8 AI Business Ideas for Entrepreneurs to Pursue in 2024 ft. Siqi Chen

Foreign.

Hey, welcome to the Next Wave podcast. I'm Matt Wolf. I'm here with Nathan Lands. And once again we've got an amazing episode for you. In this episode, we're going to break down some business ideas that leverage AI that we wish existed that don't exist yet. So ideally, people listening to this episode will go and maybe build some of these things because these are ideas that we think there's some money behind.

Today we have Siki Chen on. He originally was in the game industry, so he created a startup called Serious Business that he sold to Zynga for a ton of money back in the day. And since then he's been like, you know, one of the most prominent angel investors in Silicon Valley. He's now running Runway.com which is a great startup.

Yeah. So this is legitimately one of the most fun conversations I think we've had on this show where me, Nathan and Sicky, we just sort of bounce ideas off each other for businesses. And as we share ideas, the ideas just ramp up more and more and more and get spicier and spicier to the last one, which I think is going to blow some people's minds. I mean, I know it scared me a little bit when he gave the idea. We get into some crazier and crazier ideas as this episode goes on. So check this episode out. It's probably going to give you some ideas for spin off businesses or ways that you can use AI at your business that you're not thinking of or businesses that you can potentially go build. But it is a fun episode, a fascinating episode. I think you're really going to enjoy it. So let's dive in with Sikchen.

Using only 20% of your business data is like dating someone who only texts emojis. First of all, that's annoying, and second, you're missing a lot of context. But that's how most businesses operate today, using only 20% of their data. Unless you have HubSpot, where all the emails, call logs and chat messages turn into insights to grow your business because all that data makes all the difference. Learn more@HubSpot.com.

Hey, Sticky, thanks for coming on today.

Hey, Nathan, thanks for having me.

Yeah, I guess first it'd be good to give people some context, like why you're here, how you got on the show, and you know a little bit about how we know each other.

Yeah, funnily enough, I thought I was here because I tweeted about AI and I have, you know, people will follow me on Twitter or Whatever. But it turns out I was just reminded that the reason I'm here is you announced you had a podcast. That's why I'm here. I invited myself.

So cool. So, I mean, we were thinking today we really want to try something new, just like share AI business ideas. You know, some of them may be great, some of them may suck.

Sounds good. I came prepared.

Awesome. Let's, let's go ahead and talk about some of these business ideas. I think we should start with Nathan because I know Nathan's got three lined up. And again, if, if they suck, the comments section will let us know.

Yes, my first idea was just, you know, it actually comes back. I think the last time we met Ziggy was in Hong Kong. Is that the last time we met in person? And that was back when I was, when I was partner with Barry Osborne, you know, the producer of Lord of the Rings, the Matrix. I was trying to create this crazy crypto funded movie studio and you know, we were going out to New Zealand and Hollywood a lot together, like me and Barry and I got this tour of Weta when I was out in New Zealand and it was amazing. I was blown away by it. You know, I got to hold Aragorn Sword, WETA Workshop and also like seeing their business, like how much money they make just making all the special effects and stuff for movie studios, it's like, that's a great business. And now that you've got AI Video, it feels like there's an opportunity for like almost like a Weta for like AI Video. Right. Where all the, you know, whether it's in movies or TV or games, to go out there and be like the shop that specializes in creating AI videos for films. But yeah, I don't have much to say about it beyond that. But that's the idea. So good. Interesting. Sucks.

Not sure how much I could stay. There isn't much, but I think it's a great idea. Yeah, I am aware of some new companies that I'm not allowed to disclose doing this with very credible people. But a thing I can talk about though is related, which is there's a company called NIM Video and it started by this guy Yuri, who ran Superdao and very credible founder, raised a lot of money. And the entire idea is the state of the art for generative models is going to be open source. And so you're going to need professional grade workflow tools to orchestrate all of these. So you know, you want something more fine grained than what Runway ML alone can provide. You and he's getting started with that. It's got a bunch of traders on a platform. So this idea of professionalizing, in general, AI tools and special effects, I think it's huge.

I guess my question, though, is, like, how close are we to not necessarily needing an agency where you just enter the prompt of, like, here's the video I need, and it gives it back to you that, you know, the agency is just like a middleman that people don't need anymore.

There's a. It's a very related question how we think about AI and our own product at Runway. So, you know, we build, like, this finance platform, and, you know, there's a lot of people who use AI to make finance better, whether it's, like, planning or modeling. And our very particular philosophy on this is that the best use of AI in its current capabilities today and I think also in the future as a tool for thought, rather than a tool that does the thinking for you. And I think when I think about, like, anything creative, and I think inherently, most things that you do are creative, whether it's like finance business or like, creating a movie like you. You would rather have something that helps you offload your intent, your creative intent faster. And that's, I think, what AI can do at its best. I'm a little bit less bullish on the AI is just going to do it all for you until, you know, when ASI happens. Like, I think we can live in that world, but until that point, tools for thought is my bet on good AI products.

Yeah, yeah.

And I mean, if ASI happens, we probably don't worry about having an agency or money anyways. Right? So it's like.

Yeah.

And I mean, like, right now, you've got AI art tools, you got AI video tools, you got Runway, you got Gen 3. There's some people who somehow managed to get really amazing stuff out of it. And then a lot of people who, no matter what they do, can't get anything good. So, I mean, even using AI, there seems to be some level of skill to get really good stuff out of them.

You know, I should clarify, since you mentioned Runway, I am not the owner of Runway. ML, which is the right one, makes the videos. Again, I run a company called Runway.com and we're a finance platform.

The real Runway, you've got the better domain name.

Yeah, people get the two companies confused all the time. It's actually really wild.

So I guess I'll share. Like, my. My first business idea is actually the second on my list, but it's related to video. So I'll share this one first. And it's something that I wish I had and it may exist. So if it does exist, feel free to share. But I want a tool where I can take like video footage out in the wild. Let's say I'm like going on vacation with the family and I take video with the family and I record like a roll of me talking to my camera or I find stock footage on a website and I just want to be able to take all of this video footage that I have loaded into some sort of platform, have I like scan all of the content for me and, and then give it tags, give it descriptions. And right now it's easy to do if there's like talking, right. It can obviously transcribe the audio and then figure out tags and stuff. But I want something where I can throw videos of just like me filming the Eiffel Tower or me filming my family running on the beach. And I throw it in there and it watches the video and organizes with tags and descriptions based on the content of the video. That way, when I'm going to like make my YouTube videos, I can go in and say my family on the beach. And it'll like find that video for me and surface it. And it feels like something that should exist, but I haven't found it yet. And that's something I really, really want to see exist. It's just something that I can just dump tons of video content in and have it sort organized, tag, and make searchable for me.

A few things come to mind. I, I, I saw where we were going is you just want a bunch of clips and have it just edit the video for you.

No, I actually like the editing process. For me, that's like where the creativity lies that I enjoy doing. But I have a hard time when I shoot like hours and hours and hours of footage while I'm at an event and then come home and then look for the specific clips that I need to pull my video together.

So I don't know of specific products that do exactly that. The closest thing I can think of though is Google Photos.

Yeah, I was going to say that I think Google will do this right, Because Google does it for photos already.

So yes, it does work for video too already.

Oh, does it?

So you can search for like sushi, text beach.

I have not tried Google Photos for that, so maybe. So it could be a horrible business idea if Google is already doing it.

Yeah, I mean, I do think the idea of like, generally you have a bunch of Clips and will make stuff for you or help you tag. It is a great one. I know captions is, you know the most. There's a closest company to that. I'm not sure how much it does of organizing or trading for you, but the tools it has like really good and it's doable on your phone. Yeah, that's the first one I would think about.

Captions may have even been the company because I, I actually put this idea on Twitter. I don't know. It's been several months and a company jumped in and said oh, we're working on this, wink, wink. And I feel like it might have actually been captions.

Yeah, I. I'd be surprised if they weren't. That sounds like right up there. Yeah. Ali.

Yeah.

Yeah.

Cool. Sneaky.

Your turn.

Okay, so not spicy one but very practical. I use a service called Stainbox and it said email prioritizer filter and it's really good. I just apply a label to my emails and it'll know okay, this type of email isn't that person. I'm going to put it later or I'm going to black hole it, never see it again. Or I can create a custom label. It's just really good at what it does. It's been around for like 15 years. Got all kinds of like little features. What I love about it is a primary interface is Gmail because you don't actually interact with the website really. You just have once and then you just apply labels and it'll just automatically learn and it'll pick up like and you're training it right. And the way you untrain it just like we looked into your inbox and it was like, oh, this is actually important. So what I want and I wanted this so bad, I actually built a crappy version of this myself using Zapier is I wanted to be able to categorize my emails automatically. Unstructured ways. So I want to say, hey, all of the really intellectual smart substacks should go here and everything related to my kids should go in this folder and my kids school stuff and homework should go here and everything from our investors should go there. And I don't want to, you know, you don't want to maintain like a running list of this stuff. You just want to say this is my intent. So I have a zap that that does this. It reads every single email and I have a giant prompt that describes all the labels and all of the descriptions of the labels and it'll actually go and apply to labels. It's not Great. The thing that it's missing is and it's on my to do list of work on this, but I just wish someone would build this for me, which is I want to learn and rag it as I'm changing the labels. Right. So if it gets something wrong, I want it to just like, oh well this is. I got this wrong. I'm going to update the prompt. I think that'll just be super useful. Instead of sorting emails by like domain or person, I could just describe the vibes of the email I want and I'll know what to show me where to put it. And that's my not spicy idea that I'm still. Sometimes no one has done.

Yeah, that's actually I tweeted about something similar to that a few months back and then somebody responded to me saying, oh, I'm building that. And I guess they like raised a little bit of money. But then it's already kind of like it didn't work out or something. I'm not sure. I've also been wondering like, why is Superhuman not done this? Maybe it's just too different from their existing product. But it feels like they should be doing that. I don't know.

Yeah. Usa. Yeah.

Yeah, I can see that being super, super useful. I mean I run into the same kind of stuff, right. Like I use the same email address for almost all of my stuff. So I get emails from my kids school and also it'd be great if it worked across multiple inboxes. Right. Like one of the things that I like about Superhuman is that I could just kind of quickly switch between tabs of all my email accounts. So it'd be cool if it was just like these are all my email addresses. Sort across all those.

Yeah. So the, the thing that works and it's already like pretty valuable is I can say, and I do say, if it's a cold inbound sales email, put in a folder and ARP have it.

Yeah.

Like that's just something you can't do. Right? Like you have to read it like, oh, is this selling cold emails?

Yeah, yeah, yeah.

But if it's like a friend that I know and it's word, then like put us in an inbox. That's the level of intent I want to offload. Right. And right now, even with SaneBox, you have to train it over time and says, well, these are the specific important email addresses. And it can't really figure out if it's a cold, cold email. It sort of does, but it's not always super reliable and an AI to read an email and do a pretty good job of telling.

So if anybody's listening to this and they've tried to cold email you with a pitch and you never got back, now you know why?

Just pretend like you're his. Just pretend like you're his friend. Just like, I do read.

I do read. Kachin.

Oh, got you. Got you. Yeah.

Yeah.

Well, damn. So my next one's, like, very similar. My next. My next one is basically like custom newsletter where, like, you know, you take different sources, you're getting inbox, whether it's newsletters or other things, you know, new podcast episodes or whatever, and kind of consolidating all that into a single newsletter. Now, I don't know what the business model would be there. Cause, like, you're kind of. I guess you're screwing over the newsletters. And, you know, I have a newsletter. Matt has a newsletter. So we probably don't necessarily like this idea personally, but I as a, you know, as a business owner, I don't like it as an individual. I would love to have something where I can just like, here's all the stuff that I actually care to learn from, you know, the substacks, the different newsletters, and just give me a summary in one email every day or every week or whatever. I do.

I want that. Yeah, it's actually not bad if you make it so that I still have the subscribed substack.

Right.

Yeah, I feel like that exists, too. I feel like I've seen. I don't know the name of it too, off top of my head, but I feel like I've seen something like that before.

I know that, you know, is 18 kind of like, summarizes, like the. It gives you a snippet which is, like, kind of related. But the idea of, like. Yeah, I also remember there was something that, like, I'll create a podcast, like a daily briefing out of your stuff. That was Jelly Pod.

There's one called Jelly Pod. Yeah.

What does that do?

So that one used to. You. They give you a specific email address. You go and subscribe to newsletters with that email address, and then it gives you a daily podcast that sort of summarizes all of those newsletters into a single audio. Now, there is one caveat. They only. They only let you subscribe, I think, to 10 newsletters. And I subscribe to way more than 10 newsletters. But if you could subscribe up to 10 newsletters on it and every day it'll, like, submit to your podcast feed. So wherever you listen to Podcasts, It'll, you know, it creates an RSS feed so you can actually listen to that podcast episode and it breaks down everything that was in those 10 emails in a little 10 minute briefing.

That's a great idea. I, I, I'm, I'm, I, I literally have it open right now. I'm going to sign up for it. It reminds me of like one of the tools our team built. It's a thing that just goes through the last week's worth of customer success calls and takes a transcript, summarizes it and extracts insights. And he turned it into a podcast using Leva Labs with my voice as a deep fake. So like every Monday you press play and the likes of me telling him about what happened to all the shit our customers experienced.

Oh, that's awesome. For me, I think Feedly is another tool that it's like pretty underrated that I don't hear enough about. But Feedly is a tool where you can subscribe to RSS feeds, right? You can subscribe to your favorite blogs, but it also gives you a custom email address so you can go and subscribe to newsletters with Feedly. So then every day you just open up your Feedly platform and it'll say, here's all the newsletters that came across. Here's all that you should know from all these various blogs. So like, you know, I make a, I make a video every week of like, here's all of the AI news that came out for this entire past week. I pretty much just watch Feedly every day to keep up with that stuff. So I'm subscribed To Google's blog, OpenAI's blog, Anthropic's blog, Stability, AI's blog, you name it, I'm subscribed to their blog. And then I'm also subscribed to like every existing AI newsletter out there. So every day I get sort of like, it's not like a briefing. I still have to kind of click through each one one at a time. But I have like a single point where I can go through and like look at everything real quick and get caught up.

I'll leave that.

All right, so this is something that I think is sort of inevitable, but I've got these like frame glasses from Brilliant Labs. I'm not sure if you guys have seen these before, but they have like a little teeny tiny camera right here on the front and they have like, you probably can't tell on the camera, but there's like a little heads up display on the right eye of it where like, it'll like scroll a little bit of text across in front of your eye so you can see text. But one thing that I would, and this is open source too, so anybody can develop with it or build on top of it however they want. One thing I'd love to see, first of all, better form factor. These look like they're straight out of Harry Potter. But the second thing I'd love to see is a face remembering app, right? I wear these, I go meet somebody for the first time. When I meet somebody, like there's a code word if I say nice to meet you or something, right?

Yeah.

Snapshot of their face. Remember their face in the AI. Then the next time I run into that person, a little heads up display shows in my right eye. Oh, this is sticky chin, right? Like so I walk up and I'm like, hey, good to see you again. Sticky, right? Like, but I feel like you can build that with this, this thing here because it's got a little camera and it's got a little heads up display in one of your eyes. So you know, a little image recognition of people's faces. Use, you know, maybe Google's API that, you know, it's got the image recognition in it. You know, take pictures of people, remember their faces next time you run into them. Right across the little heads up display. This is this person's name. So you never forget anybody's names again. I want to see that being made and I feel like with these brilliant labs, like open source tech, somebody should be able to make that pretty easily I would imagine. I'm not a coder, but I feel like it wouldn't be too hard.

I love it I guess. Only question is like, does meta get there first with the Ray Bans?

Yeah, I'd be fine if meta did it too. I love the Ray Bans.

It sounds like meta's next thing that they're going to do with like the next iteration of Ray Bans is some more like augmented reality stuff. They were going, they were going to do like this giant, this like new VR headset that was supposed to compete with the Apple Vision Pro, like a really high end version of the Quest. Well, they just scrapped that and it sounds like now they're pushing because of how well sales have done on the meta Ray bans. They're now pushing more into the ray ban area and it sounds like they're going to start doing things like AR and heads up display inside of the Ray Bans instead of making higher end virtual reality devices.

I Mean that's going to be a huge opportunity for some company to build.

Yeah, that's killer. Absolutely.

I've always wanted that where like you could just see, you know, who they are, like if it's good for your business or not like some kind of context there. But you know, it gets into like a really weird territory though, right? It's like almost like a black mirror where it's like, what's their score? And like is this a, is this like a person I should talk to? Or they're like totally irrelevant. They're like not worth.

Yeah, yeah, yeah. Not.

Not important. Important. You know.

Yeah. Can we be honest and say like the real reason why we want that is like every other day we go out and someone says hi and you're like, I don't know who it is.

Yeah, yeah, yeah, yeah, yeah, yeah, yeah.

Well I mean it can also save like additional data, right? Like if you're having a conversation. But it does kind of get into that creepy area of like how much is it recording and, and that kind of stuff. But it could like remember kids names, it can remember birthdays, remember, you know, different stuff like that. So when you're having a conversation it all just sort of pops up on a little heads up display for you. So. So you can easily remember that stuff. But I run into that all the time. Like I go to, I go to like at least one conference a month kind of thing. Like I go to a lot of conferences and I don't know how many times I've gone to a conference where somebody's like, hey, great to see you again. I'm like, yeah, you do. And I have to look down at their name badge and try to be discreet about it.

Yeah.

Do you wear the glasses in public and do people get angry? I'm.

I wonder. I wear the Ray Bans in public all the time because they just look like normal Ray Bans.

Yeah.

I see Google Glass for like a month.

Yeah, well, I mean more and more devices are popping up like that too, right. Like there's like earbuds coming out that are going to like record you and they've got the.

People just get used to it. Probably like there's not much, you know.

Depending and all that kind of stuff. But yeah, I mean some of like. I know.

Same way, you know, you get used to the cell phone when you're out. Like. Yeah, yeah, you're gonna get immediate picture probably by accident. Sucks for me.

I mean literally everybody you see in public has a device that could record your voice. Take pictures of you, take videos of you. Everybody, I have another one.

I'm not gonna go spicy yet. I'm gonna save it. All right, so this is a general category of products I'm interested in, which is like ways of exploring the latent space of Mini LLM. So websim, you know, is like an example of that where hallucinates the Internet and people were building really interesting things around that thought it was interesting, was this twofer. There's a company that I recently invested in called 01 computer or adaptive computer. And it's kind of like a higher level Devin. It can just make apps for you. And one of the things you can build just by asking to build was a hallucinated Wikipedia. So basically you can start with like anything and it'll create a Wikipedia article and all the links work and you just keep on browsing it and holding it. Say that. I just think like, generally those categories of things are interesting. Like being able to explore the latent space of a generative model, right. You have an image. I'm like, I want to go over there, right? And this is where things are going eventually become real time, right. For example, there was a thing that I saw yesterday where someone trained the Fusion model on Doom 1.

Things get killed, you could fire the gun. And we're going to go from rendering pixels, just losing the entire image on your computers. I think that entire category of things is interesting. I think someone making a generative Wikipedia is also interesting. There's a few companies that are like that, but not quite like that. But I kind of like that idea quite a bit.

Yeah, Yeah. I mean, you could do something similar right now with perplexity, right? Like they've got their pages featured where you could basically throw in any topic and it essentially builds like a Wikipedia page. But it's not. It's not like hallucinating it. It's just going and finding the data on the Internet for you, essentially.

Yeah, I think what, what fascinates me is like more of the mode of interaction, right? Like that still requires you to ask something. And this idea of like just being a little click or look around to browse inquiry interface is like super interesting. That's what like, I think Websen is the first thing that I've seen that did this, right? Like just made it a hyperlink so that you click instead of having to type anything. I just think there's like a whole area of unexplored product possibilities around this idea of other ways to generate prompts and explore the latent Space.

Yeah, the doom one's so interesting. It's like. It's almost like you're moving through a dream, right? Like it's like dreaming up everything in the world as you're moving through it. It's like. That is so amazing. We have nothing like that currently.

Well, I mean, some people would argue that's what's happening right now.

Right.

Right now.

Like right now. The simulation we're in, you mean?

Yeah. There's a wall in front of me right now. I think. I know it's on the other side of the wall, but I'm not on the other side of the wall, so I don't know.

Yeah, don't say simulation three times. We may get exited out of it. So I think that's why it's the same.

Yeah. There's an AI art tool. I don't think I'm allowed to say the name of it because I don't think they publicly announced that this is coming yet. But there's a tool that's essentially an, an exploration of like a diffusion model where you, you enter a prompt into it. And when you enter a prompt, it will basically generate like an infinite number of images with that prompt that you can sort of infinitely scroll through. So in real time, it'll just keep on popping up new images generated by that prompt. And then if you click on one of the images, it now uses that image that you just clicked on as an image reference. And then it starts a whole new page with this as your image reference. And then you can sort of infinitely scroll new images based on this new image reference. And you can sort of get lost in it for hours just clicking on, oh, I like this image. And the idea being that eventually you can find the absolute perfect ideal image for what you were going for with that prompt by just scrolling and clicking on the ones that are closest to what you're looking for. And then once you find the one you're looking for, you now know the seed, the prompt, everything you need to know to regenerate that image again and again.

I need that.

Yeah, it's coming, it's coming.

Can you DM me the thing after the show?

Yes, for sure.

That's Hallucinated Pinterest.

Right.

And I think like, in general, like this idea of taking a known product and product interaction and just hallucinating it, whether it's a browser, Wikipedia, Pinterest is amazing. Yeah, I, I love the whole class of ideas.

Yeah, no, that, that's going to be fun. It's also just going to Be a new addiction for a lot of people, I think. All right, Nathan.

The, the next one would be, you know, back in the day, Codecademy was really popular. I don't know if it maybe still is, I have no idea. But it feels like that was an idea where it was a great idea. You know, it basically was a, a site that helped you learn how to code. Kind of walked you through everything, you know, and really simplified down learning to code. It was always a great idea, but I feel like it never was fully properly executed on. Like, I don't think it reached its full potential like of the idea. And it feels like now with AI you could, you know, like there. I don't know if you saw that video from a few days ago, like the eight year old girl using cursor to code and make like a little Harry Potter website and stuff. I was like, that is so cool. And that girl must be very intelligent because like, because even, you know, at a. Using cursory is still hard. I think for an eight year old it's very hard. She's definitely a very intelligent child. And so it feels like there's something for, you know, more mainstream audience where anyone like, hey, you can actually code now. Like it can actually help you and like kind of walk you through that of making some simple product or simple website and just showing you you can do it and teaching you. It feels like there's some kind of opportunity there and yeah, kind of surprised I haven't seen you. I love to do that yet.

Yeah, yeah. And it would be like generative, right? So like everybody who uses the program is sort of developing something different as opposed to like Code Academy, which Code Academy would walk you through. You know, everybody's basically building the same thing, right? It's just a set tutorial. I feel like you can have something that's generative where everybody that goes through this program builds something unique, right? It's something completely different every time. You might even at the end of the program have something that is, you know, launchable, something that you actually put out into the world.

And you can have a community aspect too or maybe like you actually share what you built and everyone kind of, kind of see all that stuff could be kind of cool.

I like this.

I've been trying to, you know, do all kinds of things to get my older daughter is 10 now into coding. So we were on hopscotch, we did Code Academy, we have ChatGPT. I got gave her WIX to make a website and over Christmas I was traveling with a Friend and they have a kid and they decided because we're in Japan, they wanted to make a video game that is a cross between Gundam and transformers, called Transdim. So to use ChatGPT asked, all right, how do we make a game? What are the roles of the game? What is a cross between Transformers going to look like? Dall E generated image they downloaded. Yeah, they downloaded Unity download Unity and started working on this game. Then I don't think they got very far, but they got way further on their own than I think any 8 year old would have without AI. I think that's great. My fourth idea, I didn't talk about a third, but just piggy off what you said. I think this general class of almost like an AI nanny where you have these tools that are kind of passive, but if you can have something that kids bored is like, hey, let's learn about X or what do you want to do? Right? And it'll just like interact, talk with you and like it'll let you hallucinate with the AI. I think that's also super interesting.

Yeah, yeah. I mean something like that's a lot better I think than you know, setting them down with an iPad or a Nintendo Switch and just saying, here, go, go play Roblox.

I think there's like some companies are doing dolls that are generated and more interactive. But I'm very pro this idea of like inverting the ChatGPT interaction pattern which is having to talk to you, not waiting for you to talk to them. I think there is so much untapped potential in just listing that.

Yeah, yeah, absolutely.

Yeah. I would love to have that for my son and like have it where, okay, you can't use your PC until you've actually like done this. Like you can't like here's a little project. Go do it now you. Now? Yeah, you can watch YouTube for 30 minutes or whatever.

There you go. Yeah, some sort of locking mechanism where the games unlock once you finish your tasks.

Yeah, that was good.

Very cool. Any other, any other ideas we want.

To throw.

Out now?

Time.

All right.

It's actually not that spicy and it's going to sound stupid, but hear me out.

Okay, okay.

AI girlfriend. But, but, and I know that's not new.

Okay. I'm about to say yeah, okay.

Okay. But I guess. Cool. So there's character, AI and friend. All these things are like doing AI girlfriends, right? So here, here, here is a thing that I think is going to print money if someone does this. You sign up for your AI girlfriend, but the interface and all the interactions look real. So it's not like character AI where you're like, I'm going to create my ideal waifu and this is what she's going to look like and I'm going to generate some images and some personalities instead. It looks exactly like Tinder and you're swiping left and swiping right. And very similar to your idea, the other product, where you're finding the exact right image, if that, but for your waifu and you're selecting, you're swiping right and based on swiping left or so right, it's learning about what your ideal WAIFU looks like and is like, because it's a generated personality, It's a generated image, generated location, everything. And then once you try, right. You match and then you have to actually have a conversation with them. And then in the background, the system will create a, generate an Instagram profile, create a iOS or WhatsApp account, and then it will actually text you as this person. It will DM you over Instagram as this person. And now you have a virtual girlfriend. But it feels real in every virtual channel that you have. It can even over time you can have a zoom call. And what happens when you call, who knows? But like you can real time generate all this stuff and not have it. It's be like this weird bot thing that you're making and it feels pretty.

Yeah, I feel like all of the technology exists right now to make that happen. It's just combining all the parts.

Yeah, exactly. Purely product.

Yeah, yeah, because you got like render net right, where you can upload like some images of a face and it'll make sure that that character stays consistent every time. And you've got the tools that can turn them into videos and you've got the tools that can do the voice. I feel like it's just a matter of somebody mashing all the tools together to, to get that thing.

Yeah, it's like Instagram, right? You're browsing Instagram and all the attractive people are attracted to you. Yeah, that's basically.

Oh, I like it. Yeah. I mean, it does, it does bring up some additional questions of like, like, where, where does that, that take humanity if, like everybody just gets their ideal digital girlfriend? Because then, you know, reproduction could potentially.

I mean, I, I do think my sense about my mental model of how value is created in the PO ASA world, I think humanity is going to adapt pretty quickly.

We have infinite adult videos online, and I think in that world we are biologically, we're just evolved to play status games and I think it's very hard to override biology even with the best of tools. And I think we're just going to value things that are real more. Yeah, yeah. I actually don't think it's going to be like as bad as people think. You know, people are like, oh, we have VR headsets. We're doing the via VR and looking at wife who's like, no, if you weren't doing that.

Yeah, yeah, yeah, you know, yeah, there.

Are probably things too that use AI to actually help connect people too. Like I think that'll probably be a thing like probably in the future. The government's be like, oh shit, this is like a big problem. We need to like have huge incentives for having, for getting married and having kids. Yeah, they'll probably make huge incentives and then there'll probably be AI apps that like help you. Like, here's your personality type and here, go chat with this virtual version of this girl who's like a real girl, real person and, and you're a good match. Then maybe you should actually meet them. I think we'll have stuff like that too. So maybe that'll kind of counterbalance all the AI waifus.

It could be like a net positive for humanity too. Because when you think of like the adult film industry, there's a lot of exploitation, a lot of human trafficking, a lot of that kind of stuff. Right. So I mean, maybe, maybe it does turn out to be a net positive in that sense. I mean you look at adult websites, nobody ever looks at them. Yet somehow Pornhub's always in the top 10 of all most visited sites. So weird how that happens when nobody uses it.

Looking at these videos, it's crazy.

Well, I think, I think these are all great ideas and I think anybody listening to this episode, if, if you go and invent any of these, I think sick. He's going to invest. That's what I'm hearing.

Yeah, that was the agreement.

So that's why he's here.

I just like, I talk founders, I talk to founders. Like, maybe you should do this. This seems like a much better idea than what you're doing.

Before we do wrap it up though, is there, is there anywhere you want people to go check you out? I mean, is, is X the best, best place to follow you? LinkedIn? Like where should people go after, after hearing about.

All of my LinkedIn content is ghost written. So probably you want to follow me on X instead.

And what's your X profile?

I'm Blader B L A D E R. And my company is Runway.com not onewayml.com the Runway.

Yeah.

Yeah. So if you're mid size larger company and you want how better your finances. We run the finance of the departments. The finance. The finance models of Superhuman and Stable Diffusion actually, and Angellist and Lambda AI and Superhuman, a bunch of other companies like that.

So very cool.

Very cool.

Well, check those out. Everybody who's listening in. And yeah, I really, really appreciate you spending the time with us. If you are tuning into this episode and you want to hear more conversations like this and hear more business ideas and AI use cases and all that fun stuff, make sure you subscrib, subscribe over on the YouTubes or wherever you listen to podcasts and sticky. Thanks again.