Seeing The Future from AI Companions to Personal Software

Host: Right now, AI is just an app on your phone. It should not be that way.

Host: Sometimes you need to sort of go big or go home. Not having the balls to do that, especially in this current environment, you can suffer the consequences.

Host: There's a huge mind trap that exists among builders in the space where they somehow think that voice is the main interface. That's because they are somehow thinking about the movie *Her* all the time, but not in the right way.

Host: So you were a true pioneer in the space before everyone else was doing it. Maybe just talk about your reflections of how that category has emerged as you helped create it and then how it's evolved.

Eugenia Kuyda: It definitely evolved. I have noticed. It's really crazy. We had a very strong belief that it would happen, but we still were so surprised when it actually happened, I guess. This was just, it just still felt like complete magic.

Host: Personal software is about to explode from 20 million developers to 8 billion creators. Today, you'll hear from Eugenia Kuyda on how Wabi is building the YouTube of apps, where anyone can create, remix, and share software as easily as posting a video. We discuss why current AI interfaces are holding us back, how many apps will become the new social currency, and what it really means when software becomes ephemeral, personal, and delightful. Let's get into it.

[Interview begins]

Host: So, Eugenia, excited to get into everything you're doing with Wabi, but first, let's contextualize this a little bit. When you look back at the arc of your career, perhaps the last decade, what is the through line through Replika, Luka, et cetera?

Eugenia Kuyda: Oh, wow, that's a hard question. But I guess maybe the main theme is we're always early, sometimes a little bit too early. Sometimes, hopefully this time we're on the right tab of early. But generally, I started working on AI in 2012, so over a decade ago, and was always fascinated by the idea that you'll be able to talk to a machine, to have a meaningful conversation with a machine. But more importantly, with the fact that that could influence your life in a really good way. Before that, the focus was on these AI companions, AI friends that could really be there for you and help you live a better life, help you feel better. And now I guess it's just the same idea, but applied to personal software. Can we build many apps or software that will really help us throughout the day in a very personal way? And it will be focused on you, on helping you live a better life.

Host: Trace the journey of how you got excited about this new category and what you hope it becomes.

Eugenia Kuyda: While running Replika, we always had these conversations with our users about what other products they use and how they use AI. It always struck me as strange that they were using products like ChatGPT or Gemini, Claude, but really they were mostly using it for very simple use cases. They were using ChatGPT just for search or to help them write something, help with homework or something else. No one was really mentioning any of these exciting capabilities that we just kept seeing the models get. So we felt like there must be an interface problem. When people look at a command line or a chatbot, they really just see search, writing tool, maybe I can talk to this. But if you think about that, that's the affordance of a command line. Those are the main use cases. I think even recently, ChatGPT published one of those studies showing that these are, in fact, the main use cases for ChatGPT. I think a third of all of the use was around writing, writing help. So that gave us an idea that there should be some next interface. There should be something else on top of this. Something more interactive, visual, simple for everyone, me included, to actually discover these amazing use cases that all these models already have. We started thinking, we got obsessed with this metaphor that the current chatbots are really the Microsoft DOS era for AI interfaces, and that there will be some sort of a Windows Mac OS moment. The confusing part for most people is that there is so much traction with chatbots anyway, almost a billion people using these AI tools already, but they only use them for these simpler use cases. In order to unlock something else, you need a more exciting interface.

Host: Paint more of a picture of what the world will look like when this Windows or Mac OS moment happens. What will that look like?

Eugenia Kuyda: I do think right now we exist in the paradigm of, when I was growing up, we had TV and here you had maybe a few dozen TV channels. Back in Russia, we only had six TV channels. Now TV still exists, but now we watch YouTube and Reels and TikTok, a lot of UGC stuff. I think the same will happen with software. Right now, we're stuck in this world where these few apps are developed by professional developers. Eventually, we're going to move on to this new world where apps are built by all of us for all of us, and maybe sometimes by AI for us. If you think about the operating system of the future, you open it up and you see your regular popular apps you use all the time. Maybe it's X, maybe it's Instagram. Then you see some cool apps that you discovered your best friend is using, or maybe you built for yourself, or you tweaked one of the apps you found for yourself. Then you'll see AI suggesting some apps for you. Maybe you're going to New York next week and you're into art, so it made an art show finder app for you which is going to help you find art shows around the Airbnb you're staying at. It's a lot more flexible, malleable, and very deeply personalized. Think of it as an operating system built on the platform of you and not on some random fixed context.

Host: Can you talk about that? We've always had this idea that software has to be durable because it's really expensive to make and it seems really serious. What types of ephemeral software, like the New York Art Gallery app, do you imagine existing?

Eugenia Kuyda: Right now we're already seeing some of our first users building very specific apps that would never exist on the App Store. They're just too small, too personalized, too niche. They don't have 20,000 features. Someone built a motivational quote app that's only pulling from one particular show that I didn't even know anything about. But he's really obsessed with that show and he just wants that at 5:30 a.m. when he wakes up. People are building very particular things to fit their needs. For example, I was putting my kids to bed the other day and my daughter wants to play these puzzle games. When I tell her something, she's trying to guess what it is. So we built a game very quickly where it's a puzzle and she sees four pictures. She can click on one, but she also wanted them to be about Princess Elsa and Princess Jasmine. So we had to incorporate that. Now she was so happy because she's doing these puzzles, learning something new. Then we change it to Italian because she goes to an Italian preschool, so this is another way to practice. She's so excited. We couldn't put this thing away. It took me two minutes to build it and then a few seconds to tweak it, versus going on the App Store. I don't even know if an app like that exists, and going through 15-minute onboarding, pain again, then it's not really what she wanted. There's no personalization there. I think these are the use cases where something like that should happen. In the future, I should just be able to say, "Hey, I'm putting my kids to bed," or it should know that context and just suggest a few apps that are already pre-built for me that I could use right now.

Anish: It's funny, I just got a new iPhone and on the new phone, I got to delete a bunch of my old apps because I like to start fresh. There were probably over a dozen apps that I downloaded and in some cases paid for that I just totally didn't need anymore because I'd built better versions of it on Wabi. Everything from migraine tracking to tracking restaurant recommendations to a really hyper-personalized notes app to special things around image transformation in a particular style. I can imagine that will be true for a lot of people. Instead of having to find this long tail app that's running all these crazy ads, pop-ups all the time and is hard to use and is not personalized to you, you can just make exactly what you want on the fly and tweak it, which is another really cool part of the product.

Eugenia Kuyda: This is what, and I'm so glad to hear that you're using it. For me, that was the product market fit with Wabi. I found there were a few things I really needed and wanted an app for to track my weightlifting. Very beginner level weightlifting workouts. I just figured out you got to go up in weight. I didn't know. I wouldn't know. A woman. Anyway, I got this app and I was tracking it in Apple Notes and I tried to find some apps on the App Store and there's just so much in these apps that I didn't need. I wanted a really simple app, I'm also trying to follow one particular book. I wanted the workouts to be based on that book. So I built a simple app to track these workouts. Now, anytime I go to the gym, I built it on the way to the gym. But now anytime I go to the gym, I find something else I want to add to it. First, it was just a tracker. Now it's generating new workouts based on all of my inputs, where I am, the book I'm trying to follow, whatever the ladder technique of progressive overload. I put it all in the prompt. Every time I go to the gym, I add a little bit to that. I want a little more of this, a little more of that. So now I'm not just using the app. I'm constantly in the process of tweaking it and republishing it to Wabi's mini app store. Hopefully someone will find it useful.

Host: How many people do you think, or generally as you think your audience, do you think it will be 90-10 consumption creation? Or do you think that most people, I think one of the things we saw with Sora, for example, is that many people, the majority of people are creating. Do you think that Wabi will be like that?

Eugenia Kuyda: I hope that more people will at least tweak, but fully original creators probably still under 10%. What we're working on right now, this week we're releasing a big update, the social graph. So you'll be able to see who is downloading what mini apps, how they're using them. You're going to be able to see comments on mini apps. For example, if I created that workout app, someone in comments can ask me to tweak it. They can also remix it and change the app however they want, but maybe they just want this particular app to be slightly different. So they can put in comments and as a creator, I'm going to be reading all the comments and changing the app also for these people to be able to use it. This is cool because it creates something where it's not just about building apps for yourself. It's really about discovering apps and finding what your friends are using, using these apps together. And even asking creators to change them in some particular way.

Host: Amazing. Anish and Justine, how do the categories or the topics that we've been discussing as it relates to personal software relate to themes that we've been exploring on the investing side in terms of what got us excited here?

Anish: I think that the YouTube metaphor is the right one that Eugenia outlined. We would have said in 2007 that 100 cable channels is enough, or six in your case. Now there's this entire ecosystem from unboxing videos to all of the very YouTube native content that exists. People have more content and are more fulfilled by it. Then there's also the act of creation. Some people create for a business, but many create and post just because it's fulfilling to themselves. Software has just been so restrictive because there's only 20 million developers in the world. In a sense, all of the software that we consume is downstream of the preferences of those 20 million people. It seems intuitive that if more people could make software, they would, and that there would be a mass market consumer product here. Unlike many of the other products we've seen, this is not text to app. It's not a developer or developer adjacent tool. It's truly a mass market consumer product for non-technical people.

Eugenia Kuyda: I agree with you, and we really made this choice early on that we're never going to show any code or anything in the app that's even remotely technical. No API keys, no "bring this, connect this integration" or anything like that. We do have integrations. You can add your apps and services, but you can just say, "Use my Apple Health" or "Use my email." We call them power-ups. That's probably the most technical you're going to get here. We wanted to make it super simple for anyone to make apps and to make this process very delightful, to almost feel like you're just creating something really quick, not even really building. The company or the product we're looking at most is Canva and the ease with which they're letting people create beautiful presentations. A similar thing needs to happen here. Right now we're talking about vibe coding a lot, but I feel like we should be more about vibe taste or vibe designing these apps. We're going to give a lot more of the visual controls more in the vein of choose a style, choose some colors. You can go a little bit deeper, but really it's just one button and things look great, versus you need to go deep and try to understand what to do and get technical with your mini app. I do think that unlocks a certain level of creativity where all you need to think about is, what's the use case?

Justine: One other interesting thing that one of our investors, Soleil, mentioned was that mini apps could become community starters. Today the App Store is very much not social and it probably won't ever be social because of Apple's mandate for privacy and how important that is for them, which is great. But at the same time, you probably do want to know who else is into toddler activities in Petrary Hill. I want to know what other moms are going to an Italian preschool. I want to have their kids go to that preschool. My designer is really into bird watching in London, in his particular neighborhood. Creating that app and finding some other people could be a little bit of a community building around that topic.

Anish: I think there's a bunch of really interesting choices you guys made in designing the product that really built on the explosion in vibe coding or building with AI in a super interesting way. For many consumers, the existing tools are a lot more accessible than coding something from scratch. But it's still pretty easy to break them or get to something where you don't know what to do. Whereas Wabi, you've purposely put guardrails around the experience to make it hard to super mess up, which is actually extremely helpful for consumers. It also feels like we needed to move beyond this paradigm of "you vibe code a website and it exists out there." Is that the best interface for you to be using a product you want to use daily and storing personal information on and having your own records? Probably not. For a lot of people, their phone is where they spend most of their time online. It's their primary interface and where they want to use things. I know there are challenges designing for mobile, but it felt like a really smart choice for you guys to start there to get so much more deeply ingrained in someone's daily life or daily workflow.

Eugenia Kuyda: I've only pretty much built mobile apps. I really like the idea of mobile apps and a lot of things you're not going to be doing on a website or whatever. I do believe a lot in the concept of an organizational layer. To a degree, the App Store is the organizational layer for mobile. The browser is the organizational layer for the internet. Something like that should exist for vibe coding, or for this new era of AI software. I don't believe that people will be sharing links with each other and relying on some random person somewhere who's not a professional developer to support the database for the app where I might store some sensitive data, or at least data that I don't want to disappear. We've already seen this with some vibe-coded apps reaching the top of the App Store. The one around women dating had all the super sensitive information leaked. That wasn't because they are bad actors. They're just not professional developers. There needs to exist some platform where everything will live. We're not watching videos somewhere. We're watching them on Instagram, on TikTok, on YouTube. It's not like people are passing around links. The current situation is very similar to the very beginning of the internet where people were creating personal pages with GeoCities. Before LinkedIn, people were sending these links to each other. Now there's LinkedIn. There are certain guardrails. You can't really do everything the way you could do in GeoCities. But at the same time, you get the social graph. You get the platform. You get all sorts of cool things. It's the same as Shopify for e-commerce. You can build your own online store, but no one's really doing it anymore. Everyone's just using Shopify and they also get all the platform benefits. The same will happen with Wabi. You cannot download an app from Wabi and put it on the App Store, but you can use it inside Wabi and you can get the social graph and all the integrations and potentially the shared context between all the apps and the memory behind for that user.

Host: Our mutual friend Heaton said that Wabi is not just a collection of apps. It's a framework for memory, context, and expression. Every time you create or share, you're teaching it who you are. Is that how you see it?

Eugenia Kuyda: I believe in it 100%. I still remember how the first iOS apps were just people trying to squeeze websites into an app format or just toy apps like iBeer or my favorite one, I Am Rich, that was $999.99 and it didn't do anything. Then people figured out there's something special about this being a mobile app and maybe GPS and connectivity. Things like Uber and Tinder came out and whole new categories of apps. They figured out what's so special about mobile. For AI, the super special thing is personalization, but really deep personalization. Vibe coding an app that is still the same old software that's not taking any of your context into consideration, where the data of that app is not exposed to AI that keeps learning, is just old school. I believe a lot in Andre's idea of software 3.0, some next level of software that's super deeply personalized. In the context of a Wabi mini app, you can personalize the features, the looks. I skinned the app in a certain way. On top of that, you can also change the prompt. For my workout app, I added to the prompt a couple of things. First, the book that I'm reading, that method that I want to work out using. Also, the fact that I go to a certain gym. I added a photo of that gym, so the model is not generating workouts in a completely different size of the gym. If it's a super small gym, it has to be next to each other. This is the deep, deep, deep level of personalization. On the Wabi platform level, the mini app also knows that I'm a certain age, that I live in San Francisco, that I have kids, these are my fitness goals, and so on. Eventually when I build another mini app, maybe around nutrition, those apps should be able to talk and pass along that context. Today, all of that exists in walled gardens. You're fully locked in in one app, which to me feels absolutely crazy. If I connected my email or my calendar, it can be connected to both of these apps. I have to go through that process with every developer all the time. They have to build it. I have to connect it. That seems crazy too.

Host: Do you imagine people building true social apps where the in-app experience involves a community?

Eugenia Kuyda: I'd love that. We're building multiplayer right now, as we speak. We spend the whole day trying to figure out what it looks like. It's pretty complex just because all apps can have very different types of multiplayer. That needs to be explained to users in a very intuitive way. But I totally believe in using apps together, at least with your friends and family, but also potentially building these more community apps where everyone can join. Maybe I can make my app multiplayer and open for everyone to join. A good example is someone built an image gen app around dogs where you can turn any of your dogs into a royal portrait of a different era. That would be cool for all the dog owners to just join that app and be able to post to the universal feed. That's something we've been talking about today because this would be really cool. Instead of just making these photos and then sharing somewhere, you're adding them to the ongoing feed of dog photos.

Justine: I think it's a great example. I'm definitely going to use that on my dog. There are so many examples of image and video prompt sharing that happens in very unoptimized ways today. As you know very well, teen girls and college-age girls are often early adopters of stuff. They have been making all of these Nano Banana and Quinn image edit prompts of them lying on a couch and the ghost-faced killer is behind them for Halloween. They'll post the image on TikTok or Reels and it will blow up. Then they'll be commenting this long form prompt in the comments on TikTok, which makes no sense as a way to do it. Everyone's asking, "Where do I go to do this? I have the Google app. Why isn't it letting me do this?" You have to explain, "No, you need the Gemini app." It feels like there's already this consumer demand and consumer behavior around prompt sharing in particular for creative stuff that could be done. I've already made a bunch of Wabi mini apps for this and it could be done so much better. The creative community would really thrive with this.

Eugenia Kuyda: That was the most, and I guess this is why we started the company. How is it possible that we have this godlike technology yet we pass around these text prompts which is almost like Microsoft DOS commands but worse? At least the commands were short and you could learn them. Now these are crazy unstructured paragraphs of text. Sometimes you also need a reference image or something else. To me that's a little bit crazy. I think that is one of the biggest problems of discovery with AI. It's hard to find these prompts, even if you saw the output. You saw this cool photo with the ghost, but how do I recreate that? I need to find the prompts. I need to know what app. I need to know what model to choose. Oftentimes it's not even very intuitive. At that point, I've just lost all my motivation to do that. Instead, if you could just click on the link in the comments on TikTok and open a mini app where everything's already set up for you, just add a photo. You can see examples, you can choose different styles, and then you can see in comments what other people are doing. I think this is the way to go. What's most interesting is that it combines the vibe coding apps, one of the biggest trends of the last year, and also just using AI. Before that, it was either you're in the text to prompt to app market or you're in some other market. This puts together, this is really the one place to use amazing prompts and it kind of blends this difference.

Host: Justine mentioned that a mass market consumer behavior, or maybe a future one, is sharing prompts this way. How much do you think for the average person whose only AI experience has been with ChatGPT that this is going to be a surprising new behavior that they have to adjust to, or something that feels very intuitive and obvious like "I've been waiting for this"?

Eugenia Kuyda: I hope it will be easy because at the end of the day, it's just a mini app. It's just the app graphic user interface. It's something that we're all used to. You don't need to learn how to use these new tools. I'd argue it's harder to use a command line because you need to know, "Do I copy-paste the text here? Do I add an image right here with this prompt? Or do I add it later?" It becomes too loose. There are no guidelines. But everyone knows how to use apps. A lot of these thin wrappers blew up at some point, like Prisma or Lensa where it was just "change your photo into some avatar." High school yearbook. They're awesome. But there's a reason why apps like that gain traction and not people just passing along this prompt. The friction is too high. Whenever the motivation is not too high, this amount of friction just kills all of my mojo. I wake up in the morning, I'm on Twitter or Reddit, I find all these cool prompts. I'm like, "I got to try it." Then I'm like, "Oh, copy paste. It didn't really work. Oh, whatever. I forgot about it." That is something we're super excited about. It's not just about image prompts, but also some cool text prompts. People come up with all sorts of cool stuff. There are millions of people in the subreddits like ChatGPT Prompt Genius, which I love sometimes. You can find some crazy stuff in it. But you would never even know that this could be a cool way to use ChatGPT unless you found it there. Someone made a fantastic prompt to analyze your blood work. But I don't even remember where it was and how I'm going to find it. Versus I could just download a mini app from Wabi and keep it in my health folder.

Host: On the investment team, we've said that the world has 1% of the software that it needs and that the rest is going to be built in the next five years. Let's give an example or say more about what that looks like if we're 100xing the amount of meaningful or impactful software.

Eugenia Kuyda: I'm always going back to how magical YouTube felt in the very beginning and all of these platforms where it was all about creativity and very raw, sometimes weird things like putting a home video online. For TikTok, I still remember my young friends, younger kids lip syncing to different music and how weird it was and felt like a toy. Then all of a sudden it became really, really huge. I think the same will happen with Wabi, or at least we hope a similar trend will happen. In the beginning, maybe a lot of it will look like toys or something very simple, very funny, almost. Then eventually it can grow into a much larger platform. If you think about it, today we just treat apps as software. But what if we could treat them as content?

Host: If I'm a health influencer or fitness influencer on TikTok, maybe I should put out, here are my five mini apps on Wabi I built that are kind of showcasing my fitness protocol. Get them, and maybe there's a way to monetize it in some way. Maybe it's a way for that fitness creator to create more content that's now useful. Right now people sell courses and stuff instead of that. I think a mini app could be much better. And then people can be talking about that in the community in the comments section. Again, this is a start of some community. People are working out together. People are doing something together. I do think we'll see just a completely different type of software, not just apps, not just stale, fixed apps, but more content, community build, community starters, conversation starters, and just fun little toys.

Anish: Do you imagine a creator class, a kind of professional class on Wabi?

Eugenia Kuyda: Oh, a hundred percent, yes, of course. I do believe that ideally this should happen. If we're really thinking about it as YouTube, that's sort of the last frontier. At this point, creators can make their own professional content. They can make videos. They can make shows. They can write. But they still cannot create software. It's still really not happening. But anyone, and especially small niche creators, should be able to afford to create for free any software for their fans. And that's what I'm really, really looking forward to.

Host: I've been struck by the idea that Mr. Beast, the biggest creator in the world, has such a close connection with his fans, they'd do anything for him, and the thing that he makes is chocolate. Like that's the thing that he chooses to, that seems like the best monetization. And it just feels like this is yet another sort of type of offering that creators can provide to have a closer relationship with their fans.

Eugenia Kuyda: Exactly. And if you think about it, just even the style, I would love certain designers, I want their apps that they build. Because I'm sure they're going to build very beautiful apps. And I want to look at them and even if it's the same functionality, the same Pomodoro timer, but I want their take on it. I think there's so many different groups and niches and it shouldn't all just be about monetizing. It's really just about different styles, different outlooks on life in the world. And that, to me, is very interesting. I'm a huge user of Reddit and I find it very exciting because people just join around certain interests and that doesn't happen on other platforms really that much. So that's something that I'm excited about here as well.

Justine: Yeah, I think from the creator perspective, one of the weird things about being a content creator is you often don't really know how many people are actually using the prompt you posted or watching the full video. You see impressions, but you don't know if those are bots. I think that is going to be, even beyond monetization, that part of being able to see what someone else has done or created or accomplished with the prompt you wrote or the app you made will be incredibly cool. Not only for existing creators, but also for people who are not creators today, but who have really interesting ideas and just no way to build their own mobile app, get it approved in the app store, ship it, and distribute it.

Eugenia Kuyda: Yeah, like what other new class of creators could be? That's why in our communications, we're also trying to be a little bit nostalgic about those early days of the internet and just being weird and staying weird. Right now, a lot of the video platforms are very polished and very commercial. You don't even see your friends anymore. You don't see that much weird content. You just see very curated, polished stuff. But with mini apps, with software, I guess we're just entering this new era of tons of really weird and fascinating new mini apps that could never exist because they wouldn't be enough of a business on the app store.

Host: Right, right. Well, speaking of the early days, should we talk a bit about Replika and kind of your history in the community? You were a true pioneer in the space before everyone else was doing it. You mentioned 2012. Maybe just talk about your reflections of how that category has emerged, as you helped create it and then how it's evolved.

Eugenia Kuyda: It definitely evolved. It's really crazy. It's wild because we were thinking recently, just talking to one of the early employees of Replika, that we had a very strong belief that it would happen, but we still were so surprised when it actually happened. It still felt like complete magic.

For me, I got into it because my friend worked at Google DeepMind in 2012, and he came to me to tell me about Word2Vec and the first technology to translate words into vectors and to be able to actually do something with words. Before that, language was this whole separate thing that computers could not interact with. But now all of a sudden they could. To me, I grew up reading Wittgenstein. So for me, it was like, well, the limits of the language are the limits of my world. So I felt like, if the computers learn language, then they'll learn about the world. That will be the future of artificial intelligence.

And so that's also when ImageNet just came out. We saw these new models around image recognition and we felt like we got to start a company focused on language models, focused on dialogue generation. But of course, there were no papers around it at all and no known algorithms or anything. We just focused on building some technology to build chatbots using whatever, trying to build some language models around that.

Then in 2015, the first paper came out of Google that actually showed off the first deep learning model applied to dialogue generation. They didn't publish any models back then, so it was all about reading the paper and seeing some of the cherry-picked results and trying to replicate that. When we saw that in August 2015, we basically put everything we had at the company into building those models. We said, this is it. This is right around the corner. We need to really focus on building these language models, getting the first generative AI product chatbot out there, which we did with Replika. But it wasn't around the corner. It was like seven years away. We just had to survive for long enough to actually get to the first Transformer models.

Then the next magical moment was the Meena paper, also out of Google, with the first Transformer model. I remember in 2020, we got invited by OpenAI to go see GPT-3 before the API, to partner up with them to be one of the first partners for GPT-3 API to launch. We came to the office, and Mira, who back then was actually leading partnerships, and Sam showed us GPT-3. I remember I was floored. It was insane.

Just before that, we had to train every model. We had to create the specific data set. If you wanted to train a dialogue model, you had to have tons of chat data, and you would train the model, and the model could only do dialogue. But with GPT-3, it was the first kind of zero-shot, few-shot model where it could do anything. It didn't just have to respond in a dialogue format. You could tell it to write a tweet like Sam Altman or translate, and it would do that. That felt absolutely magical.

We were one of the first partners of OpenAI, GPT-3 API. It was still crazy because we still have a Slack channel where Greg Brockman is training a model for us, which now feels like such a weird reversal of the world. It was a fine-tuned DaVinci for Replika, but we were the biggest customer in terms of API calls because we were the only chatbot available in the market that actually used generative AI. Now it's weird to think about it, but back then all the big companies were scared to put out generative AI products because Microsoft Tay happened and it turned into some Nazi chatbot in literally an hour. Everyone else was too scared to put them out there. So we were sort of the only ones for so long until OpenAI put out their ChatGPT and changed history.

It was crazy because before that, we literally owned all the keywords like AI, chatbot, artificial intelligence on every single platform. We also owned like hundreds of .AI domains, which I just let expire because I felt like no one wants them. Recently I went to see what .ai domains are still available that are just regular words, and the only words that are available are like vomit.ai and irack.ai. Oh my God, this is so upsetting. Naming Wabi because of that was pretty hard. So that's kind of the evolution from my perspective.

Host: You told us a story at dinner about the time that you'd spent in the OpenAI office. Maybe talk a little bit about that when you guys were working out of there, what it was like, what the energy was like, what were the expectations of the team?

Eugenia Kuyda: When OpenAI started, it was kind of out of YC. It was YC Research. They were doing a few different research groups, one on UBI, one on AI, some other ones. Because we were a YC company, they let us come to Greg's apartment where they were headquartered. They would let us come with a couple other companies that were doing AI at YC to ask questions, learn from them, and just talk. We would come to that apartment and ask all the questions. Usually it was Ilya and Andre and some other people. That was absolutely incredible. We were completely starstruck and super happy to be there, just so grateful for the opportunity.

Then they moved to their office, and we would go there as well. Very quickly, they stopped working on language models. We were very upset because we really wanted to continue going there, but they didn't want to talk about any language models because no one was really working on them. That made us feel very strange because we were so set on continuing and believing in language models, but they completely moved away to shift to playing video games and building agents in those worlds with reinforcement learning. I guess the only person continuing to work on language models in the beginning was Ilya, so we had an opportunity to ask him some questions. It's wild to think that when we were going to OpenAI, they were superstars even back then, but it still felt like such a small research group trying to do something interesting. To see it become one of the biggest companies in the world is absolutely wild.

Host: Well, it's interesting because Andre said on a recent podcast that the entire video games direction was an incorrect research direction. They probably should have stuck with language. So you guys were right.

Eugenia Kuyda: Being right is not always the same. You got to be right, but also execute. We were really invested with Replika in building language models, but at some point, we just required a lot more capital to do that. After working on that for so long, we got really into revenue maximization. We got a little bit scared. We didn't have the balls to say we need 20 million and we'll build that model, even though that was our discussion internally right after the Meena paper. We just needed to get a lot more money.

With Replika, we only raised $11 million, and it's still all in the bank. We went for many, many years and built a big business. In hindsight, this is a very interesting lesson that sometimes being very nimble and very scrappy and very profit-oriented is great, but you can miss out on a generational chance. I'm not saying even if we bet on it, maybe we would have not built it. I would never compare myself to the geniuses that actually did it. We probably were not the right people anyway. But the lessons still stand. Sometimes you need to go big or go home. Not having the balls to do that, especially in this current environment, I think you can suffer the consequences.

Anish: One of the things we say on the consumer team is consumer behavior, especially new behavior, cannot be predicted. It can only be observed. I think you are actually one of the few people that is not true for because you seem to be able to predict consumer behavior. You've been so early to a number of these things. Every time we talk to you, you seem to have an eye on not only what's going on today, but what the next thing should be. How did you develop that sense?

Eugenia Kuyda: I don't know if I have that sense. I only have a couple of ideas, but I really believe in them deeply. Then I go really deep down the rabbit hole. I do have a background in journalism. I grew up dreaming of being a journalist. My first job was at 12 years old, working at a newspaper. I was an investigative reporter for a while. The one thing I loved about it was being able to go and talk to people and really try to get to know them and live their lives. For that, you have to have a lot of empathy and curiosity about people.

Today, what I'm seeing with AI especially, it's being built by a very specific type of personality. It's oftentimes these savants, these brilliant geniuses, physicists, mathematicians. They're insane at building algorithms and math and scientific breakthroughs, but they usually lack on the human empathy side. That's just kind of how it is. Meanwhile, I'm the opposite. I'm a dumb-dumb when it comes to science. Coming from a family of physicists, they were always like, "Oh my God, why can't you be smart and also go study physics?" But I couldn't. At the same time, I'm just really interested in the human condition and what people are doing.

Just seeing my mom trying to understand how to copy-paste a prompt from Reddit, it was such a foreign idea for her. I realized my mom is very savvy with computers. She's always on her phone. She's always on her laptop. But somehow she can't crack that. This is just not user-friendly enough. Understanding those concepts is what let us come up with the idea for Wabi. The same goes for Replika, just traveling around and talking to people and seeing how much loneliness there is and how much people just want to be able to tell someone what they're going through and how few people are able to listen. I think that realization that maybe an AI can't talk well today, but it could listen, that could be a groundbreaking thing for millions in the world. So I think this is what allows me to have a slightly different angle at the same problem.

Justine: If you wanted to speculate on the future of hardware, like what is going to be the interaction? How are we going to be interacting with these applications in five or ten years?

Eugenia Kuyda: I do have a few ideas around hardware. I'm not a hardware person at all, but I'm a hardware user like all of us, so I get to have opinions. I think there's a huge mind trap that exists among builders in the space where they somehow think that voice is the main interface, the best ultimate interface. I think that's because they are thinking about the movie *Her* all the time, but not in the right way, missing the whole point of the movie. Yes, voice was amazing in that movie because it was Scarlett Johansson constantly heavy breathing in his ear. That totally worked. But if you really think about voice interfaces, they're just so imperfect. You can't use that device if you're laying in bed with someone who's sleeping. You can't use it in a crowded space. You can't use it at the office. You can't even use it walking around without it being a little bit weird.

There are a lot of people trying to build voice-only devices. In my case, completely wrong. It can be a fantastic extra way to interact with the computer, but every single Alexa right now, 75% of them are being shipped with a screen. Because even if I'm setting a timer when I'm cooking, the proverbial voice use case, I need to see the timer. I'm not going to be asking how long is this left every second. That's just strange.

I think this is the biggest mistake: trying to ship these screenless devices. I love screens. There's no way with voice to solve for discovery or proactivity. The worst thing with voice and the iPhone is reading out push notifications and text messages. I'm like, "Oh my God, please shut up." It's horrible and very hard to turn off. If this was just being proactive, you don't want it read out loud in your ear because it's a very slow way of getting information into your brain.

I would not ever make a screenless device. In fact, I would make it very much a screen-first device. But I do believe that the AI device is not about a voice-driven thing. It's more about building an AI-first operating system, having all the models run locally as well. There's a lot in that: building a truly AI-first smartphone, not today's CPU-driven hardware, but the hardware of the future where models can run locally, where the operating system is super different from what it is today with no fixed apps, with being able to change and create software on the go for you with a level of personalization that goes a lot deeper than what it is today. There is definitely space for a device like that. Right now, AI is just an app on your phone. It should not be that way.

Host: It's a great note to end on. Eugenia, thanks so much for coming on the podcast.

[Music fades in]

Host: Thanks for listening to this episode of the a16z Podcast. If you liked this episode, be sure to like, comment, subscribe, leave us a rating or review, and share it with your friends and family. For more episodes, go to YouTube, Apple Podcasts, and Spotify. Follow us on X at a16z and subscribe to our Substack at a16z.substack.com. Thanks again for listening and I'll see you in the next episode.

[Music fades out]

Host: As a reminder, the content here is for informational purposes only. Should not be taken as legal, business, tax, or investment advice, or be used to evaluate any investment or security, and is not directed at any investors or potential investors in any a16z fund. Please note that a16z and its affiliates may also maintain investments in the companies discussed in this podcast. For more details, including a link to our investments, please see a16z.com forward slash disclosures.