EP 497: Inception Games Round 1: Who's the Top NVIDIA AI Startup?

This is the Everyday AI Show, the everyday podcast where we simplify AI and bring its power to your fingertips. Listen daily for practical advice to boost your career, business, and everyday life.

The madness isn't over. Actually, if you're an AI fan or a fan of startups, the madness is just getting started. Welcome to a special edition of Everyday AI. This is the Inception Games. We have a tournament, uh, a fantastic lineup of 8 awesome startups out of the NVIDIA Inception program, and we're going to be handing it over to you all. Uh, we're gonna quickly on today's episode give you 8 fast pitches from our awesome 8 from NVIDIA Inception. And you, dear listener and livestream viewer, are going to decide which one moves on. So maybe your team is out of the big tournament and they weren't dancing this year in March. Don't worry, maybe one of your favorite startups is in this very competition. All right. I'm excited for this one. It's going to be a fun time. What's going on, y'all? My name is Jordan Wilson and welcome to Everyday AI. This is your daily livestream podcast and free daily newsletter helping everyday people like me and you not just learn AI, but how we can leverage it to get ahead and to grow our companies and our careers. So if that sounds like you, welcome. You're in the right place. We do this every single weekday, you know, on our website, which is where you need to go, youreverydayai.com. On there, you can listen to now more than like 500 episodes or almost 500 episodes anyways from some of the world's leading experts on generative AI. And, you know, one thing I noticed is a lot of times we don't bring a lot of startups on the show because sometimes I'm like, hey, you know, sometimes startups are super advanced and, you know, they have great, fantastic products and sometimes they're not. But with the NVIDIA Inception Program, which I partnered with for this series of shows, it's a legit startups, some great ones. So when I was at the NVIDIA GTC conference, I was lucky enough, like I said, to partner with NVIDIA and to be able to go interview 8 awesome AI startups. So for our livestream audience, I hope this one's going to be a lot of fun. But if you are listening to this on the podcast, you can still get in on the voting action. So real quick, here's how it's going to work. I have kind of on my screen here 8 different quick video pitches. They're between 3 to 5 minutes long. All right. So you're going to hear who the pitch is from. You're going to hear a little bit about their company. And, you know, I'm going to be interviewing them, you know, on the GTC floor here. So we recorded these about, about a week and a half ago. And I want you, you know, for our podcast audience to listen in. Which one is the best? Which one would you want to use, or which one do you want to hear from again? Because essentially we're starting with 8. It's an elite group, but only 2 are going to move on to our final show next week. And like I said, for our podcast audience, I know that's where our bigger audience is. I need to hear from you. All right. So you can come and vote in 2 different ways. So pay attention. And hey, livestream audience, you know, Dr., Dr. Scott, Michael, good to see you back. Big Bogey, Brian, everyone else, Sandra, you actually get 2 votes. Okay, so hear me out. You can vote in 2 different ways. Vote number 1 is you can vote once on the livestream. So on either LinkedIn or on YouTube. So what you are going to need to do to officially cast your vote is you are going to put hashtag and then followed by the company name.

Okay.

So as an example, the first company in our pitch, the first of 8, is called DeepChecks. All right. So if after all 8, you're like, yes, DeepChecks is the one I want to hear more of, maybe they're the type of company, you know, that would be great for your business, right? I think a lot of these companies that we're going to be going over today from the NVIDIA Inception program are probably great solutions that your company has been looking for. Because let me be honest, I talk about AI every day. I love startups. I follow the startup scene very closely. When I was at the NVIDIA GTC Inception Pavilion, yes, they have so many startups in the NVIDIA Inception program that they had their own essentially dedicated expo hall. I had only heard of maybe 10% of them. Right. And as I'm going around doing these interviews, I'm like, wait, this, this software is amazing. This startup is amazing. It's going to solve so many people's business problems. Okay. So number one, you can vote on this, on this livestream. So podcast audience, I always leave the link to the live stream. Okay, that's number one. Number two, you can vote in our newsletter. Okay, so if you haven't already, go to youreverydayai.com. In our newsletter, at the very top of today's newsletter, which you can always access actually on the web as well at read.youreverydayai.com. So in today's newsletter for Friday, April 4th, we're going to have all 8 our, our Awesome 8 group, uh, and you can vote on there as well. Okay, so again, 2 votes: 1 in the live stream, use the hashtag, uh, and then 2 in our newsletter. And then the 2 companies with the most votes move on to the finale, which is next week, and we're going to hear some updated, uh, hopefully pitches from them, answering a lot of your questions. Okay, so that's the other thing Livestream audience, as we go along, ask questions. What more do you want to know from these startups? I'll probably have some additional questions, but I'm going to give them your questions as well, because the two that move on, we're probably going to do a quick, I don't know, like 8-minute secondary pitch, right? So all these questions that you have, still get them in. All right. So, you know, as an example, DeepChecks is up here first. It would be helpful for me, livestream audience, if you say, hey, What is DeepChecks' ideal client? Or if you want to know how much does DeepChecks cost or whatever, right? Get those questions in as well. But that's why for your vote, use the hashtag. All right. It's going to be a little bit easier for us to tally them in case there ends up being dozens of comments in the live stream. All right. I hope that makes sense. Real quick, I need a little help from the live stream audience. Let me know if you can hear the audio real quick.

All right.

So I'm going to like essentially be sitting back listening to these pitches again as well at the same time as you. But I want to make sure that you all can hear them. All right. So here, here we go. Let me know, livestream audience, if you can hear this audio and then we're going to start it over. Don't worry. All right, so I am here with Philip from DeepChecks, another NVIDIA inception company. Philip, tell us a little bit about DeepChecks.

Hi there. So thank you. First of all, thanks for having me.

All right, livestream audience, I just started, played about 10 seconds. All right, so hopefully y'all can hear. All right, thank you to a couple of our YouTube audience. This is also how I found out. Which livestream platform is the best because, you know, we just got confirmation from our livestream audience on YouTube, but I'm guessing the— there we go. All right. Our LinkedIn audience just chimed in as well. All right, here we go. I'm excited. So we're going to hear first 3 quick 3 to 5 minute pitches from our first group. Then I'm going to come back on, you know, ask questions of you all, uh, make sure to see what you guys are liking, what you're not, and then we're gonna get the second group on. All right, here we go. I'm excited. Let's kick off the Inception Games. Uh, you're gonna hear real quick, uh, quick pitches from 8 amazing, uh, companies that were at the NVIDIA Inception Pavilion. Here we go. Inception Games round 1, let's get it.

Go.

All right, so I am here with Philip from DeepChecks, another NVIDIA Inception company. Philip, tell us a little bit about DeepChecks.

Hi there.

So thank you.

First of all, thanks for having me. I'm Philip, co-founder, CEO at DeepChecks. At DeepChecks, what we do is we're giving you everything you need to make sure generative AI systems are doing what they should be doing. So the main thing when you're building an LLM-based system, it's really hard to know how they're doing. There's no, like in the traditional machine learning systems, there's something called a test set. You heard of a test set before? No test set, nothing of this sort exists for generative AI systems. So what we're trying to do is enable defining, measuring, and validating the progress of systems like this. So as you're building it, you're making the prompts better. You're changing the setting of your React system. You're having a better, you know, you want to change the different model from, let's say, Gemini to Sana 3.5. and you don't know if you're doing better or less because that's, that's how we help.

So what do we do?

We have a suite of proprietary models, small language models that we combine with a kind of no-code option for creating LLM judges. And then we orchestrate all this together. We call it the swarm of agents to determine per individual interaction, was this a good interaction or did it fail at one of the different criteria? So we're kind of starting off by saying, did the system work or did it not work? Did it have a hallucination? Did it give irrelevant information? Did you know when you're talking to ChatGPT, it says, I'm sorry, I'm an AI chatbot. Yeah, so it takes all those into account and then it tells you, did you manage or not manage? And that way you can score a version. So you could say, Sonnet was better than Gemini on this, and here's why. And it shows you examples. So it's really end-to-end evaluation. Of generative AI systems.

Very cool. You hit all of my favorite words, all my favorite things, right? But tell me a little bit, who are your average customers or clients at DeepChats?

So first of all, any company that's building a generative AI system, any company that's using OpenAI in the background for some sort of textual interface, they're a potential client. The three verticals I'd say that we're working the closest with are healthcare, government, including defense and financial institutions. So those are, I'd say, the 3 largest verticals. But there's really a long tail. Any startup could be using it. Any Fortune 500 company could be using it. And we're proud to say we started with startups and now more of our new business is coming in from some of the larger enterprises.

Very cool. So you kind of told us some of the features. What are the benefits, right? What do, you know, your customers or clients have to gain by using DeepTracks versus if they didn't use it?

So I think the number one benefit you're getting is a higher probability of success for the entire GenAI project. GenAI projects today, when you start them, their chances of success are well under 50%. And I think just by having something of our sort where you can actually iterate quickly, check the next versions, understand what we call the AI progress. Are you actually improving your system? So by having that in place, it's kind of like test-driven development. You're raising the chances of deploying. And then number— the second best benefit is you're improving the timeline. You're going to release more projects and so forth. So there's, there's no real way around it. It's just the question, are you going to be doing evaluation with, you know, you know, hand labeling using CSVs, sending them by Slack or Teams, or are you going to have like a kind of more robust automatic system that's helping you do that? There are many, many different side benefits. So we said, talked about version comparison, giving the go/no-go per version. We can use our AI to assist human annotators. We also have monitoring in production. We have a whole flow of end testing of checking, you know, kind of checking of all the different, you know, risks that happen within these types of systems for the malicious prompts. So we kind of try to give everything you need in one place. But if I have to talk about the number one benefit, It's actually shipping, raising the chance of shipping, and shipping faster. Perfect.

And then real quick, what value has the NVIDIA Inception program provided to your organization?

So first of all, the NVIDIA Inception team is amazing. I really don't know what it's like in other hubs, but we have the luck of working with the Israeli hub, and it's really in almost every aspect we're getting some sort of assistance from them. So they'll give us feedback. Even if they see an event of ours that's not related, they'll give me feedback. They'll say, hey, you should have changed this. Joint blog posts, working together on integrating within, you know, NIMS, within the Nemo ecosystem, helping us figure out how to reach out to, let's say, new verticals like the telco sector is one that we had less experience with, and they really give us a lot of know-how, these specific connections. So at the beginning, actually, when we signed up, it was just like, oh cool, you know, some other program we could join, but it turned out to be a really good choice.

Awesome. And real quick, if someone, a viewer, listener, they're like, wait, I need DeepChecks, What's your quick pitch to them to get them to sign up?

I think the first use case almost always is if you're, you have one, at least one person that's trying to, a few different versions for how to have an LLM application, then you can see which one's better. That's usually the first hook, not always, but basically the main idea is, wouldn't it be amazing if you're building a generative AI system and then you could get a score for every version, like a test set, like in the classic machine learning? It's really once, once you try it, it's really hard to go back to having this kind of voodoo and kind of manual CSV thing.

So awesome. Philip, thank you. If you want to hear more from DeepChucks, let us know. All right, here we go with our next startup. All right, I am here with David from Expander AI. Uh, David, Tell us a little bit about Expander AI.

Sure.

Love to be here. My name is David. I'm from Expander AI. Expander helps organizations to connect their internal systems and build sophisticated multi-agent.

So, I mean, what's this best business use case for Expander AI? Because I know, you know, agentic AI is all the buzz right now, right? So what's kind of an easy-to-understand use case for that scenario?

So customers are now looking to build custom AI agents instead of buying AI agents that are already configured. So for example, support AI agent that takes care of support escalation and answering instead of you to a ticket. That's something that now customers are now doing this. So instead of them building and investing in fine tuning and connecting to different AI applications, we give them a platform that they can easily choose the connectors and design the workflow that they want to build and then run it.

Okay. So is it, I mean, is this for, you know, technical clients, non-technical? Do you work with big enterprises, startups? Like walk us through kind of like what your average client looks like and also kind of the benefit that they get.

So the average client is, uh, with the developer team that would more than 100 developer teams. Uh, so, uh, we're talking about small, medium enterprises, but we are also working with startups. We're gonna release open source really soon. For any developer that are building AI agents, they can just use the platform. But our ideal customer is someone that is focusing on building internal AI agents.

What is, you say, the biggest value that you can add to companies? Is it more of, you know, less time, more potential revenue, right? Like, what's that one big value from using Expander AI?

So it's engineering time. Right now, building AI agents is so expensive and requires a lot of skills that not a lot of developers has. An organization that wants to move fast and build internal AI agents, they need to invest a lot in bridging the knowledge gap and/or going all in on one platform, and then they're losing the other capabilities of the second platform. So we give them a platform that can automatically use all the good worlds of all the platforms. We work with NVIDIA, we work with Anthropic, we work with OpenAI, and developers don't need to compromise about how they choose the technology. So organizations that choose to use Expander, they move much faster. They're focusing on business challenges instead of technical challenges, and their developers are able to complete AI agents much more quickly.

Yeah, speed, especially when you're trying to take advantage of everything Agentic AI has to offer. I think speed is huge, right? Let's talk a little bit about the NVIDIA Inception program. How has this helped your company?

Wow, a lot. So we worked with Inception since we started the company. One of the first use cases that we did together was to release a benchmark. So our main technology is to generate connectors, enterprise, there's like thousands, system. So we built the technology that generates connectors to private APIs and the Inception program helped us to do benchmark with NVIDIA experts. So we did the benchmark with NVIDIA in Tel Aviv and they benchmark our connectors and they published public-facing article about how good the connector is and they proved that using Xpander connectors, AI can do the job 3 times better without Expander. So that's like a game changer for us as a startup. It's an acknowledgement from NVIDIA that we do something that is there with a value, and it's all thanks to the Inception Program.

I love it. I love it. What would you say is kind of the next big challenge that you're working on to help clients, or maybe the next opportunity that Expander is working on?

Yeah, so now we're going to talk about multi-AI agents that are doing multi-step. That's like the next big thing. Everyone knows or should be able to know how to build AI agents that perform up to 10 to 20 operations, but it becomes very strategic when you have an AI agent that can perform a human-level task. And that's a very novel challenge to do right now. And we are focusing on that as we speak. We have an MVP for just this problem. On how to build a multi-Agent with a graph system that can do a very sophisticated, very complex task that goes into human-level complexity. So that's like the very focused area that we are focusing on and that we are reading from customers that they want to solve.

All right. Last question. If someone in our audience heard this and they're like, I need Expander, right? What's your kind of sales pitch to them on why they should use it?

Yeah, sure. So any organization that have internal systems, why spending the time generating connectors instead of just using a platform that can generate connectors and gives you the ability to design graph and state machine with all the frameworks that you have currently available in the market? So instead of going all in on one framework, LangChain, QoEI, NVIDIA, OpenAI, and Tropic, we give you the ability to really use all the frameworks in your private APIs and design a state machine that works across all those frameworks. So this, this technology allows developers to really focus on business challenges instead of technical challenges. And you as a business leader, no, don't need to do, uh, investment in one specific, uh, uh, framework.

You can enjoy all of them. All right. Uh, David, thank you so much. So if you want to see more from Expander, let us know. And now let's take a look at another. Startup. Y'all, these are, these are some good ones. Go. All right, we have our next NVIDIA Inception startup. We have Sharon here from Beamer. Can you tell us a little bit about Beamer?

Sure. Beamer built a technology to optimize video at large scale and it runs on NVIDIA GPUs, so we gain acceleration of video encoding from NVIDIA GPU, a component called NVENC, the NVIDIA encoder. And we make it so much better. We make it efficient by about 40% on average, or 30 to 50%, depends on the use case. And what you get is that you can take these huge video repositories for autonomous vehicles or for user-generated content or media and entertainment and make them so much more efficient running on NVIDIA GPUs at large scale.

Okay.

What would you say is the one biggest problem that you solve for your customers or clients?

I think that every customer that has large volume of video, then all of the associated cost of the video handling has to do with the tonnage, with the amount of videos. So think about that cost half size.

Okay.

You know, that's a huge benefit.

Okay. So who is your average customer or client at Beamer?

So we are approaching 3 different markets. One of them is the traditional markets that we've been there forever is the media and entertainment market. And now thanks to the NVIDIA Accelerated platform, we can also actually approach markets that much larger volume. One of them is autonomous vehicles, which you can see over here. So in order to train an average model for autonomous vehicles, you need 100 petabytes of video. Now think 50 petabytes.

Okay.

That's a big deal.

Yeah.

Right? With user-generated content, we are talking about 80 to 100 man-years of video captured every day. These are hundreds of millions of video clips, short video clips. Thinking about that, you know, distribution, the amount of networking that it requires, 50% of the internet out there is occupied with video. So think about that, half the size.

Yeah. So, so who is your, your average client? And are you mainly working with large enterprise, smaller startups, a little bit of both? Who is your, your average client or customer?

Traditionally larger enterprises, but as of a year ago, we launched what is called Beamer Cloud. So we have a cloud service available on AWS and OCI, Oracle Cloud Infrastructure. And that means that everybody you know, can use it with a very low friction. You open an account with your email and, you know, you can start working. Very easy. No code.

Very cool. Yeah, you gotta love no code and saving time, right? So the NVIDIA Inception Program, how has that helped your success so far at Beamer?

Oh, I don't know where to start. So last year we've been a part of the Inception Program and we are now graduate. Of the Inception Program. So at the very beginning, I think, you know, having the envelope of the Inception Program was huge for Beamer because you are actually coming to a place where everybody comes to see what's new. And it is pretty much, you know, hosted by NVIDIA. So everybody comes to see NVIDIA and what's new, and then you're right there. So the opportunity is mind-blowing. So that's the initial benefit. But also, you know, helping us to take the word out there and helping us with introduction to prospective customers. NVIDIA is an amazing ecosystem. It works like a huge startup. So, so many opportunities are coming our way and every Inception member way.

And this is amazing.

And now, you know, being a graduate of the Inception program, you know, we are also benefiting, you know, from, for example, being getting additional exposure from Inception when they are now offering, you know, all of their partners to get promoted if we are offering discounts to newcomers to the platform. So there was a platform launch just today and we're part of that announcement. Yet another, you know, big thumbs up to Inception.

Love it.

All right.

And last question here. If someone listening out there is like, oh wait, I think I need Beamer. How do you convince them? What's your kind of one sentence pitch?

Um, if you have a lot of video and, uh, you want it to move faster from one place to another, if you want to save on your cost, if you want to have better user experience and marry that with AI, that's Beamer running on NVIDIA GPUs.

All right. Love it. So if you want to see more of Beamer, let us know. All right. And here we are with another startup.

Love that.

All right, what are you guys thinking so far? So that is 3 down, uh, we have 5 more to go. So for our live stream audience, if you joined us halfway through, we have 8, uh, our awesome 8 group in the Inception Games. Uh, you're hearing their quick pitches. Uh, we are 3 down, we have 5 more to go. Y'all, each time, like, obviously I did these interviews, uh, about a week and a half ago, the GTC, uh, show, and now I'm listening back to them and I'm like, oh wait, I'm thinking of so many amazing questions that I should have followed up on. So, uh, 2 are gonna move on to the finals, so make sure to get your vote in and also get your questions in for these companies as well. Uh, hoping to pull out another quick interview for our 2 finalists. Uh, enough with that. Let's go ahead and listen to our next round of startups. Because I'm switching tabs here, uh, doing this live is always tricky, y'all. So let me know if you can hear our second video here if you could, y'all. Here we go. All right, let's now talk to our next Inception startup. This is PlyOps. Tony, tell us about PlyOps.

Absolutely.

PlyOps is a solution accelerator for GenAI applications, very complementary to help people maximize what they can get out of their GPUs and their investments in NVIDIA GPUs.

All right, so tell me about who is your, you know, client or customer, your average client or customer?

Yes, anybody who's putting infrastructure together using GPUs and is looking to maximize what they can get out of those GPUs and overall lower their OpEx.

So like tell me maybe, you know, you can pick out one actual customer, but you know, what is kind of the before and after benefit? You know, are these companies that have mountains of data and they're just trying to figure out what to do with it or what does it actually look like? Before they come to you and then after they come to you?

Absolutely. So as Jensen mentioned, the biggest driver of data nowadays is after the model is trained, how do you want to extract data from it and interact with it? So that's creating a ton of data that is not normally saved, and we help do that. And what does that do is allows these extra cycles as a result of these savings to be used to serve new users and new applications. So precisely that's the reason people are deploying PlyOps.

What do you think, you know, so far? Well, first of all, how long has PlyOps been in business?

PlyOps has been around over 5 years and we have been into GenAI business in the last year and a half, 2 years. So we've really accelerated that program. It was a perfect match with our IP, core IP, which is also a match with what NVIDIA is envisioning for data centers.

So yeah, even let's talk about that. So, you know, generative AI, you know, our audience, that's really what they care about. How has generative AI changed what you do at Plylops?

Completely, 100%. Just like it is changing the world now and it's changing the data centers, how people are deploying gear in their data center, how what people are employing, all of that is, we are very a reflection of that. We have completely changed our product roadmap to match what GenAI needs. And so that means being very much aligned with what GPUs are doing, which are becoming the center of the compute now in our world and all the good things that they're doing now, but really just the start of what they will be doing more. Which is mind-blowing.

So, you know, talk a little bit about the NVIDIA Inception Program. How has that helped, you know, what you've been able to accomplish so far with PlayOps?

Well, for one thing, we are very thankful for NVIDIA to giving us a space here and letting us showcase our solutions. It's amazing to be part of this community. This show has just been so amazing over years, but especially now with where we are with the AI era. So the fact that we are here and we can interact with participants, solution providers, potential customers, and really ordinary people that are, their lives are impacted by what NVIDIA is doing has been tremendous.

What would you say is the biggest problem right now for, you know, everyday businesses that PlyOps solves, right? If you had to say, here's our number one solution that we provide, What is that?

Uh, not for you to having to hire PhDs to figure out how AI works. A simple plug-and-play solution that is, uh, that will give you the boost that you need and avoids having to have the, the kinds of, uh, services that, uh, uh, that are required unfortunately now to, to bring people up to where they need to be with AI.

What is next for PlyOps? What are those next big problems that you're hoping to solve? Right. So maybe I come back in a year and ask you, what's that big, you know, next problem that you're looking to solve for your customers?

Right. So right now we're showcasing maybe 2 GenAI applications. I think next time, next year, this time we would be showing 10 different GenAI applications and we will be offering it as a service in multiple clouds.

AI moves too fast to follow, but you're expected to keep up. Otherwise, your career or company might lag behind while AI-native competitors leap ahead. But you don't have 10 hours a day to understand it all. That's what I do for you. But after 700+ episodes of Everyday AI, the most common questions I get is Where do I start? That's why we created the Start Here series, an ongoing podcast series of more than a dozen episodes you can listen to in order. It covers the AI basics for beginners and sharpens the skills of AI champions pushing their companies forward. In the ongoing series, we explain complex trends in simple language that you can turn into action. There's 3 ways to jump in. Number 1, go scroll back to the first one in episode 691. Number 2, tap the link in your show notes at any time for the Start Here series, or you can just go to starthereseries.com, which also gives you free access to our inner circle community where you can connect with other business leaders doing the same. The Start Here series will slow down the pace of AI so you can get ahead.

All right.

Our last question for you. For all of our listeners out there, maybe you piqued their interest, caught their attention. Why should they work with you or engage with PlyOps?

Simply because we improve your dollar per token cost, improve your margins. We bring, we let you maintain more money out of your AI operations and your GPU operations. And so that's really what's driving things. And so if you come to us, we'll help you make your GenAI data center better for you.

All right.

Sounds good, Tony. Thank you so much for introducing us to PlyOps. And make sure if you want to see more out of the PlyOps team, let us know both in the comments and in the newsletter. And here we go. We're going to go into our next one.

All right.

I'm loving this so far, y'all. That was PlyOps. And here we go with our next.

Got it.

All right. So I am here with Billy from Glia Cloud. Billy, tell us what Glia Cloud is.

So we're a video surfacing process and we're doing automizing the video ad creation. And mostly we serve late, late adapters within sort of like, think like government sectors and your regional, let's say, tourist sectors or a little bit of, uh, I'd say small businesses. Which they're selling physical objects and we help them create optimized ads, things 15 to 30-second YouTube pre-rolls or the ones you get in, let's say, news ad slots, stuff like that.

Okay. So yeah, and like, and I know we kind of have it going on here in the background. So, you know, this is for, uh, you know, companies that maybe don't have access to huge creative teams. So you're using generative AI video. to help them create ads online. Is that right?

Absolutely.

We actually started pretty early on, uh, the early, I think it's around 2016.

Oh, wow.

Okay.

Yeah. I think like GPT-2 times.

Okay.

So at the time we have to make a lot more infrastructure. So like to, so I get around the little like GenAI that was going on at the time. And we actually did our own, uh, video sort of like render engine. And a lot of the work at the time was figuring out how we break this up.

Right.

So yeah, how, how has the product changed? So first of all, 2016, the very early days, love it. I mean, how has the product changed over the years as generative AI just gets more and more powerful?

Yeah, it's definitely having a great impact on the content side, like maybe just a precursor because sort of like our client base are less sort of like knowledgeable on more like technical side of things. So we don't really have a high standard, not to say it's bad. They're looking at our videos more like a product.

They just boost their Sims. Sure.

But within the time, we also, we're trying to make ourselves like really like ahead of the curve. We have like a really tiny task force. It's like staying ahead of like what's new and looking at what's not in your model. So what you're seeing, if you can see on the big screen right there, is one of our experiments. So. Talking about GenAI, we're trying to leverage what's happening, like what's so exciting. We just announced yesterday about like Cosmos and how we can actually leverage the controls for like 3D environments. The models actually understand 3D scenes that we can retain, let's say camera controls. We can initiate camera controls within the first stage, we can output that like 3D accurate video and we put it into video to video. Models that will retain that information, and we can actually create more like live-action-looking transcripts. This is where we will head.

Oh, that's awesome. So real quick, I know we kind of touched on this, but who is your average customer? And if you had to say it in like one sentence, what is that problem that you solve?

That's a great— so average customer, I think legacy media. I think that's a really good one, and they want to boost their let's say traffic through their articles. So we have partnered with some publishers and then we create automatic video for them. And big part is their engagement rate. Also, there's also tourism sectors within governments. Those, we have a really great case study, which ended up like boosting their CTR is for, I think, 4 times of their CTR.

Wow.

Okay. So real quick, the Inception program from NVIDIA. How has that helped your success?

It has, it has been really helpful in terms of the resources we're getting, uh, from, I think we actually started contacted by NVIDIA One Life to the Fabrels from NVIDIA and they told us about this program. And a big part is attending this and connecting with so many other startups and then sharing a lot of, because like both of us are the more technical kind of guys here. So the big part of us is just like sharing experiences, working on this kind of like all over the place is like in clouds.

Okay.

And then if, uh, someone in our audience, if they heard what you said, they're like, that's us, we need this. What would you say to them? Why do they need GLEA Cloud?

I guess things, maybe AI could be a little bit intimidating for maybe someone who's looking into this, this kind of product, but Think of us as an agency, as an agency who's focusing on, let's say, the volume of that video you can create within a short time. Also think of the cost. We are genuinely operating within one-third of the cost of a normal, like, human-led agency.

All right, great. Thank you, Billy. So if you want more from Gliacloud, make sure to let us know. All right, a couple more. Here we go. Give me the little 3, 2, 1, Amy.

Go.

Also, I have to shout out Amy from NVIDIA helping me film all these. So if you hear me say, if you hear someone say go, shout out Amy. And thanks, Danny, as well for helping us hunt down some of these startups. All right, here we go. Contextual AI. All right, so here we are with our next startup, Contextual AI. John, tell us a little bit about Contextual AI.

Yeah, Contextual AI, we're the world leaders in RAG. So we help large enterprises and fast-growing teams build specialized RAG agents for knowledge-intensive tasks. So anyone who's building RAG should be looking at Contextual AI. Our CEO co-invented RAG at Meta and then left Meta to start Contextual AI.

All right, that's awesome. You just hit in all the buzzwords people are talking about, right? You know, agents, RAGs. So maybe just explain for maybe some more non-technical people, you know, why do they need retrieval augmented generation in their company? And then how do you all make that happen?

Yeah, yeah, yeah. At a simple level, you know, we want to have LLMs have access to current relevant information. And of course, a major differentiator in the enterprise or for any business is their data. So being able to connect that data to LLMs and get very grounded responses, that's where we excel. And these can be very complex tasks. Like we work with Qualcomm for their support engineering team, other large enterprise tasks. And we also work on both structured and unstructured data, which is actually a really hard problem. We've taken the top of the text-to-SQL benchmark recently.

Yeah, that's huge, right? Being able to make use of both structured and unstructured data. Tell me a little bit, who are your average customers? Is it just enterprise clients? Is it more medium-sized? Like, who all do you serve?

Yeah, so we typically work with large companies or even fast-growing teams. We do have a free trial offer, so if you want, you can do a free trial here. And scan that QR code and get started and try out our component APIs or even our full platform. You can do a 30-day free trial. And so we really try to make it accessible for developers, but also working with those large enterprise teams. And we have forward-deployed engineers that could help teams get their RAG projects into production. So that's really where we come into play is getting that RAG project into production. If you're struggling with quality, then we'd want to talk to you.

So you kind of, you know, answer this in one way, but If you were to say directly, what is the one biggest problem that you all solve, what would that be?

Biggest problem we solve? I mean, I think RAG is a very big problem. So I hope that we're solving that problem for any team that's thinking about RAG. But then as you're starting to think about agents, you know, moving towards more agentic experiences, you want to have knowledge at the core of that agent experience. That's where we can really come into play as well. So it's connecting that data, connecting that valuable resource in your organization to your LLMs. And then I think the other thing that often gets overlooked with contextual is we actually tune based on feedback and based on ongoing kind of behaviors. So your model's going to get better. Your agents are going to get better over time. So it's not just a static thing. It's something that's going to improve over time. And I like to think about that as your institutional knowledge is being re-encoded back into the AI. So that's how I would talk about it. But I think there's a lot of value for anyone who's thinking about RAG or agents.

Speaking of improving over time, what if, you know, if we're having the same conversation next year, what's that next big thing that you all are looking to solve or improve upon?

Yeah, I think for next year, I think, you know, this is the year of agents. We're very focused on 2025 being the year of agents. So I think that will be our focus and really thinking about how RAG works relative to this agent execution we're all heading to.

All right, real quick, we're at the Inception kind of pavilion here. What has the Inception program from NVIDIA meant to Contextual AI?

Yeah, it's been huge. I mean, it's given us presence here in the pavilion. So, you know, this is an amazing group of peer companies to be with. Additionally, being able to market through the NVIDIA team and presence on the blog. So, you know, for us, we just went GA in January. So this has been a huge kind of boost to our ability to go to market. And that's the team I work on. So it's been wonderful having NVIDIA as a partner. And working with them as part of the Inception program.

Awesome. Last question. If you caught someone's attention, what's your pitch to them? Why do they need to check you out?

Where other AIs are like your intern, we're like your best analyst or your best researcher. So that's where the knowledge-intensive tasks come into play. And that's where we really work on very domain-specific knowledge for hard problems. So as my CEO would say, we want you to be ambitious. Think about those hard problems, those really high ROI problems that you want to solve with AI. And bring those problems to us. That's where we're going to excel.

All right, thank you so much, John. So if you want to see more of contextual AI, let us know. Thanks. Get ready for the next one. All right, y'all, 6 down, we have 2 more to go. Again, only 2 are going to move on to the finals, so you can vote by leaving a hashtag and the name of the company in the live stream on LinkedIn or on YouTube, as well as vote on— vote in our newsletter for Friday, April 4th newsletter. All right, here we go. Our last 2. Again, this is our awesome 8 of the Inception Games. Only 2 are going to move on and they're going to answer more of your questions and we're going to determine one winner next week. All right, here we go. We have 2 last pitches, some great ones here. Again, livestream audience, if you could let me know if we can hear the audio on this last group. Two more great Inception startups. Here we go. All right, now I am here with Jack from Democratize. Jack, can you tell us a little bit about your company?

Okay, so Democratize is actually also an AI company, but the A stands for Apparel Intelligence. So we do have provided body scanning and also textile AI to turn body, human body, and also the textile into digital twins. And then we use these digital twins to create your personalized apparel. And right now we focus on performance wearers, so we help all the professional athletes to create really precise and tailor-made clothes for them to enhance their sports performance.

Very, very cool. I love it. So aside from maybe professional athletes, Who are some of your other customers or clients for Democratize?

All right. So actually, this is the new topic that we have. Previously, we are a tech company, fashion tech company. So most of our customers comes from fashion brands or like fashion supply chains. So our bigger customers like Under Armour, they use our textile digital solutions to digitize all the material they have and then use this material to the digital design workflow. So previously we would have customers from fashion brands, from supply chains like apparel supply chains or textile supply chains. And right now for this new project, we are actually working with a cycling fashion athlete. So we are making the very tailor-made and very precise cycling jersey for this athlete. Very cool.

What would you say is the biggest a problem that Democratize solves.

Okay.

So we try to kind of re-engineer this apparel ecosystem, right? Because right now, because of fast fashion and people love new things. And, and so it's actually caused the overproduction issue there in this industry. So it's pretty, um, I don't know, this is a huge problem, but no one really want to address this. Um, so by leveraging AI and all the information, um, technology we have to try to kind of, kind of re-engineer this process. So if you are the consumer, if you say, all right, if you place order and you can get the tailor-made clothes within 5 days, why don't you do that? Because you can get, you don't need to care about any sizing. You just tell me, oh, you want a slim fit, you want a loose fit, and I tailor it for you. So if we can kind of build this process, then we can reverse the whole ecosystem, right? You decide you want to go buy this clothes and we tailor-made for you and then ship it to you. So we don't do the production of demand to reduce all the overproduction issues.

So yeah, it's getting loud in here in the Inception Program about to open. But maybe real quick, tell me a little bit about what the NVIDIA Inception Program has meant to your company so far.

Uh, I think India, they are looking for a lot of AI new startup and then we are pretty, as you understand, we are pretty focused on the fashion technology. So it's pretty niche for them and they like to this kind of vertical end-to-end solution. So I think they, when they listen to our pitch and they are really like it in this area and also it's also related to like sustainable, um, to issue that sort of problems. So I think they gave us a lot of resources to use their XP Canvas software to help build these digital twins and AI photos.

So outside of, you know, maybe is there a next iteration? Are you looking to bring this concept to bigger or wider markets in the future?

Yeah, so as I mentioned, we just pivoted to this B2C direction. So we are going to launch our first prototype set. So we're going to release our first collections by end of this year. And then next year we are moving forward to like extend to different kind of clothes types. Like right now we're working on cycling, maybe next one will be the running. And yeah, so gradually to expand the whole collections.

Very cool.

All right.

So if someone just heard you and they're like, wait, I need Democritize, what's your quick pitch to get them on board?

Good question.

All right.

So I think we just try to introduce a new way to purchase clothes. So we give consumer another choice to be more sustainable and choose the right fit for your own. So you are the, you are the brand, not you trying to fit into the brand's clothes, but you are the brand. So we design for you. You are, you are, you can decide whatever you want to wear and you don't need to care about all the sizing problems.

All right, Jack, thank you. If you want more from Democratize, let us know. All right, and our last one.

Go.

All right, so we are here with, uh, Ina from Lumex. Uh, tell us a little bit about Lumex.

Lumex is a self-service access for data analytics for business users and enterprise. Think about banks, pharma, financial services. All of them would like to have business intelligence as their daily practice. Ellumex enables that in less than 7 days with 80% savings on the top.

Okay.

So, yeah. Okay. So it's about token optimization or like what's the actual, what's the benefit? It's just, uh, just more efficient, uh, tokenization. Like walk us through what that looks like.

The most, the most exciting benefit about Ellumex is actually trust. So business users from our perspective have hard time to actually understand agentic answers, especially when it's black box. And make decisions based on that. So what Inomax does is also handle the quality of underlying data and make sure that the data which is going into our GenAI is of high quality and compatibility from one side. On the other side, we provide the full explainability about the answers so any users can understand how the question was interpreted, which data is mapped to, and what logic is implemented. This is like ultimate open the black box.

Awesome. Walk me through who is your average customer or client for your platform?

The buyer would be chief data officer of Fortune 500 company. Also lately we see lots of giants from Silicon Valley are super excited about the solution as well. But the users are average support center or marketing or a product analyst. So basically any business user in enterprise or any organization really can ask that data-related question and have explained and hallucination-free answer.

That's great. So essentially it's just providing more confidence in the answers you get out of agentic systems.

It's full-stack solution. So we handle data quality, we handle governance and trust, and we also handle interaction within system that you like to use. So itomax is not a new interface. We already embedded into your CRM, your Power BI, your Slack or your Teams whenever you already are, and we provide those answers immediately.

Very cool. Talk a little bit about the NVIDIA Inception Program. How has this helped your organization?

NVIDIA Inception Program was exciting so far. So we recently had PR just yesterday about how we use NVIDIA NeMS and other underlying technologies to basically scale for those enterprises as we serve. Right now, we have systems which has hundreds of thousands of tables and millions of queries, and there is nothing like NVIDIA to to enable us to have 7-day setup, up and running for those massive companies.

Awesome. And then, so if one of our listeners or viewers, if they heard that and they're like, wait, I need this exact thing, what's your quick pitch to get them to sign up for your platform?

If you have silent data sources in your system, you have your SAP, your BI tools, your Azure, and you would like to have A genetic AI that you can trust. Contact Illumex and we'll make you happen for 80% less cost in 7 days.

Awesome. Love the great pitch. All right, so if you want more of Illumex, uh, let us know and, uh, let's dive into another startup. Wow. All right, so that's it. That is our awesome 8. That was Wild, y'all. That was a lot of fantastic startups, right? Like, I wish, in all honesty, I wish I had more time, you know, at the NVIDIA GTC conference. If you've been listening to the show, you know, I've already had like 6 interviews with some of the brightest minds in AI, both from NVIDIA and other companies. So maybe next year, if you all like this format, if you, you know, heard something you liked, maybe we'll expand the field, uh, from 8 to 16 for the Inception games. But that is a wrap, y'all. So, uh, a couple questions that kept coming up. Um, I'm gonna go through all of the questions that came in from our livestream audience, uh, and, you know, for our two finalists, I'm gonna make sure to ask some variations of those questions, uh, to our two finals, uh, that are going to come back for another round, the final round. So here we are, we have our awesome 8 of the Inception games. Uh, yes, a couple of questions that came up. We are gonna have a recap in the newsletter, okay? So, you know, in case you're sitting there jotting down, uh, you know, notes with your pencil on what they all do, we got that. That's what the newsletter is for. That's why I always say we learn, uh, on the podcast and the live stream, and we leverage it in the newsletter, right, to grow our companies and our careers. So, you know, maybe Maybe the startup that you like most from the Inception program isn't going to be the one that makes it to the finals. And that's okay because we're going to have links to all of the startups in the newsletter so you can go find out. Maybe they're going to solve a huge pain point for your company. So as a reminder, you know, because a lot of people are like, hey, I need a quick recap at the end.

All right.

I'm not going to re-pitch them. But as a reminder, we had Deep Checks. We had Expander AI, uh, we had Beamer, we had PlyOps, uh, Glia Cloud, uh, Contextual AI, Democratize, and Illumex. So, uh, make sure you get 2 votes. Use them wisely, right? So maybe you're torn between 2. You can vote for one in the live stream and then a different one in the newsletter. You can vote for the same company once on the live stream once in the newsletter as well. But we're only counting, you know, one vote on the live stream. And, you know, on the newsletter you can only vote once anyways. So get them in now if you, you know, kind of were sitting on your vote and waiting until the very end. Again, for the podcast audience, maybe you want to vote twice, maybe you don't just want to vote once in the newsletter. We always put the link to the live stream in the show notes for today's podcast. You don't have a lot of time. We are saying voting ends Sunday night at 11:59 PM Central Standard Time. So you got a little bit more than 48 hours until we go from our awesome 8 in the inception games to our final 2. And we're going to be bringing them to you all to answer your questions. So I can't wait. So I hope this was a fun one, y'all. Like, this is the first time we've done something like this. You know, actually, it was last year at GTC when I partnered with NVIDIA. I went through the Inception area of the GTC conference and I'm like, wait, our audience needs to hear more about some of these startups because I can tell you, I can tell you already there's been a lot of startups in the Inception program that have gone on to become literal household names. Like, as an example, Did you know ElevenLabs, the leader in text-to-speech, they're in the Inception Program, right? So this is where tomorrow's biggest AI players are starting out today. So I can almost guarantee in a couple of years, a lot of these, maybe you heard about them for the first time, but a lot of these companies that we just talked about, I think they're going to continue to grow, continue to change how we all do business. So I hope this was a fun one. Again, shout out to our partners at NVIDIA at the Inception Program. This was a great one. I love startups, I love AI, and I love kind of this basketball format of the brackets and the games and all that. So make sure to get your vote in. Go to youreverydayai.com if you're looking to sign up for the newsletter just to vote. That's where you can do it. So make sure you go look at today's newsletter, April 4th. So thank you so much for tuning in. Hope to see you back later for more Everyday AI. Thanks, y'all.

And that's a wrap for today's edition of Everyday AI. Thanks for joining us. If you enjoyed this episode, please subscribe and leave us a rating. It helps keep us going. For a little more AI magic, Visit youreverydayai.com and sign up to our daily newsletter so you don't get left behind. Go break some barriers and we'll see you next time.