Welcome to Building the Future with your host, Kevin Horrick. Get ready to explore the cutting edge of innovation and entrepreneurship. From startups to industry giants, we're bringing you the visionaries shaping tomorrow. With listeners in over 150 countries, we're one of the fastest growing shows worldwide. Join Kevin, a seasoned software designer and tech expert, as we dive into the ideas and inventions changing our world.
Intro/Outro:Building the future starts now.
Ad:Building the future is brought to you by iClerk.ai. Are you a small business owner, a professional, or company looking to automate? Do you have difficulty hiring staff to perform mundane knowledge work like reporting, form creation, or information analysis? Leverage the power of AI and the automation of agents even if you don't have the time or experience. Try us out with little risk and no upfront investment.
Ad:It's time to work smarter, not harder. Visit iclerk.ai to learn more and get started today.
Kevin Horek:Welcome back to the show. Today, we have Eddie Weinworm. He's the CEO at Avvious Future. Eddie, welcome to the show.
Eddi Weinwurm:Hi, Kevin. Thanks for having me.
Kevin Horek:Yeah. I'm excited to have you on the show. I think what you guys are doing is really innovative and cool. But before we get into that, maybe let's get to know you a little bit better and start off with where you grew up.
Eddi Weinwurm:Yeah. So getting, my my biography is quite confusing. So, I grew up in Vienna, Austria, very classical, very atypical. I grew up without TV, just reading books. It was quite boring back then.
Eddi Weinwurm:You know, no Internet, no TV, no books. And, but it, I don't know. It it it teach it it it teach me one thing, like, curious curiosity needs work. Yeah. You have to Yeah.
Eddi Weinwurm:Go out and discover instead of, expecting that things get presented. Yeah. So I had to go to the library every week to, collect a bunch of books. Else, my life would have been boring. Yeah.
Eddi Weinwurm:So, yeah, classical, European education background, I would say.
Kevin Horek:Okay. So what did you take, in university and why?
Eddi Weinwurm:So, yeah, university. I I I studied philosophy. So that, was the my my my father told me, like, as a as a main subject. You know? You do you want to become taxi driver?
Eddi Weinwurm:Yeah. And so philosophy, teach me to read, to write, to understand, and to reflect on things, especially to doubt things, doubt their own bias and so on. And, actually, that's that's that's that's something which helped me all the time. Yeah. So I I always was in I always was in between two worlds all my life.
Eddi Weinwurm:On the one side, the logical mathematical side, which I really love, and on the other side, the creative, parts, the more fuzzy, yeah. The, yeah, the the creativeness and, the the the softness. Yeah. So hardness and logic and mathematics and softness on the, communication side and on the creative side. So I was always, in between these two sides.
Eddi Weinwurm:And, so, yeah, philosophy was a subject for me to where could unite both things because it has, the one side, it has logic, and on the other side, it had all this poststructuralistic, theories and so on, which are very, very, creative. Yeah. So that, that was my, university target.
Kevin Horek:Okay. Very cool. So walk us through your career up until becoming CEO of AviosFuture, and let's dive into that.
Eddi Weinwurm:Yeah. So, honestly, I I I I'm not often talking about, this career because it confuses people. Because, people Okay. Needs to they they, you know, they want to put you in boxes. Yeah?
Eddi Weinwurm:And you
Kevin Horek:Carry. Yeah.
Eddi Weinwurm:Around box in a in in in in this, yeah, nerd box or in a creative box or in a management box. Yeah? So my career, I I started, fresh from university at radio as a journalist, later switched to TV, made my career there, directed, directed commercials and so on. And, at one point, I don't know. I was getting, you know, to write software was always for me, like, for others going to the gym.
Eddi Weinwurm:Yeah? Because, you know, in this in this creative world, you're in this creative world, you're always, you know, everything is fuzzy. Yeah. So every discussion, you cannot prove anything. It's it's all speculation.
Eddi Weinwurm:Yeah. Why did a commercial fail or why did it succeed and so on? It's it's all relative. Yeah. It's all a big discussion.
Eddi Weinwurm:And the nice thing on writing code is actually, when I came home, I sit down, write some code because it's so binary. Either it works or it doesn't. Yeah?
Kevin Horek:Yeah. That's fair.
Eddi Weinwurm:Yep. This was like going to the gym. And at some point in Berlin, I was hired as a consultant for a TV station, and they had this challenge of, getting their storage managed. And I was sitting in my my big office, board. I had to attend a few meetings that weekend.
Eddi Weinwurm:That's all. So I sat down and started to write this. And this was actually this started my my first company. And very quickly, I switched from the creative side and from the the the the, yeah, visual creative side to the hardcore tech side. Of course, it was still connected.
Eddi Weinwurm:This product was, from, yeah, was handling TV stuff. But yeah. So that's, where I started my first company. And, like, seven years ago, I, you know, I was always interested in AI. Like, from my childhood on, like, talking with the computer, like, what we do now, chatting with Chechi, it was always my dream.
Eddi Weinwurm:Yeah.
Kevin Horek:So That's awesome.
Eddi Weinwurm:Yeah. So, even at university, I I took some subjects back then about AI and so on. So, yeah, so I grew. This company, it was quite successful, but I I always wanted to go into the AI direction. And my partner, Spectan, seven years ago, did believe in this issue.
Eddi Weinwurm:You know, there was Interesting. AI will never be I can never and AI blah blah. So yeah. And I really had this problem, that I I I didn't want to administer a company. I I'm a creator.
Eddi Weinwurm:Yeah?
Kevin Horek:Sure. Yep. I get it.
Eddi Weinwurm:So I I stepped down as CEO, and sold my shares and focused on AI before everybody else jumped on the train. Yeah. And it was quite difficult, back then to to to convince people to join and so on. Yeah. And now this, big AI explosion is happening, and we were prepared, with our research, with our technology to take off from day one.
Eddi Weinwurm:And yeah. So it's it's funny because, you know, all these, people who always doubt, AI and very, very, well, laughing about it, they are now AI experts. Over there, they became AI experts because they can write some prompts. Yeah. Mhmm.
Eddi Weinwurm:So what we do in in our companies, we we we do really hardcore AI research, and, yeah, because we have the knowledge to do it. Yeah.
Kevin Horek:Okay. So let's dive a little bit deeper into how does it actually work, and how do companies leverage your technology?
Eddi Weinwurm:Yeah. So, our big technology is to, think against the grain or go against the grain because, it's, you know, every AI these days is happening in the clouds. So, all these models are cloud APIs. You know, when you chat with, chat GPT or you have some Google APIs, it's all happening in the cloud. And this this is actually the easy way to go because if you want to improve your model, you just add new parameters.
Eddi Weinwurm:You make it larger and larger. So, AI is currently, like, exploding in size, and every model is is bigger and bigger. That's the easy way to go. But, what what what I discovered early on is, like, you know, cloud is nice and great, but in a lot of lot of cases, you don't want to use cloud services because you have really sensible data. There there are use cases where there's not even an Internet connection.
Eddi Weinwurm:You know? It's it's like a brick firewall air gapped. Yeah. So, of course, there is also in this case, you also want to use AI. So what our first product, is, for example, Caravan, which I placed in the market where I come from because I really understand this market very well.
Eddi Weinwurm:It's, audio visual AI. Yeah. So used by, for example, major Hollywood studios. Why it's important for them? Because their production environments where the edits, the movies
Kevin Horek:Right.
Eddi Weinwurm:Is completed. They have cameras watching the editors. Mhmm. They record the screens of the computers and so on. They have a a a security like the NSA.
Eddi Weinwurm:Yeah? So, of course, it's impossible there to connect to the Internet or do anything, like, with this big model. So, we came up with solutions which can run, the build this, very deep, machine learning, stuff and tools, really on prem isolated from the Internet, all self contained in hardware we ship. So that that they don't need to use, really, like, a supercomputer there, but it can run on on on normal hardware. And and that's the success we're currently having.
Eddi Weinwurm:Yeah? So media production, film production, TV studios, but also, I don't know, big retail chains, which which use it for customer interviews and so on. So everywhere where you want to keep the stuff safe and secure and, and and, on premise. Yeah. That's where we fit in because it was really hard engineering to make the models a lot bigger, but smaller.
Eddi Weinwurm:Sure. While, getting them, yeah, optimized.
Kevin Horek:No. That's that's very cool. So how do okay. So, like, if I'm in a movie studio and I'm editing, for example, how am I using your technology in that process? Or, like, how do I, like, how do I leverage your tech with we you know, in as a customer?
Eddi Weinwurm:For example, the simplest, the the simplest example is, you want to edit this interview now.
Kevin Horek:K.
Eddi Weinwurm:So, and then you remembered I told something about my childhood, in Vienna.
Kevin Horek:Okay. Yeah.
Eddi Weinwurm:You have no idea how I formulated this. Yeah? What words are used? Yeah. So So we're talking about Vienna now again, but probably that's not what you're looking for.
Eddi Weinwurm:Yeah? So, there you can search with this context of our search, and we are talking about searches in really, really huge archives. Yeah. We have, for example, years of archive or fifty years. Yeah.
Kevin Horek:Right.
Eddi Weinwurm:Searching there really with all this soft skill, like an assistant would for a footage by describing it, but describing it like, really, like, an emotional phrases. Yeah. Not what, will happen, what you really want to see, but what feeling it should represent or searching in in in interviews. That's, for example, one fractional idea we have. Yeah.
Eddi Weinwurm:And, you know, when you search, for a fast car, you usually in in in media production, you want to illustrate something, and you don't want the fast car parked in a clip in front of a restaurant. Yeah.
Kevin Horek:Right. Yeah.
Eddi Weinwurm:Yeah. So being able to search with, like, a human description, that's one major feature we have. And then, of course, we have features which just help the editors. They help, like, to predict how this story can visually continue.
Kevin Horek:Okay.
Eddi Weinwurm:And then you have different options, and you don't have to, you you don't have to, like, go through your material and find the right footage, when, the AI can already help you. So, you know, that that's really, there's a hot topic in media production, of course, that AI takes jobs. And, we really take care in our product development that we develop something which is, actually taking away the boring part.
Kevin Horek:Yeah.
Eddi Weinwurm:So, you know, like, when you edit this interview to shuttle around and to search for certain, it's super boring. Yeah. Super annoying. Yeah. So we, have a lot of tools now included, which just take away the boring part that the creatives can really focus on on on on on what's fun.
Eddi Weinwurm:Yeah.
Kevin Horek:No. Yeah. I I % agree. Like, I've been using AI for basically everything these days. Right?
Kevin Horek:And it gets you a certain part of the way maybe down the road of an idea, but you really need to execute it or you have to constantly, like, walk it through step by step to do certain tasks for you. And it's not really any different than starting with a template back Yeah. Like, before AI. Like, it's just the next evolution of that. And I think you're right.
Kevin Horek:It gets rid of those, like, mini minimal or, like, kinda boring parts of the job that I don't wanna do. And well, nobody wants to do. Right? So I can focus on the, like, the the better and more fun stuff. Right?
Kevin Horek:And I think we're so far away from it really eliminating people's jobs in my opinion. Sounds like you kinda share the the same, opinion.
Eddi Weinwurm:Yeah. The I mean, the the the, I see there are two approaches to it. Yeah. Of course, you can create AI which acts as like a person and which is let them in competition to you. Yeah.
Eddi Weinwurm:Yeah. But there is this other perspective which I prefer, and that's where I also develop my products too. It's what I call symbiotic AI. Like, it's it's enhancing you. Yeah.
Eddi Weinwurm:It's advancing you and your abilities that you can even compete with an AI, and and you can compete on the market. Yeah. So these are two different aspects. Of course, some companies will replace chops with this, really acting AI, but I think if we use AI clever, and and and develop a clever AI, which can be really built at the symbiosis with us. Yeah.
Eddi Weinwurm:Enhancing our perception, enhancing our thoughts, and enhancing our expression. Yeah. I think that's really what we should aim for, and that's, really what I can recommend, to to everyone. Yes. Start using AI.
Eddi Weinwurm:Start using it to be really, in a symbiotic relationship and get support and get the best of it. Yeah.
Kevin Horek:Yeah. No. No. That makes makes a lot of sense. So do you wanna maybe give us a few other use cases in other industries of how people are leveraging your technology?
Eddi Weinwurm:Yeah. So, of course, we said we started in this one vertical in the media and entertainment industry, and then we soon discovered, oh, this problem, yeah, it it exists in a lot of markets. Yeah. And, honestly, you know, everywhere where you deal with trade secrets, internal communication, contracts, and so on, that's where you don't want to send it to a third party. Because not that you don't need to trust a third party, but, you know, it it it all can be hacked, all can be abused and misused.
Eddi Weinwurm:And, you know, you have a lot of regulations. We have, for example, in Europe, this GDPR data protection and so on, and this all is a is an administrative administrative hell if you if you really interact with third party services. So yeah. But what what we offer, is, for example, you know, in in in the health industry, you know, patient datas are the most sensible data you can have. Yeah.
Kevin Horek:Yeah.
Eddi Weinwurm:And, there, of course, the AI can be quite useful or, like, monitoring of of a facility. Yeah. You you have a camera which is watches which watching which is watching your top secret research lab. Do you actually want to livestream this? Yeah.
Eddi Weinwurm:Of course not. Yeah. So, I think, you know, with every new technology, we have these two phases. The first phase is, yeah, it flies. This this, Wright brothers, okay, it flies.
Eddi Weinwurm:And we don't care the first car used. I don't know how many gallons. Whatever. So we didn't care. But soon after, then the phase of, engineering and optimization starts because then it's, the question, okay.
Eddi Weinwurm:We we cannot just pump more gas into it. Yeah? Yeah. Or or really build I don't know. Now now now we're building atomic power plants just to run these AI models.
Eddi Weinwurm:Yeah?
Kevin Horek:Yeah.
Eddi Weinwurm:So we should start to to to go into this engineering and say, okay. How can we optimize it, that it's really interesting for business?
Kevin Horek:Yeah. Interesting.
Eddi Weinwurm:Of course, in this optimization, we can also go further and say, okay. Can we optimize it that much that we can run it on prem on location? Yeah.
Kevin Horek:No. Fair enough. So how do you basically get in front of a customer and then figure out what their needs are? Because that health care space that you just obviously, there's a million things you could do in the healthcare space. Right?
Kevin Horek:So how do you, like, kind of because people are I think people are still scared of AI, and then you're going into something where it's, like, very, like, patient data. There's a bunch of laws around this stuff. So, like, how do you go in, sell your services to these, you know, big organizations and governments, and then still be able to create, like, a product out of it?
Eddi Weinwurm:Yeah. Yeah.
Kevin Horek:And service.
Eddi Weinwurm:So, honestly, it it it it we get approached. You know? The people see what we do now now in the media business or what we achieved in the media business. Yeah. And for example, we have one big retail chain here in The US, One of the biggest, actually.
Eddi Weinwurm:We started to use our product in the media production because they have in house media production, social media production and so on. And then, for example, the customer research passes by and says, oh, what's this nice AI? We also want it. Because, you know, we have this endless Zoom interviews with with with customers where we check the market trends, and then we want to search, okay, peanut butter recipes or whatever. Yeah?
Eddi Weinwurm:So, you know, you want to this customer interviews where they check really the the the customer values. They are very they have to be very open because you want to have the customer talk freely. So, for example, then then the approach is, yeah, we this is a good idea. So it it's currently working this way because, you know, a lot of companies are missing the expertise. We are talking now about chief AI officer and so on, and then they are these these are future jobs.
Eddi Weinwurm:Yeah. So the they're they're really also huge companies lack, the the the real expertise, and then they often hire some young college kids. Our script kiddies would think, okay. I download some models from Hugging Face and I have my product ready, but they never experienced engineering and how much it really what's the difference between a prototype and a prototype?
Kevin Horek:Yeah. Yeah. Totally. Interesting.
Eddi Weinwurm:I mean, this really surprises me because I'm I'm I'm talking with huge companies. They sometimes really lack, understanding about the whole topic. Yeah?
Kevin Horek:So how do you bridge that gap, though, and how do you start educating them about AI and what they can and can't do with it? Because there's so much misinformation out there about what AI can do today, what the promise people are hoping for can do in the future. So, like, how are you actually working with these companies to actually say, like, here's what we can do, and here's hopefully what we can maybe do in the future? Yeah.
Eddi Weinwurm:Yeah. So, you know, when we have a a ready, you know, a really finished product like, caravana in the meat industry, it's it's I always prefer to make, you know, these POC installations, which is proof of concept. You know, they get something delivered, and then I say, oh, but, you know, don't upload as technician test material. Let the end users play with it.
Ad:Yeah. Then it's solved.
Eddi Weinwurm:Then it then I then I have it solved because the end users don't want to give it out of hands again. Yeah. So I get them hooked on it. Yeah. That's that's that's actually when you have really good product, the the best approach to say, okay.
Eddi Weinwurm:We ship a unit there. You we install it, and it's it's an end and then it's sold. Yeah.
Kevin Horek:Okay. So are you building, like, a physical box then that houses your technology? Is that what you just kinda said?
Eddi Weinwurm:Yeah. Yeah. So we don't build them ourselves. Of course,
Kevin Horek:we have
Eddi Weinwurm:partners who do this. Okay. And but, of course, it gets, preinstalled. That's what we call an AI node. And the nice thing on this AI node is you can cluster them.
Eddi Weinwurm:So, if, if the the the customers want to use this in other sectors or want to expand their dataset or whatever, they can buy additional, units and linear clusters. Yeah. So we want to keep it as simple. That's our goal, as simple as connecting to an API. And, I don't know if you played around with these APIs of of Yeah.
Eddi Weinwurm:The AI and so on. They're quite easy to implement. So
Kevin Horek:Yeah.
Eddi Weinwurm:Our goal is to to to to to come to a similar, yeah, setup. You know, it it's plug and play. Of course, it needs a bit of setup, currently. But, in the end, yeah, it's it should be as simple as possible. Yeah.
Eddi Weinwurm:That's that's the reason why we really ship this as boxes. Yeah.
Kevin Horek:Okay. So I wanna dive a little bit deeper into that. So you called it Carol one? Yeah. Okay.
Kevin Horek:So and and what does that actually do? Like, walk us through from what it like, becoming a customer to actually getting it and using it and the whole setup? Because I'm curious to know how the hardware and software kind of fit into that.
Eddi Weinwurm:Yeah. Yeah. So, most usually, that it works this way that, you have some assets, you have some archives or you have some production assets. Depends. For example, you have a Netflix show documentary.
Eddi Weinwurm:You you worked the film for three years. So, you you install our appliance, connect it to the storage. AI automatically starts to, analyze the storage, starts to watch these, clips, the assets which are there. And and then you can just, have a user interface, which is very intuitive where the, you know, the greatest then can quite easily work search for them getting, get, I don't know, recommendation how to continue stuff and so on. So, that's also again, you know, we want to keep it as easy to, yeah, to run.
Eddi Weinwurm:Of course, there are then use cases when we go really into big installations, like, we we did now for a major science fiction franchise, training on their universe.
Kevin Horek:Okay.
Eddi Weinwurm:Our models really learn to identify spaceship by parts or planets by their, trees and so on. Yeah.
Kevin Horek:That's cool.
Eddi Weinwurm:Yeah. That's, you know, for for for my, engineers and, data analyst. It was really fun because they love it. Yeah.
Kevin Horek:Sure. Yeah. Totally. Yeah. That's cool.
Kevin Horek:It
Eddi Weinwurm:it was highly motivating them. Yeah. So but, that's an additional services we provide. Yeah. So because, as I said, we are really these models we are having, we're really developing them ourselves.
Eddi Weinwurm:We are training them ourselves. We know how to to to to come up with new things. Yeah. So, even if they want to, have new, functionalities, AI functionalities implemented, we sit down with our engineers, and we we come up with solutions. So yeah.
Eddi Weinwurm:So, we offer more than just this plug and play box. If they really want, we can integrate this into existing infrastructures and so on or, come up with custom models, custom training, and so on.
Kevin Horek:Okay. No. That makes sense. So the hardware box itself, you you mentioned you partner with people. Like, is it basically just like a Linux computer or or kinda what's in it exactly?
Eddi Weinwurm:Yep. So, it's it's it's a Linux box. Yeah. So it's a Linux box with a strong, GPU in the end. Yeah.
Kevin Horek:Sure.
Eddi Weinwurm:So, you know, we we we we want to keep this also as efficient and, you know, as, yeah, as affordable as possible. We don't you know, it it's for the customer in the end, it's much cheaper to buy this. Even we would also offer this, like, as a as a as a service. Yeah? But it's it's much cheaper to buy this than to go into this use all these APIs because, you know, these cloud services, they, the costs, they are really, significant.
Eddi Weinwurm:Yeah. Yeah. Now imagine you have fifty years of archive. Yeah?
Kevin Horek:Yeah.
Eddi Weinwurm:Sending all this, the sunburn is gets really expensive. Yeah.
Kevin Horek:So, you know, I I understand agree with you. So roughly, what's the cost of a box?
Eddi Weinwurm:I'd say that between 40 and 50 k.
Kevin Horek:Okay. That
Eddi Weinwurm:that's that's the starting point. Yeah. And, I don't know if we have customers which then have, like, really a lot of these boxes, but
Kevin Horek:Right.
Eddi Weinwurm:We really build them so that they are easily to cluster, easy, comfortable, and, hot swappable and stuff like this. Yeah? Because we really that's that's the thing. Yeah. We want to bring AI down from the, yeah, it flies to, okay, here it it runs.
Eddi Weinwurm:Yeah.
Kevin Horek:Yeah. No. That that's really cool. So walk us through maybe some I see on your website you have, some government use cases. Do you wanna maybe give us some of those as well?
Eddi Weinwurm:Yeah. That's no. The the government use cases, they they are now, some projects we are having, but they are not really like customers yet. But it's quite complicated. Sure.
Eddi Weinwurm:I can imagine. Bureaucracy squared. But I don't want to complain. But the thing is, like most governments, they're really limited in AI news. And this creates a huge imbalance.
Eddi Weinwurm:For example, you know, patent lawyers, they are now allowed to use ChatGPT officially. Yeah. So they they they they research, okay, what does the what does, do other patents say about this and so on? And, you know, the examiners, they're not allowed to use it.
Kevin Horek:Right.
Eddi Weinwurm:So this creates a huge imbalance. Yeah. And then and this is in many sectors. Yeah. Like, it's it's even like, I don't know, for tax or for for everything.
Eddi Weinwurm:Yeah. Usually, government agencies, are not yet allowed to use it. Yeah. So there we see, of course, the market. Yeah.
Eddi Weinwurm:The government itself wants to invest now. I don't know if you're here. It is about 10, a hundred billions and so on because they see this problem. But, you know, if the government built something, it takes some time to it's really running and, yeah. So, yeah, these are these, yeah, the, different verticals we see, where we fit in.
Eddi Weinwurm:And in the end, we want to, expand to, yeah, to a lot of sectors. Also, which is easy to use, you know, we ship your hardware, you plug it in, and you get a product. Yeah.
Kevin Horek:No. That's that's really cool. So I wanna talk a little bit about what can you actually do with AI, and what are your kind of thoughts around AI right now? Because, like, I wanted some of the gadgets that were, you know, like, came out last year when they were promised, you know, a bunch of stuff, like, I really wanted some of that stuff to be true. And I think the perfect example is, like, the rabbit r one.
Kevin Horek:Like, I wanted that thing to be what they said it was. Yeah. I pre ordered one. I got it, and it's the biggest piece of garbage, and it's sitting on my desk collecting dust.
Eddi Weinwurm:Yeah.
Kevin Horek:And so to me, that was the perfect example of something that, like, I wanted it to be true, but it's completely nowhere near what was demoed on stage.
Eddi Weinwurm:Yeah.
Kevin Horek:And I think a lot of people kind of have that. Like, they're, like, they're terrified of it because of, like, what's promised, but the reality is so disconnected. So can you maybe talk about where we are? Because you've been in the space a long time and and
Eddi Weinwurm:what your thoughts are. So that's actually the the future vision I have with the company because, you know, we have Moore's law that the hardware gets more and more performant.
Kevin Horek:Yeah.
Eddi Weinwurm:And this is now especially through in in all these, GPU applications. And on the other side, we are working against this because we try to make this thing smaller. And I think one big problem we currently have, and even if I use, like, chat GPT, the voice chat, Yep. You know, you will have this really awkward delay of replies, and you're sharing the resources, this model with millions of others. So you can sometimes wait really long for the reply.
Eddi Weinwurm:And, actually, you know what I would what I'm dreaming of glasses, which where I'm while I'm looking point at different things, like see, or even, like, darken down some parts of the vision and, enhance, the point where I'm where I should look for. So, you know, this part of AI, which is really, like, responsive Yeah. And, and therefore can just run on your local device. Yeah. So, I think that's that's a lot of, that's a huge problem in the experience.
Eddi Weinwurm:Yeah. Because Yeah. If I have the smart glasses, which I really love. Yeah. And I really wish I had it, like, soon in a working.
Eddi Weinwurm:Right? Yep. Yeah. Yeah. But this responsiveness and you you're developing UX.
Eddi Weinwurm:Yeah. You're a UX developer. You know how important responsiveness is. Yeah.
Kevin Horek:%. It it it's almost like the number one thing.
Eddi Weinwurm:Yeah. So and and there I see the crossing point. It's also a question of privacy because, you know, when I have this symbiotic what I call symbiotic AI, I want to have it always on. Yeah? Yeah.
Eddi Weinwurm:It's like I don't want to give my brain when I when I enter the house, I don't want to put my brain on the on the, yeah, on the cupboard.
Kevin Horek:So
Eddi Weinwurm:but I also don't want to stream what's happening in my bedroom, you know. Yeah. And there we again at this point, yeah, but we say, okay, it's nice to talk with Chat GPT about business and whatever, but, you know, when it gets really about private issues and when maybe really this AI starts to watch my daily life and it's on twenty four seven Yeah. Want to have it on me. Yeah.
Kevin Horek:Yeah. Yeah. That's interesting. Well, the other thing too is I think you also don't like, especially if you're wearing glasses, you don't want them like, again, it's kind of like Google Glass was, like, the first kinda thing with that. But if you wore those out in public, everybody was, like, so uncomfortable.
Kevin Horek:Right? And you also looked kinda ridiculous. So you need something that looks just like regular glasses. And, obviously, some of this like, some of the the amount of actual surface area where you can even stick a battery is basically nonexistent. Right?
Kevin Horek:Unless you connect a cable to a battery pack like kind of the Vision Pro does. So we're kinda, like, it's we're in a weird place, right, where the tech's not really there yet compared to work with the hardware.
Eddi Weinwurm:But there is a lot of, investment going on in this sector, and I think we will be there. And I I don't know. I I wish you know, I I I really bad in remembering names, and I'm on a lot of conventions. Yeah. I always need my assistant next to me, which whispers in my ear, oh, this is not Tom, whatever, and asking me about his, children.
Eddi Weinwurm:Yeah. Right. So having something like this Yeah. Of assistant, I would love it. Yeah.
Eddi Weinwurm:And and, you know, just a little plug in the ear, which is transparent to to to the normal audio and gives me some, enriches my perception. Yeah? Yeah. This alone would would be amazing.
Kevin Horek:Interesting. Yeah. No. I % agree. So as somebody that's building hardware for AI, like,
Eddi Weinwurm:how
Kevin Horek:and I know I've seen what, like, kind of Facebook and Meta are doing with their glasses and the snap glasses and and whatnot. And all these like, I played with the Vision Pro. Like, how soon do you actually think we are out from that? Are we three years, five years, ten years? Do you know roughly?
Eddi Weinwurm:I would say, you know, it depends what what you what you demand. Yeah? So if you want to if you want to really run a a GPT four model on your phone, it this takes at least more than ten years. Yeah?
Kevin Horek:Okay. What
Eddi Weinwurm:we do for example, what's our trick? Why how we get our AI smaller is, because we are using compounds instead of monolithic models. It means, like, for example, this media solution we have there, like, now over 10 different models.
Kevin Horek:Okay.
Eddi Weinwurm:Which makes sense because, you know, when you ask, about, cookie recipe, Yeah. Just to pass through all these layers with all this information about quantum physics, Shakespeare, and so on just to give you the the right amount of sugar to add. Yeah?
Kevin Horek:Right.
Eddi Weinwurm:Waste. Yeah. So, we use, for example, expert models. Yeah.
Kevin Horek:Okay.
Eddi Weinwurm:I plan on models, which, or routing models, which say, okay. This model is handling this and that. Yeah. So climate engineering in that. And, what I see earlier is in hybrid.
Eddi Weinwurm:Yeah? Hybrid. So that you say, okay. You have some models running locally.
Kevin Horek:Right.
Eddi Weinwurm:And then you have, if you it's really the heavy lifting, reasoning, it kinda happen in the cloud. Something like that.
Kevin Horek:This. Okay. Kinda what Apple's trying to do with Apple intelligence and failing miserably at it. Yes.
Eddi Weinwurm:Kind of. But, that's that's what I what I I meant before. You know? It's easy to build a prototype, and it's fucking difficult to to build a product. Yeah.
Kevin Horek:Yeah. No. That's that's fair. And it's interesting because, like, if a big company like that is struggling to build, you know, simple things, then, you know, like, but the caveat to that though is you can build some pretty incredible things with it if you kind of walk it step by step through, you know, prompts. Right?
Eddi Weinwurm:Yeah. Yeah. So I did end it. It's it's all engineering, and you can you know, a lot of companies are now, like, you know, the thing, like, okay. One, woman needs nine months to to give birth to it, to a baby, let's hire nine women and get it done in one month.
Eddi Weinwurm:Yeah? So that's that's the right philosophy philosophy. Yeah. Yeah. So just, throwing a lot of of clever scientists onto a problem, doesn't give you a product.
Eddi Weinwurm:Yeah.
Kevin Horek:Interesting. Yeah.
Eddi Weinwurm:They are more reacting than acting. Yeah. And our advantages, because we started seven years ago with all this, you know, we are acting. And we were the first to, for example, introduce this semantic search, in in in the media sector. Yeah?
Eddi Weinwurm:Yeah. You know, before everybody else. And all the others, the big companies like Adobe and so on, they had to react. And
Kevin Horek:Yep.
Eddi Weinwurm:And that's the state of the industry. And then you get something like, you know, that these products which are half baked monsters. Yeah.
Kevin Horek:Yeah. That's that's interesting. What what about the, the AGI stuff that we keep reading about? What are your thoughts around that?
Eddi Weinwurm:AGI, like a general intelligent.
Kevin Horek:Yeah. Yeah. Yeah.
Eddi Weinwurm:I don't know. I if you ask me as a philosopher Yeah. Okay. No. I think it's coming.
Eddi Weinwurm:Yeah. And I think, we humans are not so unique. Our language is not so unique. You know, that that's that's really the the weird thing that, language models were much quicker, trained and then much quicker it did show up much quicker in research than, all these image models. Yeah.
Eddi Weinwurm:We really understand visually or robots are still bumping against stuff. Yeah? And language models writing academic papers. Yeah? So we we we understand that maybe, yeah, of course, we are the best, which is known to us as humans.
Eddi Weinwurm:But, where should the why should there be a limit, of of intelligence which can't get better than us? And I'm thinking I I'm I'm thinking, we're heading there. Yeah? And, probably, you know, with the next generation of, OpenAI models, we will have really, PhD, reasoning. And yeah.
Eddi Weinwurm:So we have Interesting. If it's good or bad, I don't ask me. It's, weird. I didn't I didn't know when I when I had Yeah. When I had this wish, I want to talk with my computer.
Eddi Weinwurm:I didn't expect this. Yeah. That's, yeah. But I'm living this dream actually. Yeah.
Kevin Horek:No. That's that's really cool. Do you have thoughts or advice for people to maybe stay relevant in their careers or things that they can do to make themselves more valuable? Because I especially in tech right now, it's it seems to be coming out of the doom and gloom a bit. But do you have any advice for people that are maybe looking to get into tech or are kind of panicked about all this stuff in tech?
Eddi Weinwurm:So, my advice is, the the wise, advice I give to my developers, use AI as much as possible. And, yeah, it's I, you know, that's I know software development from before, AI, and I know it now. And for me, the difference is, we are just much faster. We can achieve much more in the time. Yeah?
Eddi Weinwurm:That's Yeah. Changed. But my team size didn't change. Yeah? So I don't know how it will be in the future.
Eddi Weinwurm:I I don't know. But maybe, maybe, AI plus human is still of more value than AI without human. Yeah?
Kevin Horek:So I I think so.
Eddi Weinwurm:Yeah. So having having this human creativity and maybe, even human, weirdness of creativity. And there is again where, you know, I mentioned in the beginning, I'm in between this world of creativity and mathematics. Yeah? Yeah.
Eddi Weinwurm:So, having this weirdness. Yeah? Additionally, as an input, maybe, helps, yeah, makes our the the the our product, what we produce when we work together in synergy with AI, maybe this little better that we can beat pure AI. Yeah. So I I can just recommend every everyone keep up to date, with AI.
Eddi Weinwurm:Use it, learn to use it, and, learn to handle it, yeah, to guide it. Yeah? And Yeah.
Kevin Horek:No. I I think that's actually really good advice. So other you've mentioned chat g b t. Is there any other AI tools that that you find useful in your day to day that you'd recommend people try out?
Eddi Weinwurm:No. For from I don't know. For me, I'm lazy in this way. I'm using the newest model, the one, and that's mostly what I use as a sparing partner. As I said, I I, you know, I learned to write.
Eddi Weinwurm:Yeah. Yeah. That's was my job back then. So I'm writing, but I I'm I'm using it to get feedback.
Kevin Horek:Totally.
Eddi Weinwurm:Using it as a sparing partner for ideas, that stuff. Yeah. So, I think we should just that's also what what I recommend to think more creative what you can do with it. Yeah. And not just, like, do this for me, but maybe ask me something.
Eddi Weinwurm:Yeah? Maybe make me make me reflect my buyers I'm having. Because of you know, in business, for example, as a CEO, I have to make decisions. Yeah. And
Kevin Horek:I Yeah.
Eddi Weinwurm:Else. Yeah. So turning it around. Yeah. That's, for example, what else I can recommend.
Eddi Weinwurm:So be creative also in the way you use it.
Kevin Horek:Yeah. No. I I think that's that's actually really good advice, but we're kinda coming to the end of the show. So how about we close with mentioning where people can get more information about yourself, obvious future, and any other links you wanna mention?
Eddi Weinwurm:Yeah. Obviously, yeah. On ObviousFuture.com. Yeah. One word, ObviousFuture.com.
Eddi Weinwurm:Or at me on LinkedIn, Eddie Wineworm. And, yeah. I'm always, happy to yeah. As you see, I'm I'm always happy to talk about it because I It's good. About this.
Eddi Weinwurm:Yeah.
Kevin Horek:Oh, very cool, Eddie. Well, I really appreciate you taking the time out of your day to be on the show, and I look forward to keeping in touch with you, and have a good rest of your day.
Eddi Weinwurm:Thank you.
Kevin Horek:Thank you. K. Bye.
Intro/Outro:For more episodes, you can find us on all major podcasting platforms or visit Iambuildingthefuture.com. Join us next time and keep building the future.