527: Exploring AI in Business with PrimeLab io’s Wendell Adams

Host Victoria Guido welcomes Wendell Adams, CEO of PrimeLab.io, as he talks about his lifelong passion for technology and entrepreneurship. Wendell shares his experiences, from hacking electronics as a child to studying various fields in college and eventually starting his own business. He emphasizes the importance of understanding market needs and leveraging language to make technology accessible. Wendell's drive to improve encryption and data security led to the formation of PrimeLab; a company focused on making encryption functional and accessible without compromising performance. Wendell discusses PrimeLab's strategic direction and market fit. He outlines the challenges and opportunities in the entertainment industry, emphasizing the need for innovative solutions that respect user control and privacy. Wendell also shares insights into how PrimeLab's technology can democratize data access and enhance business processes. The episode concludes with a reflection on the future of AI and encryption technologies and Wendell's advice for aspiring entrepreneurs to think critically and creatively about their ventures. PrimeLab.io Follow PrimeLab.io on LinkedIn, or X. Follow Wendell Adams on LinkedIn. Follow thoughtbot on X or LinkedIn. Transcript:  AD: We're excited to announce a new workshop series for helping you get that startup idea you have out of your head and into the world. It's called Vision to Value. Over a series of 90-minute working sessions, you'll work with a thoughtbot product strategist and a handful of other founders to start testing your idea in the market and make a plan for building an MVP. Join for all seven of the weekly sessions, or pick and choose the ones that address your biggest challenge right now. Learn more and sign up at tbot.io/visionvalue.  VICTORIA: This is the Giant Robots Smashing Into Other Giant Robots podcast, where we explore the design, development, and business of great products. I'm your host, Victoria Guido. And with us today is Wendell Adams, CEO at PrimeLab io. Wendell, thank you for joining us. WENDELL: Thanks for having me. So, question, actually, where'd you guys come up with the name? VICTORIA: You know, I have asked this before, and I think I remember the answer. I might have to go back to the 500th episode to get it, but I think it was just robots was already kind of a theme at thoughtbot. I mean, thoughtbot, obviously, has robot in the name. Joe might have the best answer. And we have our special co-host, Joe Ferris. Who better to answer? JOE: [chuckles] Yes, I'm not sure who better to answer, probably Chad. I don't remember the answer either, but happy to be here to speculate with the two of you. It comes from the blog. We named the blog Giant Robots Smashing Into Other Giant Robots and then used it for our podcast. But I don't remember where the blog name came from. WENDELL: It kind of reminds me of the Robot Wars thing, like, where they would have competitors driving around the robots and then smashing into each other, trying to flip them over and disable them. JOE: That was excellent. I also watched that. WENDELL: [laughs] VICTORIA: Yeah, it's a pretty great name. I really enjoy being a host. And, you know, I go out to local San Diego events and meet people and introduce myself as a co-host of Giant Robots Smashing Into Other Giant Robots. It's usually pretty funny [laughter], which is where I met you, Wendell; we met at a San Diego CTO Lunches, which was super fun. WENDELL: Yeah, I always enjoy any type of tech conversation or anything else. I thought that was a lot of fun to sit down and just talk with people and talk about what they're working on. VICTORIA: I love that, yeah. And before we dive into the tech and get to hear more about PrimeLab, I just want to start a little more socially question. What did you do last weekend, Wendell? WENDELL: It was my father-in-law's birthday party at Legoland. We took my daughters my mother-in-law, and we all went to Legoland. It was a lot of fun. Although, honestly, I prefer the San Diego Zoo over Legoland, so... VICTORIA: Can you please describe what Legoland is to people who may not know? WENDELL: Okay. Legoland is based in Carlsbad, and it's really ideal for, like, four to nine-year-olds. And they have, like, miniatures of all the different cities. Actually, the SF miniature that they have is crazy detailed with Chinatown and everything else. They did an amazing job there. They actually...I think they just redid the San Diego part of it. But the miniatures are really cool, seeing all this stuff. They have different rides performers, but it's definitely, like, one of those things that it's more for kids to go and kind of experience. If you're an adult, you're going to love a lot of the processes that go into place, like how they built things, but mostly, yeah, it's very much kid rides and stuff like that. VICTORIA: I imagined it to be, like, life-size Lego buildings, but maybe I'm...that's very interesting all those other things you could do there. WENDELL: Well, like, they have the One World Trade Center, and I think it's, like, 25 feet tall. It is, like, the replica of it. It's kind of interesting, too, because not all the Legos that they build, they're huge, are solid Legos. So, it's like, they'll do where it's like, on the outside, they'll do a base, and then they'll build it. There's a replica of a Lamborghini. That one's life-size. But it's heavy. It's, like, 2,000 pounds, something like that. VICTORIA: Is that as much as a regular Lamborghini weighs, too, 2,000 pounds? It can't be that far up. WENDELL: I don't know. No, I don't think it...no, it couldn't be. VICTORIA: I have no idea how much cars [laughs] weigh. What about you, Joe? Did you do anything fun this weekend? JOE: Not a lot. It was supposed to be my son's first soccer game ever, but it rained here in Boston, so they postponed it. Sunday he went to my parents' house for a grandma day, and so I did nothing. I ate cookies. WENDELL: [laughs] VICTORIA: Wait, what kind of cookies were they, though? JOE: They were chocolate chip cookies. VICTORIA: That's so good. JOE: They were good. They were brown butter chocolate chip cookies, I should say. VICTORIA: Were they homemade, or did you get them somewhere? JOE: They were. We made them in this home. VICTORIA: Oh, that's the best. Yeah, love that. I got some fancy cookies that someone else made, and they were also [laughs] very good. And then, yeah, I've just been having cookies pretty much every day. So, that's been my time. WENDELL: My mother-in-law recently made me peanut butter cookies, and those are my favorite kind of homemade cookies. VICTORIA: Okay. Noted. You'll get a post-podcast gift of peanut butter cookies [laughter]. I love that. It's so great to hear a little bit more about each of you as, like, in a personal way before we dive into AI. And tell me a little bit more about your background and what led you to PrimeLab. WENDELL: I've always kind of, like, been a hacker, so to speak, just from a technical standpoint. My one grandfather was an engineer. He worked for GM designing, like, assembly arms and stuff like that. And then my other grandfather was a master electrician. So, I've always been the person that, like, just worked on things, got stuff together. You know, there's a lot of stories. Like, there's the story about when I broke my grandmother's workbench, rocking bench out front, and it was all aluminum. I remember telling my grandfather, and he's like, "Oh, what are you going to do?" And I was like, "Buy a new one?" He's like, "You got money?" I said, "No." And he said, "Well, you better figure how to make it then." So, ironically, it's half aluminum, half wood. We took wood, sanded it down, and stuff. So, it's just like I've always been an entrepreneur. I've always been interested in this kind of stuff. I used to hack VCRs, and PlayStations, and all kinds of stuff. I always liked parts and components and rewiring things. And as I got older, I also really liked math and all those things. And I wanted to understand more about how the world works, so to speak, like why it works the way it does, not just from a technology standpoint. But why do people think the way that they do? Why do things behave the certain way they do? So, initially, I started going to college. I thought I might be a math professor, and then decided to get degrees in business, economics, finance, marketing, consumer product goods, and comparative religions. So, while I was in college, I started working on, like, hacking, different video games, writing JavaScript, writing Java, all kinds of stuff. And then, eventually, even writing mobile applications early on, and then just analyzing because I always liked to build phones, too. I would take apart phones. And I really was curious about, like, how to make things faster, more efficient, and better. So, now to bring it down, like, how to make things accessible, where it benefits some of the smallest people and make it where it's a greater opportunity for someone to come out ahead of something. Like, one thing that I learned from my marketing degree is language matters. So, it's like, all the marketing it's not anything special. It's just they intentionally create language barriers that cause people not to feel as accessible with it. And then, like, you hire a consultant or something to just basically teach you about those language barriers. And I think every industry has, like, SAT, or LTM, or something like these abbreviations that mean a lot of different things. And it causes bottlenecks if you don't speak the language. So, understanding the language but also learning about how was very helpful from a standpoint on the marketing side. And I always try to figure out how do I make this accessible to people who don't understand that language? VICTORIA: And what was the turning point where you decided to start PrimeLab, and what made you realize there was a company there? WENDELL: It was a project I've been working on since at least 2011, honestly. And just as a heads up, PrimeLab as a whole works with encrypted data for AI models and to speed that up and everything else. So, early on, I was very obsessed with how advertising works through, like, stealing user data, which stealing is different, here or there, the sense of privacy, the sense of, like, how things could run, and the sense of messaging. And initially, a lot of it was using encryption as an overlay in, like, the pixel application space, which is always a way to hack or get into it. And it slows everything down. So, I had always been working on trying to figure out how do you speed up and embed security so it's actually functional? And it took a while to figure out, like, give encryption functionality, like, make the encryption something that you could actually execute on. And, actually, one of the things that really helped is the blockchain space there's a lot of, like, hash trees and everything else, like, where people are innovating in that. That's really helped innovate encryption as a whole from understanding, like, Merkle trees, hash graphs, and everything else to make it more functional and faster. Because people are trying to speed up distributed networks and stuff, but the actual technology that they built, like Hedera is...What Hedera has done with Hashgraphs and everything else—really amazing. I'm glad that they open-source stuff like that. But it's also really interesting just to see how things push forward. So, like, when I first started, like, RAM was, like, 256 in a phone. So now, you know, you can get multiple gigabytes, which makes it a lot more capable to do encryption, decryption, and work more in the functional space of things. The bigger problem that you have on the data part is how an application communicates because there's so many levels of abstraction. Like, you have the Swift language that communicates into something else that then communicates into something else. Like, right now, we're talking on a system that's recording us over the internet through a browser, all those different things. And it's an approximation of what the data is and what we sound like. It's not an absolute. So, I was really interested in when you have absolutes, and you can verify those absolutes, what can you do with that? A few years ago, I felt like we got to a point where we could actually execute those things and actually deliver on that. So, therefore, I decided to start PrimeLab with my co-founder, who I really liked and enjoyed. And we've had a lot of really great advisors, where people have helped us continuously. Over, you know, the decade-plus of working on this, I've gotten a lot of input from some of the smartest people I know, from people who have designed full server racks for AWS to literally a good friend of mine that built cloud storage. His name's on the patent for it. So, that kind of stuff has really helped me understand and build this where it can communicate the lowest possible level. VICTORIA: Yeah, and to just recap and reflect that back a little bit, it sounds like you were always interested in how to make encryption faster and lighter weight, and so you could build it in and build in security without impacting the performance of the applications. And then meeting your co-founder and the advancement of technology, this time a couple of years ago, led you to think, okay, let's really go forward with this. WENDELL: Kind of rephrasing, I was always interested in control. So, like, one of the things that really interested me...so, I started a video game store buying and selling, like, video games and trading cards and stuff when I was roughly ten and a half or so, and then sold it roughly when I was 17, which is how I paid for quite a bit of college and likewise. But the things that really interested me about that is it went out of business three to four months afterwards because the person who basically bought the rest of it bought too much of Madden. And Madden, at this time, the margins were, like, a buck, as you go all the way through, and the price drops immensely. So, I wanted to really understand why that happened. What you kind of get to is, like, they didn't have control over it, just, like, the bulk orders methodology, where they would buy the whole entire supply. And what I've seen over the years, be it Apple, Google, or anything else, is, like, that was...in that example, that's a game publisher, EA, flexing control, right? But more and more companies are flexing control on a platform like now with Facebook or advertising. If you think about what Google used to do, Google used to provide a lot more insights when you had your own website. You used to know your own keywords. You used to know a lot of things about your users who come through. More and more, Facebook and Google try to stop that. And they're really the ones determining your own user personas for you. So, you become dependent upon them. So, I wanted to say, okay, from a business standpoint, how do you implement control and privacy where it's permissioned? And encryption was one of the answers that I came to. But then it was, how do you make encryption functional then to actually execute on control? Because unless the system is secure, faster, cheaper, better, it's never going to get adopted. VICTORIA: That makes sense. Thank you for sharing that. And you mentioned your founder. I'm curious, how does your founder kind of complete what you needed to be able to get the business up and running and off the ground? WENDELL: He has a robotics degree, so he had launched several products that had failed. And he wanted to learn marketing after they had failed. So, we have a similar like mindset about, like, control and functionality for how something may or may not work, and that allowed us to communicate well. So, like, I have a lot of friends and stuff. But the thing that allows me and my co-founder to work really well is that we come from things in different angles, but we have the same language that we speak. So, like, that's what I was talking about before, like, LTMs or otherwise, like, language really matters from how you can move something forward when you're talking in different industries. And just with him, there's a lot of stuff that you don't have to say. You can skip a lot of filler and then go straight to what something might be or a solution or something. Or if we have to jump to a tech abbreviation, to a market abbreviation, to a financial abbreviation, he's one that can follow along with me really quickly and then teach me a lot of things about operational execution because he's great at operations. I am not great at operations. VICTORIA: That's really interesting. And I think you're making a good point about, like, a shared language. And it reminds me of any product that you're building; if you want to sell it to a company and you want them to adopt it, you have to consider their language, their belief system, how to influence change within the organization. And I wonder if you could talk a little bit more about that with your experience at PrimeLab. WENDELL: I'll give you an example of a market that we decided to go after. So, instead of just working at, like, healthcare markets where you have, like, GDPR...for people who don't know GDPR or HIPAA, HIPAA is for the United States. GDPR is the EU privacy requirements, right? For the right to be forgotten and everything else. So, these are vernaculars that you need to know. But the requirements of each one is very different, and these are markets that we've learned being in tech and likewise. But we wanted to change it up. So, I wanted to go after the entertainment market as a whole, namely because after meeting with some select people, including a stunt man, this is going back a few years ago, I started to realize that the entertainment market was getting kind of screwed over quite a bit from a tech standpoint. Basically, tech goes through this thing where...someone wrote a great article about this. It's called Enshittification. But, basically, where they go they try to take over a whole entire market, where first they're providing great value to your users. And then, gradually, you enshittify your product to provide greater value to your investors. And then, gradually, you suck all of the value out of the room for both. Right now, if you look at Sora, what OpenAI is trying to do in entertainment, [inaudible 16:08], you kind of can see that happening. They're going, "Hey, here's a great value for it." And they're really pushing that stuff off. But the thing about the entertainment market that I think is really interesting is it's basically thousands and thousands of small businesses that are constantly going, it's so chaotic. It's not like tech and startups. There's a lot of overlay of, like, you know, people are looking for that top quartile film that's going to make the money back, and then long-term royalties that they can earn off of it, right? Whereas in tech, they're looking for those huge markups as well. So, I was really fascinated by it, but it was something that, like, we had to learn. Like it was something that I didn't know otherwise. So, it was literally...how we learned it was we took our tech stuff, and we would walk SAG-AFTRA strike lines. We would walk strike lines. We would go to entertainment events, and we would demo what we were trying to do, and we would show them. And then, oftentimes, we got really negative feedback right off the bat. And we're like, "No, no, no, so, you know, this is for you. Like, you could control. Like, this is going to help you." And then, after doing that enough times, talking to the SAG-AFTRA lawyers, and everything else from there, and all of the creatives, the creatives were coming to us and giving us ideas how to explain it because there's, like, three different formats. You have tech, business, creatives in the entertainment industry. And it's like, we could talk to the tech people. We could talk to the business people. But you really need the creatives. And, like, the wording of each one, like, each group of those is vastly different. So, having the creatives be able to explain something in 90 seconds that used to take me a couple of hours to dive into became really valuable. And also, in tech, like, you have this thing where it's feature creep, where you're like, oh, I'll add this, this, and this. Just to hear very coldly and bluntly, like, "If it does X, I'm interested. If it does Y, I'm not interested." That was very interesting or refreshing of, like, "Yes, you're going to solve these problems. But I need sign-off for everything in there." And it's kind of weird in the entertainment part, too. Like, you want to solve a problem without being a competitor to another vendor because you need so many different sign-offs. And if you're a competitor to another vendor, to a certain point, maybe that's going to cause a hiccup with sign-offs because there's 18 different cooks in the kitchen, so to speak, just so many different people that need to say, "Yes," all the way through with it. VICTORIA: Thank you. Yeah, that's really interesting. I'm curious, Joe, if you have an answer for that question as well, like, any experiences about navigating change and putting new products in place at different clients, different industries? JOE: I don't think I've had the same kind of resistance. Like, I haven't been on the front lines the way you described, like, literally in the, you know, going and talking to people on strike. I think I have more indirect experience talking to the people who are doing that. And certainly, like, I think there's generally a resistance to bringing in new technology without eliminating the old way of doing things if that makes sense. Like, people want the old ways of backup. Like they want to be able to go back to paper, which I empathize with. But that's frequently been a challenge for the people I've worked with is that they don't fully embrace the new process, which significantly reduces the value they would get from using it. I don't know if that's something you've encountered with PrimeLab. WENDELL: So, we were building another company of mine many, many, many years ago. I was building a website for this lumber company, and I remember showing up, and the owner was there. But it was his son that had commissioned it, and the owner didn't know about the website. And I was like, "Oh yeah, we'll get the website going." He goes, "Oh, this web thing it's a fad. It's never going to happen. You don't need websites. It's faxes." That's how everything would happen. But secretly, what was happening is they would get an order. They would print it off, and then they would fax it. So [laughs], I always thought that was crazy. VICTORIA: I mean, one of my local bars still just writes the order on a ticket and sends it on a clothesline down to the grill. So [laughs], sometimes old is good. But I think that you know, I want to hear more about where you found or how you found a product-market fit for PrimeLab and where that AI really becomes useful and ethical in the industry you're focusing on WENDELL: How I look at PMF (product-market fit)...and if you hear me just say PMF, that's what that means. So, how I look at PMF is I'm a little different in the fact that when I look at a product, or a technology, I don't just look at, like, so you have foundational tech. Like, okay, this is encryption. This is control, right? Now, where's the market that has the biggest problems with it? So, I like to go out and actually talk to those people. Because, like, when you're implementing tech, or you're implementing the product itself, it's different. So, you're like, you have the underlying infrastructure, but whether that's a button or a simple API that you need to build so it works different to hit that PMF...are you familiar with the term build a better mousetrap? VICTORIA: I don't think so. JOE: I'm familiar, but I'd still love to hear you describe it. WENDELL: So, in business school, and likewise, they will tell you "If you build a better mousetrap, people will come, and they will buy your product." So, like, it's a common thing where they're like, "Build a better mousetrap. People will come. They'll be there." And the thing that you learn with consumer product goods and marketing, though, is they actually built a better mousetrap, and it failed. And the reason why it failed is you had a mousetrap that was roughly a cent versus another mousetrap that was three cents. And I think this is in the '60s or so. The other mousetrap was reusable, so it executed a lot better, and everything else is more humane. But what they didn't understand is that it was wives most of the time that would have to actually handle this. And they didn't want the mouse alive, and they didn't want to reuse the trap. They wanted them to actually be disposed of right away. So, by not understanding the market, even though they built a better mousetrap, they'd missed the point. Like, the main problem to solve wasn't killing the mouse or having it be reusable. The main problem to solve was, like, getting rid of the mouse. So like, if you have a solution for getting rid of the mouse, the next thing is your execution for it. Like, does it hit the actual market, which is the fit aspect? Like, every product is a little bit different where you look at, like, how does this fit in? So, in this case, fit is very important for, like, disposing of the mouse, which is why you also have, like, you know, mouse poisons are popular, even though they're terrible because they die somewhere and, hopefully, you don't see them. And it's like sight unseen, right? Now, I'm glad, like, that's changing and stuff. But it's understanding even if you have a solution to something, you need to understand what your market wants out of your solution, and it's not going to be an abstract. It's going to be an emotional, like, execution-based process. So, you kind of have to go, all right, this is my market. This is kind of my fit. But the actual product I'm building is going to change to make sure it works all the way through with this. I was advising a startup many, many years ago, and they were building this CRM software on Android for South America. And I think they were building it for Android 6 or 7 at the time. But the market that they were targeting, they all ran Android 4.1. So, they spent a little over a million dollars building for the wrong version of Android that wouldn't even work on that version of the system. Like, it was one of those things where they were required to build it for that. But they didn't understand the actual market, and they didn't spend enough time researching it. So, it's like you get the Bay Area groupthink. If they had actually spent the time to analyze that market and go, "Oh, they run, you know, an inexpensive phone. It's 4.1. It's low RAM," now you can design a product. If you want it to be a CRM, you're going to, like, chunk up the system more. Like, you're going to change all that instead of just wasting a million dollars building something that now you basically have to start over again from scratch. VICTORIA: That seems like he got off cheap, too. People make way bigger mistakes that cost way more money [laughs] because they [inaudible 24:13] WENDELL: Well, that wasn't me. That was an investor that -- VICTORIA: Oh no. I mean, yeah, not just them. Yeah. WENDELL: He's like, "What would you do?" And I was like, "You should sell this company or sell your stake ASAP because that's a really bad sign." JOE: I have found that the answer nobody ever wants when you're doing product validation or testing product fit is, "You should not build this product." The idea that the software just shouldn't be written is universally unpopular. WENDELL: Yes [laughs]. That's, you know, that's part of the reason why it took me so long to do PrimeLab is because, like, it took a long enough for the software to actually need to be written, if that makes sense. Mid-Roll Ad: When starting a new project, we understand that you want to make the right choices in technology, features, and investment but that you don’t have all year to do extended research. In just a few weeks, thoughtbot’s Discovery Sprints deliver a user-centered product journey, a clickable prototype or Proof of Concept, and key market insights from focused user research. We’ll help you to identify the primary user flow, decide which framework should be used to bring it to life, and set a firm estimate on future development efforts. Maximize impact and minimize risk with a validated roadmap for your new product. Get started at: tbot.io/sprint. VICTORIA: What does success look like now versus six months or even five years from now? WENDELL: I take a different approach to this because I have so many friends that have sold their businesses. They raise and everything else. I look at success as instead of an exit or another large thing, like, literally, we turned down a billion-dollar term sheet offer. I didn't like the terms. I didn't like what it would do from the control standpoint of the technology. What I care about is go-to-market and, like, adoption and actually getting the tech out there in a way that has market penetration but, like, that adds value to every person's life. VICTORIA: Yeah, maybe say more about that. Like how do you see AI and this technology you have with PrimeLab benefiting people and benefiting the industry that you're working within? WENDELL: So, the current AI models are kind of weird. They're basically just filter systems because they communicate in pixel space and then go down to functional space. It's the GPU. GPUs are actually terrible to use for AI. This is why you have dedicated AI chips getting built. Hopefully, the RISC-V chipset does actually do something because that's a chipset that I think it's an open-source chipset, but you can actually especially build models on it. So, I think that we're going to see a lot more in the RISC-V chipset where it's like, this is just for one particular image, or this is just for explosions, or this is just for touching up all these different points in the actual individual, like, microcontroller module data that ends up compiling to move forward with it. But the AI models now it's like you took the internet, and you're trying to ask it a probability question, what I was talking about before, where it's not an absolute. So, it's like, if I want to do an OCR system or anything, I take an image. It's got to say, "This is..." letters; it's going to recognize that. So, there's, like, multiple models and algorithms that need to run on that whole entire process. You even have artificial data, but all of that information is an approximation. It's not an absolute. If you want absolute, you can get a lot of absolute data from the actual hardware devices themselves. You know, take a Sony camera. You could see the lighting. You could see the raw information, everything else there. But because of how expensive it is, people compress it. Like, take YouTube where it's compressed, and now you're training off of it. You're trying to compress it more and then run an algorithm so that you don't have to actually process those large, raw files all the way through. That's just a bad infrastructure for compute. You're trying to reduce, but you're also trying to utilize what you own for rights, same thing, contextual, or anything else there. There's no value in a model. Once a model is out there, it's just weights moving it back and forth. The value is in the data and the applications. So, the actual data itself that's going in. So, if you have just lava scenes, like, having all that data for lava, and I want to put it in a background, now I can do that, but more importantly, it's not about just adding it into the background. The thing that is often missed is contextually the output. So, like, say I want to do a financial report. Rather than having the data of all financial reports out there, what I want as the input is my financial data. And what I want as, like, a fine-tuning output is an example of the reports that were generated. And I don't want those reports as the input to inform the output because that's where you get a hallucination. Maybe it starts grabbing financial data from someone else. And I also think we're in store for a lot more hacks because with not just poisoning data, which we do in the functional space, if someone tries to access it. But, I mean, literally, there's the story...I think the guy was in Hong Kong, where they faked his board all the way through with it. Because you have agents acting and executing on people's behalf, you're going to have systems where people go onto the hardware and start generating fake financial numbers. And now that's going to get reported. Or you pay an invoice that you weren't supposed to pay because someone manipulated your AI agent. And a lot of the stuff that we're seeing now from Microsoft and everything else that's not really where the models will go. It's great to do it, but it's kind of like we're in the dial-up stage of AI. Like [chuckles], dial-up has its use cases and stuff, but it's nowhere near what the tech will look like in the future, and it's nowhere near how it will function. And one of the big pushbacks that you see, like, from Google, from all these different places, like, they want your attention. But at the end of the day, Google's an ad company. Facebook's an ad company. It's not in their best interest to have hyper-localized data that you control for your models and likewise. They want it in the cloud. They want it used there, where they can control that data, and they can monetize and advertise for you. But at the same time, like AI models work the best, and AI applications work the best when the data set is limited, so it can't hallucinate, and when the outputs are actually controlled to what it should be from an informed standpoint. So, where we're at this is just in the beginning stages of stuff. VICTORIA: That's really interesting. Thank you so much for sharing. I think if you could go back in time when you first started PrimeLab and give yourself some advice, what would you say? WENDELL: You know, I lived through the Great Recession. The Great Recession informed me a lot more. The things that I didn't understand this time...like the Great Recession, was market contributors doing stuff that impacted everyone with their spend and their adoption, and how those things were. But the Fed raising interest rates, which is, you know, Silicon Valley Bank failed and stuff like that, that dynamic of those startups and, like, how much startups power everything, like, I would have advised myself to pay more attention to the Fed and those market dynamics going forward. Because what changed is it's not just the Silicon Valley Bank failed it, you know, Rippling went down, for instance, which would pay therapists in Florida and all kinds of stuff. Like, it broke so many different things. It caused bottlenecks in business that we're still going through. Like, everyone's like, "Oh, we're getting back to normal." Really not. It's still, like, delayed all the way through it. The AI aspect is really getting back to normal, where people are really pushing AI. But if you look at SaaS and other industries, it really, really slowed down. And the reason why that matters is, like, in my field, production and timelines matter. So, when you have that plus, you know, the entertainment strike and everything else, you have things where the actual production of things starts slowing down immensely. Whereas AI is one of the few things that you still have innovations because that never really slowed down, same thing with the models. But all the rest of the industries and stuff have really slowed down. And understanding what that means from an operational execution standpoint...it's a good thing I have my co-founder [inaudible 32:24]. It matters quite a bit because it means your team sizes have to change, how you handle certain clients has to change. Because once those companies start downsizing or laying off people for whatever reason like, that's going to change how you're working with them, and their requirements are going to change as well. VICTORIA: And what do you see on the horizon as a challenge or a big hurdle that you face as a company or as an industry? WENDELL: You know, the entertainment market's really interesting from all the different sign-offs. The challenge is more execution of timeline. So, like, if you're doing something with, like, Nvidia and the healthcare thing, it could take years. If you're doing something in, like, the IoT space, you know, also years. If you do something in the entertainment space, it could take weeks to months, except the large studios. The larger studios, it could take a couple of years as well. But going to market, I think, is a very big challenge, not just for us but the whole entire industry. I mean, there's a reason why Sam Altman came down to LA to meet with studios, to try and get stuff moving forward. And I think one of the things that he's forgetting is like, you think of Netflix. Netflix is streaming. In order for that to work, they needed Roku, and they needed Kevin Spacey because [chuckles]...it's crazy to say that, but House of Cards is kind of what made it, right? And Hollywood was mostly boxing them out quite a bit. Same thing with Blockbuster otherwise. They had to drop a hundred million dollars, a large enough bankable star at the time that would really push something forward. And they had to basically really push Roku out there so that they had PMF across the board. What that means, though, is, like, Netflix is paying for content like crazy, right? So, this is kind of enshittification in a process. So, they're paying for content like crazy. So, now Hollywood's making money. They like it. At the studios, they don't love it when their stuff's going there because maybe it's less money, but now they start cutting the seasons short. They start cutting...it's a lot more algorithmic-driven. You have the ad systems that sort of come out. So, now, like, Netflix is not just doing ads where the customer experience is getting worse, but now, also, the business experience for those partners selling stuff is also getting worse, and all that value is getting driven to Netflix. Like, that's the tech system and Hollywood's learned that. But, like, when you're looking at the next adoption, like, they're hesitant for that. Just like a lot of stuff with AI, they're hesitant because they're thinking about all the power and control that they gave up. But you have to show how they're going to make money. You can't just cut costs, right? If you can't show how they're going to make money, you're not going to get adopted. That's kind of what I like there because so much of tech is about saving costs and being more efficient. In the entertainment industry, it's not just those two things. It's how can I make more money? And it's going to, like, ooh, you can monetize your content through training samples and stuff like that. So, our model goes exactly against what the large tech companies have where they want to take content, train on it, like the search engine does, suck the value off Sam Altman's Sora. Ours goes, all right, this is your content. Only you own this. You can take your own content, train it, and then perform this operation on it that is more efficient likewise. And if you choose to monetize it in any way, shape, or form, we can just take the functional space, not all the images and no one will ever see it, and take that functional space for training so that you can actually monetize from that as well. VICTORIA: I love that. Super interesting. Thank you so much for sharing. And do you have any questions for me or for Joe? WENDELL: I've noticed a lot of differences on, like, applications and how systems are built. So, I'm kind of curious about you guys' standpoint about applications, you know, the Apple Vision Pro. Facebook just said they'd start licensing out their AI system, or Meta, whatever. So, you have the comparisons to Android versus iOS that's happening, stuff like that. So, I'm really curious about, like, you guys' thoughts on the Vision Pro and that ecosystem. JOE: Well, I can't speak for all of thoughtbot, but I can say that, to me, it was interesting to see that get released. And it's been interesting to see how aggressively Meta and Apple have been pursuing the various VR markets. Like it reminds me of when television companies and studios worked really hard to get 3D movies to be a thing. WENDELL: [laughs]. JOE: Because I think they just ran out of things that people are asking for. Like, people were interested in getting better resolutions up to a point. Like, they wanted better packaging. But it got to a point where it was like, they didn't want to give anybody anything they were asking for. So, they were like, what if it's in 3D? And, like, for years, it seemed like Apple was really on top of seeing what people really wanted, and being able to present a very well-prepared version of that product before other companies were able to. And, personally, it's not what I saw with the Apple Vision Pro. Like, it wasn't the obvious missing space that was there when the iPhone or the iPad showed up. WENDELL: Yeah, I always go back to, like, the "Why?" question. You know, previously when...even just before we had talked, I was talking about comparative religions, and why that's so valuable is because it really teaches you...again, I've had this conversation before, but the comparative religions, if you think about religion as a tech company, they're always trying to solve why. Like, why did the sun come up? Why did this happen, right? And you always have to do that. So, apply that to technology, Google or Apple, why does this product exist? And when you get to, like, it just existed to make money, I think that's really the 3D thing. Whereas, like, why did the iPhone exist? It existed to solve this problem of being portable on the go and getting information in the way that we communicated, too. VICTORIA: Yeah. I think the Apple Vision Pro appeals to a very specific market segment and that that segment is not me [laughter]. I, actually, during COVID...after...it was, like...yeah, we're still in COVID. But during the pandemic, I moved from DC to California. And to connect with some old friends, I bought a VR headset and decided to go to virtual coffee with them. And it just makes me nauseous. And it actually affects...quite a lot of women get nauseous in VR. For some people, the look—the capability is really exciting. They have the extra money to spend on gadgets, and that's what they like. And it's very appealing, and the, like potential, is really interesting. I just find it for myself. Personally, I'm more drawn to tech that's not maybe cutting edge but solves problems for actual people. And kind of why I'm interested in PrimeLab, what you were mentioning is just how artists can use this technology to protect their creative work. To give that power back to people and that control over their content, I think, is really interesting rather than...I'm not really sure what I would do with the Apple Vision Pro [laughs]. Like, the early ones, I mean, it's cool. It's fun. I definitely enjoy it. Like, I sometimes like to learn about it, but it's not my passionate genre of tech that I normally go for. WENDELL: Going back to what you just said about, like, control, like, part of the thing is because of the hash IDs that we put into place, like, you don't need analytics. You don't need cookies or anything else, like the content holder. Basically, like, if you have a TV set or something and you want to stream content to it, you can actually see that information directly yourself. So, it takes the person generating it and the person viewing it. It forms...we call them function access keys. It forms a one-to-one relationship, basically, where you guys know if you want to know what you want to know, but then you choose to give access to the platform if you want to, which changes the dynamic of control quite a bit. And it's interesting because when you look at platforms like the Apple Vision Pro, and you look at Apple's whole entire system as a whole, just trying to lock in people, I think it's interesting because something like what I just described, Apple can't really stop. It's how compute works. So, if people want to use it, there's nothing they could do to stop it from being used. So, I'm really interested in the product stuff and just more about, like, how...and I'm curious what you guys think on this, too. Especially as you see phones and processors and everything else, I'm really interested in, like, how these things come about, like, how things are actually built and developed and the why for that, like, in the everyday use. So, like, the Apple Watch it started off as a fashion thing, which looked like a money grab, and then the why was, oh yeah, fitness. So, just curious if you guys have seen any other products out there that you're like, oh, this really resonates with me and the why. JOE: Yeah, I'm not really a gadget person, but I think the idea of taking some of the capabilities that we've gotten with the internet and with phones and making them hands-free was interesting. And that, to me, was what I think started pushing the development of products like the Apple Watch or Google Glass. Like, I think that hands-free capability, the trade-off became rewarding in the fitness field, but I think it's more generically applicable. I think that technology it's too obtrusive in other scenarios and too bad at its job to do some of the things it could do. And people got creeped out by Google Glass. But it doesn't really seem like the Vision Pro fits in there. Something being successful hands-free means it becomes less obtrusive, whereas the Vision Pro is like you become a cyborg. VICTORIA: Do you have anything else you would like to promote? WENDELL: I wouldn't say necessarily promote as much as like people with ideas or aspirations, like, I think it's important that you think counter to what everyone else is doing. There's that line of, like, when everyone else is running in one direction, run the other. And it's like, if you have a business or startup idea, really think about your market. Like, think about why you're doing what you're doing, and don't be afraid to just go out there and talk to people. You will get value no matter who you talk to. So, like, I'm a hugely tech-based person. My wife is a therapist, and I learn from her everyday things about emotional intelligence and all kinds of things that I would be an idiot otherwise. But also, learn, like, you can always learn something from someone. Like, take the time to listen to them. Take the time to actually, like, try and figure out what's one thing I can learn from someone, even if, you know, I learn stuff from my daughters even. Like, don't put things in boxes. Like, try to think outside of like, how can I ask a question to learn? VICTORIA: I love that advice. That's great. WENDELL: Have you guys used Suno before? VICTORIA: That's music, right? Music AI. WENDELL: All right, I got to show you guys this. We're going to create you a quick theme song. Like, this is what I mean by, like, it's an interesting solution for why. VICTORIA: That does sound fun. I like the ones...like my friend's a doctor, and she uses AI to take her conversation she's having with patients and automatically fill out her notes. And it saves her, like, 20 hours of documentation every week. Like, I like that kind of app. I'm like, oh, that makes a lot of sense. WENDELL: What's a style of music that you guys really like? JOE: Swedish pop VICTORIA: Like ABBA [laughs]? I'm down for an ABBA Giant Robots theme song. Sounds great. WENDELL: I think you're going to like this. [Music Playing] VICTORIA: These are awesome. They're super fun. Thank you so much. You can subscribe to the show and find notes along with a complete transcript for this episode at giantrobots.fm. If you have questions or comments, you can email us at [email protected]. And you can find me on X @victori_ousg. This podcast is brought to you by thoughtbot and produced and edited by Mandy Moore. Thanks for listening. See you next time.  AD: Did you know thoughtbot has a referral program? If you introduce us to someone looking for a design or development partner, we will compensate you if they decide to work with us. More info on our website at: tbot.io/referral. Or you can email us at: [email protected] with any questions.Special Guest: Joe Ferris.Sponsored By:thoughtbot: When starting a new project, we understand that you want to make the right choices in technology, features, and investment, but that you don’t have all year to do extended research. In just a few weeks, thoughtbot’s Discovery Sprints deliver a user-centered product journey, a clickable prototype or Proof of Concept, and key market insights from focused user research. We’ll help you to identify the primary user flow, decide which framework should be used to bring it to life, and set a firm estimate on future development efforts. Maximize impact and minimize risk with a validated roadmap for your new product. Get started at: tbot.io/sprintSupport Giant Robots Smashing Into Other Giant Robots