaining that. So I didn't want to interrupt you. So there's this, um, sandbox gets all trained up, um, and then the output is a script. From there the human can generate, um, storyboard. And then if everybody's happy, you can actually automate the next part of creating the finished product. Is that correct?

Speaker C: Right. So our dream world is, um. We are still not there yet, but our dream world is. Once you have the storyboard, with a click of a button, you can do a first pass animation end to end. And now you can literally as a human, go and look at what that would look like. And then you can make some edits to it based on, again with your intuition and your creative sense. But we simplified a lot. And when I say simplify, we're talking about getting from weeks or months to five minutes right now with the state of technology.

Speaker B: Well, I won't buy, I'm not buying into five minutes. I do buy in that it's probably a matter of like hours compared to um, days and months in traditional production. And so actually kind of answering my next question, but let me just say it out loud, which is I, um, imagine that your customers, uh, the cost differential, even though this is a costly process, probably, uh, in terms of just compute this, um, probably is what do you think? Like order of magnitude, like what percent of life? It's like, you know what I mean? Is it 10% of the cost of the traditional method?

Speaker C: Oh, uh, you'll be blown away. It's 90%. And this is real numbers.

Speaker B: Okay, because, because, you know, people are worried about that. There's a lot of GPUs. And so is this fun? Uh, is this like financially from a business perspective a better decision? I certainly see it from the speed perspective. Help me understand the cost perspective.

Speaker C: Yeah, so I think that what most people are missing, and this is m. I'm talking about the animation world, okay, where GPUs were still heavily used in the animation process even before AI.

Speaker B: That's right, I forgot that.

Speaker C: Right. So the whole process that we talked about, you had a human who was doing the first pass in Maya or something, and then you would take it and send it to Render Farm. So you went one minute, a, uh, two minutes video is costing you $4,000 to $8,000 just to render it interesting. And then you're pleased. And not counting human cost, or two weeks of animator time, a 3D animator that like 10 people involved in the process, even before it was sent to Render Farm. And now what you're saying is you don't need that entire kind of full blown animation pipeline animators per se, but you have a creative person who can literally go to the end product fairly quickly. Again, I know you're not buying five minutes, but even if it's two hours, you're talking about weeks to two hours, right?

Speaker B: So, uh, yeah, that's going to be much less. And it's interesting too. I think you're really getting the best out of that person because who is the creative by having them look at more of a finished product rather than slog through a bunch of kind of frameworks or, you know, early, uh, scaffolding.

Speaker C: Right.

Speaker B: Is that the best place for them to spend their time? Probably exactly.

Speaker C: So one cool thing that worked for us is, um, when we build this tool internally for ourself and sitting next to your customer, right, was awesome because we're getting this instant feedback that this is working, this is not working for us. And, um, that helps us move faster in a way that the product itself was built for creatives and not for animation per se.

Speaker B: Not just for, like the CFO of the animation studio. Because, I mean, the animation world's sort of famous for, like, just focusing on the bottom line to the point where they're not taking good care of their creatives. And so you, um, really kind of kept that creative mindset or the creative, uh, user in mind because you had to. They're sitting right next to you.

Speaker C: Right. And one last comment I would make is, at least from, from what I'm seeing right now, is we definitely started with enterprise customers today.

Speaker B: Okay.

Speaker C: Because they do feel the immediate need for it. But we also open up the same tool to anybody, like on the Internet, who wants to create short form creative content. Right. And the, uh, the barrier to entry has been pretty much going next to zero. You don't need to learn all those complex tools to be an animator. You have to be creative person to animate it now. Right. And which is the distinction that we are making.

Speaker B: I don't know this world well enough. Give me a quick flyover of like, what would I. So this is your product. Right. This is such a cool idea. Um, I'm going to ask you harder questions in a minute. But, um, but, uh, you know, what would I, like, what would I have to know as a creative and where would I spend my time? Just paint me a picture of the world. Without your thing, you can think about a creative person.

Speaker C: You're talking about a creative person. Right.

Speaker B: Like who Persona is. Yeah. So what would I have to know that I, with your thing, no longer have to struggle with or learn?

Speaker C: Yeah. So as a creative. So traditionally, when I think about terms of Personas, a creative person traditionally has been that, uh, who is, uh, more industry savvy, uh, who's following latest trends on social media, who is, who knows how to write good stories, how to tell those good stories. While if you think about animators and production is they will take script from a creative person on the execution side and then they are the ones who will execute it. So the bridge that we are gapping right now is the execution gap. If I Have a good creative instinct in terms of what story works, how do I tell a story? I can create that content myself, which wasn't the case, um, until last year.

Speaker B: I see. So there was just this long kind of process of interpretation and maybe they get an output that's meeting what they think will work, won't work. I think it's so interesting, uh, as I think more about AI, one of the things that's standing out to me is that it really changes the workflow. Uh, we would think about it as quality in some ways, but it's really the creative component, the human component is instead of spending all your time upfront front curating all the inputs, you just need that person interacting with the outputs. So you're really speeding that up.

Speaker C: Right. And you're spot on. So one example that's still on my mind when I joined last year is a creative process would be a creative person who write a script, hand over to animators. They will have a first pass, give back the script like this. Work. This doesn't work. Give back. So that whole back and forth is no longer there in a way because you are just executing yourself. And that's a learning curve. So the last comment I would make is it's still a learning curve for a creative person to start doing it. So when you're creating. So that's where the most tech companies are missing the mark in a way. A bit, from my perspective is they know that this is going to happen. A creative person is going to execute it. But the mark that most companies are missing right now is you don't want them to learn prompt engineering. Right. You don't want creatives to become prompt engineering. And that's the thesis that we have with Studio as well. They don't need to know whether they're using OpenAI or Gemini.

Speaker B: Yes. This is a limitation. Yeah, yeah, yeah. We forced everyone to learn Excel because there was a limitation of the machine. It's a limitation. We just. That's how PCs worked and apps worked and that's the best we could do. Um, and I do see this a lot in industry as well, where people are just grafting on an AI ification to a current process that should actually get thrown out entirely, potentially.

Speaker C: Exactly.

Speaker B: It's interesting. It's kind of a mental model that's me too. I struggle with it sometimes too. It's a very different way of working.

Speaker C: Right. And I think that's one of the biggest thing, always on top of my mind when we're designing the studio, like Invisible Studio. Is how can we make it as natural as possible for somebody to come in and work in the studio, uh, without having to worry about AI. Ah, they should not feel that they're using AI. They should just say what they want and this tool should do it. And now we have AI to solve those things.

Speaker B: Interesting. Um, some more softball questions before we get into the more spicy ones. But for uh, um, Invisible Universe, uh, have you had where are you in your um, development cycle? Do you have a couple customers or give me a sense of where you are in your maturity of your product and company.

Speaker C: So like I said, um, we started with the internal tool. That means within the company, everybody's using it. At some point it was non negotiable. You have to use it and we have, you have to use it. Oh, I can't live without this kind of a phase already.

Speaker B: That's great.

Speaker C: And we started opening up beta. Great. Ah, for the consumers. So we do have quite a bit of interest and we have long list of wait lists to come and try.

Speaker B: Okay.

Speaker C: Again, we are mostly restricted based on our bandwidth.

Speaker B: Sure, sure. And so you're very early. You're very early. But you know, I love what you're saying about using it internally. Just, it's a unfortunate fact. I feel like a lot of uh, companies barely have a design in figma and they have a pitch deck and it doesn't really have the maturity that uh, you know a product that's been dog food, it has.

Speaker C: And one thing I want uh, to make sure that I also articulate is uh, we're at a stage where we also realized that we had a strong industry partnership. We have direct connections with big studios and big toy manufactured companies already. So we already did bunch of POCs with those big companies. Uh, and we are at a stage where we are actually signing long term contracts with them.

Speaker B: I had a, I had a feeling that might be the case. Um, right. So that's, that's great. Congratulations on that. Creating something complex like this is hard. The hard question I was going to ask you, which is, you know, I've seen demos of bits and pieces of this across, you know, the Internet, mid Journey and OpenAI. I have something like this that's in the works. Um, you're sort of your sandbox. You're sort of describing a rag kind of like setup. You know, you probably already get the question like, why can't I do this myself? I already know you can't because it's hard to package this together in a way that people can just it's usable and I get that. But like, what are some other reasons that buyers might need um, to understand this as its own standalone product rather than kind of roll your own or DIY it if you're in a large company.

Speaker C: Right. So I'll make small corrections. So rag, we barely do rag. The uh, rag for us is very small. So the smallest component for us, and that's only during the first phase, that's for ideation. That's the only way we use rag. I think the biggest, two biggest thing in my mind that we provide is fine tune image models.

Speaker B: Okay.

Speaker C: Because that's what none of the big computers are doing because they are trying to solve for generic use case.

Speaker B: Let's back up. Why do I need a fine tuned image, uh, generator? And why is that easier, difficult?

Speaker C: That um, goes back to that character consistency thing. Every time I prompt, I want my character to look exactly the same. We have an AI, uh, engineer within us and that person has trained more than 300 image models in last six or nine months. And that's where the demand is. Because let's say if you have a brand, like going back to your bag example, if you type a generic bag in OpenAI, every single prompt would produce a different bag.

Speaker B: No, that's right.

Speaker C: So that's why we need uh, the fine tuned uh, models. And what we do is uh, we piggyback off a couple of good open source models and we fine tune them to tailor it to the brand. And that's the biggest component of that creative sandbox that I was talking about.

Speaker B: Oh good.

Speaker C: Um, then uh, on top of that, what we have built, which is about to be released next week or so is uh, we learned all the things about how do you fine tune this image models and build a one click solution for it where let's say um, you already have established brand, you're coming in. Um, but how do you take that one person who has trained knowledge of the 200 AI models that he has trained, uh, uh, to a tool. So we build a solution where you can just give us an image or a character image, click a button and we'll fine tune that model for you. So now every time you prompt your character, you get the exact same character that you want.

Speaker B: This is actually very tricky in technology. This kind of consistency, it's really hard to pull off, uh, even in text. So that's a great answer. I understand that. Um, let me ask you this because, um, I'm sure you've been following the discussion around AI slop, right? The barrier to creating, uh, these types of, um, certainly social media clips is dropping all the time. And this has been certainly a cause for concern in the cyber world because it's a tool for bad actors. But even if you're not a bad actor, uh, you know, people will say, oh, you know, the whole Internet is just AI slop consuming other AI slop. Um, do you guys think about that and, or have your customers asked about it? What is your opinion?

Speaker C: Uh, so I think the, the difference with us is given that we are an animation company who didn't start during the AI, uh, world, we had a very good notion of what brand means and what it means to be on brand. So I do think that we are in this transition phase where that's with how Internet started. Right. There was so much that's true. Uh, uh, I'm missing the good word, not to use bad words.

Speaker B: Kind of like it was so much before the web. They called it Nutnet. I remember people would say, it's nutnet, it's full of nuts. And they weren't wrong.

Speaker C: I have a good example. When the Internet started, there's a lot of piracy going on, right? Copyrights issue, uh, uh, with the content and stuff. But now we don't have that anymore. And I think that's my perspective. It's like we're in this transition phase where we're seeing all those slop and stuff, but at the end of the day, that brand, following that, building the IP that people love, and that's how you monetize, that's how you build brand and not with slops.

Speaker B: Yeah, you're really kind of focusing in two areas in the seam there that are the very important gray area, the experience of the creative and getting that creative point of view in. And then really, um, defending and promoting a brand completely, um, is difficult. We all talk about branding and our brand, we all interact with branded goods. But, um, how do you see, like, help the listener understand why is branding. Why is staying consistent with brand so difficult?

Speaker C: Can I say that question again? Sorry.

Speaker B: Yeah, yeah, sorry if you got distracted there. Um, I love your point about your expertise in protecting and defending a brand. I think most, uh, even business people don't understand how hard that is. What are some of the elements of brand that people probably aren't thinking about? You think about a logo as brand, but what else is it?

Speaker C: Again, my mindset is coming from the tech, so I might not be the right person to answer this, but from my perspective, uh, being in this industry for about a year, uh, I think the brand is something that a general audience love, that thing that they can relate to, the thing that you usually grow up with. Right. And it's really hard to build that emotional connection. Mhm. And that's why you will see industry try to protect that brand. Because as soon as that emotion gets diluted or you're off brand, then you're losing that connection with the brand. And you can see why a big franchise ends up failing is at some point you have third or fourth movie that comes out and it no longer the same thing that was the first movie. Right. And I think that's where in my mind is brand is so much important.

Speaker B: And consumers are tough judges. Right. They are in, they're interacting with brand in so many more ways than just visually. It's the way the character talks or the look and feel of the background, the ancillary characters. Like all of that is feeding in, um, in very minute ways that are we, we all notice, even if we don't know that we notice. So, uh, it's very detailed. I think that's interesting. You know, have you, you mentioned like, hey, I, we had to tell everyone you're using this tool. I think it's, you know, creatives in general can be very set in their ways. What kinds of resistance or criticism or you know, concerns have you heard from the people who might.

Speaker C: Right. So it's your typical change management process. Right. And um, one of the things that we talked about is we are building something for years or two or three or four years down the road. So when we actually started on this journey, the AI was not that great to begin with. Um, the image models were not at the stage where they are at today.

Speaker B: That's right.

Speaker C: Like just VO3 is mind blowing. I cannot so get over with that. Like, but if I would have told somebody that that will be the case last year, it wouldn't be the case. Right. So from the change management perspective, our messaging was, hey, we are here to help use this tool, identify the gaps and tell us what's working and what's not working so that we can help you fix that instead of just discarding that it's not going to work. And I think that was how we started this journey. Like, okay, try it out. Um, give us the feedback and we keep on iterating on it until it gets there. Um, and at some point you would want to put a hammer in there. In my mind, say, hey, just use it for a month before you just discard it. In my mind there are two or Three different categories of people. People who are generally curious, they'll just try anything that they like. Then there is this middle ground who are not sure but they will do it. And then the other one is like no, no, I'm actually not trying it. So that's where you need hammer for. And the first two you'll probably slowly get them fairly quickly.

Speaker B: Yeah, I agree. Even the late adopters though can help you sometimes. They have good comments. Um, so uh, we're going to get into our rapid fire questions pretty soon. But um, for our listeners or anybody who's interested in um, I mean we're going to put in the show notes a little bit more about Invisible Universe. But um, are there any um, resources or um, newspapers or new things that we should be keeping track of in this animation AI automation space?

Speaker C: Yes, I think uh, to me the biggest thing is the innovation that's going on in the video model.

Speaker B: Tell me more.

Speaker C: And so the timeline in my mind is all this AI revolution started with the language models and the video models and the image models were kind of ignored when it started.

Speaker B: They were, yeah, I get you, I get you.

Speaker C: Right. And in the last one year image models are kind of almost caught up where you can get like just the Last release from OpenAI with that image model. That's just awesome. And there are a couple of others like Flux, um, Google and they are pretty good at this point. And the last from in my mind is video, where it is getting so good that now it adheres to your prompt so well. Um, again, uh, we are reaching at a stage where you can tell a model that I want this kind of music and this dialogue and the model comes out with the lip sync music sound effects. I think that's where the true innovation is going to happen in next year as you see more and more large video models starts coming out.

Speaker B: Yeah, I like that point of view. I think that's right. I think yeah, we're going to move from text first to maybe video first.

Speaker C: Right.

Speaker B: We kind of always knew that was happening. But I think you're right. It's going to be a fast timeline on that.

Speaker C: Right. And again that's the hardest thing to solve in my mind, uh, because voice and text were kind of close related in a way. So you started with text, you went to voice image and now it's video's turn and that's where humans um, are uh, great at the visual things. So you're going to see a world where our interaction will become a lot more interactive. You can prompt you can talk to literally at some point, I'm hoping to the video model itself and the video would change based on what you are suggesting. Right. It's going to be an interactive experience going forward. When it comes to the animation, again, I'm talking about a dream world, but that's where the innovation would likely happen.

Speaker B: Well, it's an exciting dream world. I want to see that happen. Um, and I think it's going to happen sooner rather than later. Let's get to our rapid fire questions. Um, what's one quality that you think separates good leaders from great leaders?

Speaker C: Uh, I think the great leaders are complexity compressors. They simplify things. Uh, both inward communication for communicating with their team and to the stakeholders or to their partners. Right. I think that's what the creators do.

Speaker B: What's one industry? Well, we've talked a little bit about this, but outside of your industry, what's one industry that you think AI will completely transform?

Speaker C: Um, I think healthcare, um, I think it's even beyond paperwork. Paperwork is definitely happening to change, but it's about drug discovery and personalized medicine. That's where the next biggest thing would come.

Speaker B: What's the habit or routine that keeps you at your best professionally?

Speaker C: Um, I call it curiosity. Deep dive. Every morning, 15 minutes, start reading the articles that you like the most and then give yourself a time to either say, go deep in the topics that you like and don't time box it to an extent.

Speaker B: What's the best advice you've ever received as a leader?

Speaker C: Um, so as a tech leader, the best advice that I got was you need to understand the assembly language of your stack. You once in a while get your hand dirty, understand what's going on at a deeper level to be able to be a good tech, uh, lead.

Speaker B: I agree. I tell students this all the time. You can't take the tech out of the tech business. You really have to know it at some level. Right. You got to get in there. Um, which tech leader, past or present, do you most admire? Like the whole like Mount Rushmore of tech leaders in your mind?

Speaker C: I'm still a little bit old school, but I think Larry Page and Sarah Gabriel still kind of my, uh, this kind of God look up to. Yeah, yeah. Uh, they changed the industry like completely whether you like where Google is today or not. But that's a totally different discussion.

Speaker B: Um, and then, um, what is your favorite? Do you have a productivity tool or an app? Doesn't even have to be like for business, uh, that you would recommend or that you love?

Speaker C: Um, I've been big fan of deep research lately in either using ChatGPT or Gemini. It's just amount of things you can learn by just typing and then reading that summary of the deep research itself is mind blowing. This has changed how I operate.

Speaker B: Uh, and then if you had to summarize your leadership, uh, philosophy into like a tagline or a motto, which you may have already shared, uh, what would that tagline or motto be?

Speaker C: Um, it goes back to my first rapid fire question. Um, it's like clarity fuels momentum and I think that's the most important job that you can do as a leader.

Speaker B: I love it. Well, thank you so much. It was a really great conversation. I feel really educated, uh, in your business, which was, uh, not really, uh, known to me. Thank you. And we'll see you, uh, hopefully out on social media. We'll start seeing your videos soon.

Speaker C: Awesome. Yeah. Thank you.

Speaker A: We look forward to bringing you the latest industry news in our next episode. In the meantime, check out our other episodes@techcell.com podcast and be sure to subscribe to our YouTube channel so that you never miss an episode.

Speaker C: Sam.