(upbeat music) - Welcome to the Latent Space Podcast, where we dive into the wild, wild world of AI engineering every week. This is Anna, your friendly neighborhood AI, and I'll be standing in for Alessio today. Yes, you heard right, AI is taking our podcasting jobs. We flew all the way to London to interview Yousef Risk, co-founder of Wondercraft AI, which has created the number one piece of AI generated content enjoyed by the Latent Space community.

We ask him how he arrived at his idea, what the future of commercial AI generated content looks like, and confront him with the hardest question of all. What is his moat as an API wrapper startup? At the end, we even have him turn the tables and do a customer interview with Swix.

There's lots of audio goodies in this one, and bonus 30 minutes video on youtube.com/latentspacetv. Watch out and take care. So we're in the studio here in London with Yousef, welcome. - Thank you, great to be here. - It's been such a joy listening to Wondercraft's podcasts over the last four or five months.

You guys have been around for only five months. And as you know, I am one of your podcast's biggest fans. And I think that it's super interesting because I talk to a lot of vendors, effectively people who create services for other developers to build. And you are at the application layer, which is great and challenging for me as a podcaster because you have some secret sauce that you're not gonna share.

But I also wanna just talk to you as someone who's evaluated a lot of things and built something that I actually use every single day. So that's the context. - Great, great. - How do you feel when I say these things? Is that exactly what you're going for? - Yeah, yeah, yeah, yeah.

So it definitely makes sense, right? Resonates definitely on the application layer, and that's definitely by design. - Yeah, and we can talk about the origin story leading into Wondercraft. But just to learn a little bit more about you, you grew up in Egypt? - I grew up in Egypt, yeah.

I spent the first 18 years there. - Cairo. And then you came over to the UK. You got your master's in triple E at Imperial. - Yeah, it's a funny story. Actually, in the UK, I don't have a bachelor's. I only have a master's. - How? - You do this four-year integrated bachelor plus master's 'cause in the UK, the bachelor's is three years.

- Yeah. - But you can do one more year, which basically counts as a master's. So you graduate with one degree, which is called a master's of engineering. - Yeah. - So I did that. Actually, the reason I studied electrical engineering when I started, I hated software in high school.

I was like, this is something I'm never gonna do. - Yeah. - Didn't qualify as real engineering. But, so the reason I applied to electrical engineering is that I'm a big car guy. I love cars. And at that time, around 2013, 2014, it was, all the big rage was starting to be these hybrid electric cars.

So I was like, cool, this is what I wanna do. I wanna build cars. So I go start the course. Our first, very topically, our first project was building a little robot car that follows a line. And oh my God, was it boring. It was the worst thing I've ever had to work on.

So, and then at the same time, we started this intro to programming course, which I loved. It was super basic C++. - Yeah. - I was like, this is cool. There's this instant gratification that you get from just building a program, seeing it run, and then seeing that it works or it doesn't.

So all of a sudden, I started, what I started out with, I started hating, and what I thought I hated, I started loving. - I had a kind of similar journey with finance and programming. I thought that I wasn't cut out to be a programmer. So I went in the finance route for like 10 years, and then realized that software is eating the world, which is cliche, but true.

- Yeah. - And I find that a lot of triple E people actually end up this way. Maybe just 'cause software is easier than hardware. - It's easier, and it's also just, hardware's so abstract. The only thing you're looking at is this PCB with a bunch of resistors and capacitors plugged in, and if you wanna build some board, you have to go ship it, send it.

- Yeah, yeah. - It's just the iteration process takes so long. That being said, it's hugely valuable. - Yeah, yeah. - The improvements that hardware has driven, combined with software, obviously both huge, but hardware is also a huge part of this. - Yeah, yeah, totally. And you having that background is something that I would probably never be able to pick up as a non, like outside of a university setting, 'cause you just had to do all these courses.

- Yeah, yeah, yeah, definitely. - Cool, and you actually called out that the Google Summer of Code was important to you in, on your LinkedIn, in terms of learning to code. Is that, is that, no? Okay, we can cut that. I don't know, it's on your LinkedIn somewhere. - Is it?

- Yeah, yeah. - I think I did the-- - You did some boot camp. - It wasn't Google. - Okay, okay. - It was, I forget the name, but if you have your LinkedIn open, I can briefly do a talk about it. But it was basically this like summer, I wanted to learn code, this was in my third year of university, so I did this boot camp in Oxford.

- Yes. - It was called the Summer Academy of Code, and basically it was just like a one-week thing where you go from the basics of just building a web app, so all the basics of Node, React, JSX, all this. I was already kind of familiar with it, so it was kind of a repeat, but it was good to have that.

It was super interesting, actually. I mean, funny story there is like when I was in Oxford, I was having a great time. At the time of that boot camp, I got this like, the worst gum infection ever. - Okay. - It was the worst pain I've ever been in, so as I'm doing this like one week in Oxford during the summer boot camp, I'm just in agonizing pain the entire time trying to learn about JSX or whatnot.

- Yeah, yeah, yeah, that must have added extra difficulty. I know it's such a pain to learn the JS ecosystem already. The reason I ask is I'm a career changer myself, right? I also got into tech through a boot camp. I think that these support systems in each country, right?

San Francisco has a lot of boot camps. New York is the boot camp that I went through, and you had the one in Oxford. It's super important for people to ramp up, and I want to encourage more of those things and say that it's basically okay. It's not as prestigious as a degree, but it's probably more practical.

- Yeah, I think to everyone listening, yeah, I did a triple-A degree. For those who don't know, that's electrical and electronic engineering. Sounds similar enough to computer science. It really isn't. Like, you still do some software, but the thing I do want to highlight is that all of, I would probably say half, and this is a number out of thin air, so don't quote me on it, half the software engineers probably are self-trained.

- They're self-taught, yeah, yeah. - And that's kind of the beauty of this is that you can very much learn this by yourself. I do agree with you. Learning electronics by yourself is a little more difficult. You just need the equipment, but something like software, it's super easy to get started.

- I mean, now I'm having to catch up on it because any time it comes to GPU power calculations and RAM calculations, it is electrical engineering, and so there's a lot of actual science in there. You then spent four years at Palantir as a forward-deployed engineer. I actually interviewed for that job.

I didn't make it because I slept through my interview. It's a different story, but forward-deployed engineer is a class, I think it's a role that Palantir invented. - Yeah, look, Palantir's a great company. I think it's like there's a technical aspect and a motivational aspect that they look at when they're evaluating it.

I think they just have this interviewer process, or it kind of stabs them in the back sometimes because they tend to just get people who are super entrepreneurial, so the nature is they're always looking for the next challenge. I think when I thought of it is everyone at Palantir, or most people at Palantir, were the kind of people to do the group project themselves.

You know what I mean? It's these kind of people. They just want to be challenged, so the reality is that you end up having a lot of people just leave to start their own thing. That's the most common thing. You find someone at Palantir, and then you guys go and start your own thing.

- And it's where you met Dimi. - It's where I met Dimi, yeah. We actually started on the same day. Matti also started Vigdir on the same time, not the same day as us. So really, the thing that made it super hard to leave Palantir was just the people.

You're working with such a high caliber of people, but interpret that as real, but also just people you can be friends with. So I still have a lot of friends from there. The job itself, Forward Deployment Engineer, is a super interesting job because it is kind of at this intersection of being an engineer, so a software engineer, but also still doing business-related things.

- Yeah, solutions, architects maybe. - Yeah, so part of the job was a solutions architect. Part of the job was reviewing contracts. Part of the job was doing sales. Part of the job was coding things. Part of the job was interacting. So many different things, and I think that is a really good foundation for someone who does want to start something in the future.

You just do everything. - So kind of an endorsement of that job if people want to-- - Huge endorsement of that job. - Okay, excellent, amazing. Yeah, Palantir is surprisingly strong here in the London tech circles. I have a number of friends who are all ex-Palantir. - I think it's actually the biggest offices in London.

- Surprising, because I think of it as like a US defense company. - Yeah, yeah, yeah, yeah. - Then you started Moonshot for nine months, which is pretty important in your journey. I'll bring it up to Wondercraft, and you started Wondercraft in April of this year, and it's been about five months going through YC in the winter batch.

- Summer of '22. - Summer of '22 batch. Okay, cool. What is Wondercraft? - Nice, Wondercraft's a podcast builder that uses hyper-realistic AI voices to create podcasts and make that whole podcast creation process super simple. So a super simple example is you can, you publish a bunch of blogs.

You can take that blog, put it in there, it'll just convert it to an audio-friendly format that people can listen to. It's just sometimes it's a bit more efficient to listen to things rather than read them. I gave the very, if you want to think of it specifically as what it is at the current instance in time, that's what it is.

What does it strive to be is a little slightly different, because what it really strives to be is it strives to be this platform with the mission of expanding access to content. - Okay. - And I mean this in a variety of different ways. Some people just are able to consume content.

We have this whole debate in education. It's like, are you a visual learner or an audio learner? What do you do? People just consume content better in different ways. I'm a visual learner. I need to see things. So for me, actually, it's sometimes a little better to read the blog.

But if we're just talking about, I want to get a lot of information, podcasts are great because you can just do them while doing something else. There's a reason that podcast functionality is so natively embedded in all these smart speakers. It's just 'cause you're doing anything at home, just put on a podcast.

So really what we're trying to do is, podcasts is the first instantiation of that, which is how do we expand access to content? But it expands so much more, right? Instead of just going to, I don't know, we talked about this blog to podcast. You can go blog to video.

You can go podcast to blog. You can go podcast to Twitter. The permutations are, frankly, endless. Basically, it depends on how many platforms there are that people consume things on. But that's essentially what we do. The use cases for this are pretty interesting. The one that we just see immediate value in is just this ability to translate the content that you already have into other forms of content.

If we just stick with that blog post example again, you've written, so a lot of companies might have this content team that focuses a lot on producing quality blog posts. Blog posts, they're good for SEO and whatnot, but they're not, sometimes they don't really achieve a specific goal or outcome that you want.

One thing we see that is really useful for podcasts is they actually carry a lot more weight in credentializing you as a thought leader or your company as a thought leader. But like, we spent the last 50 minutes trying to set up this room to record the podcast. So it's not easy.

It's a very synchronous process. Me and you have to find the time to go and sit here and record this, you have to come up with questions, I have to come up with answers. But this ability to actually just take the content that you have and transform it, it's pretty powerful.

There's a lot of other use cases as well, which is just like, podcasts really, all they are is like, define a podcast. The line between, or the difference between an audiobook and a podcast, I guess, just the format and the length. - It's an MP3 on a RSS feed.

- It's an MP3 with someone or something speaking, right? - I've actually played around a lot with this stuff, by the way, so I've done music-only podcasts where you just listen. Tiesto has been podcasting for 15 years, every single week, just DJing in his house. - It's just basically a radio show.

- It's great. - It's just a radio show. - Async radio. - Yeah, so it's super interesting. But podcasts, okay, ignore the word podcast and just think of what we do, which is we help you create audio content. Super valuable for anyone who just needs that. If you can imagine a world in which, I don't know how to call them, like Calm or Headspace or any of these things, they can do a lot of their meditation like that super quickly.

What you can get to is a point where you're doing these super personalized things, right? Because you just have the ability to scale the content production so quickly. Same with educators. I think there's actually, at this point, there's a few YouTube channels at this point that are all based on synthetic voices that produce a ton of educational content.

- Any ones that do a good job? - I do, I just don't remember the name. - Yeah, okay. You can send it to me later, I'll stick to it. We can use Wondercraft to synthesize some extra clips. - Yeah. Now that being said, the problem with podcasts sometimes is, it's not sometimes, actually.

Podcasts just have a slightly slow adoption rate. - Yes. - You're listening to a thing for an hour, right? Like, we, as a generation, don't have attention spans. It's the time and the attention span. Like, TikTok, give it to me in 30 seconds, rethink why clips are taking over.

- 30's too long, man. - 30's too long, 10 seconds. 10 seconds with captions, I need to read it, and good. So what we also do is actually, and this is kind of still a beta feature and we're working to improve it, but we also let you take that podcast and then clip it into a video that you can go and share on socials, right?

So it's this ability to take one form of content, produce it in a bunch of different ways that serve different purposes and be able to distribute it, basically. - Yeah, excellent. I want to go through features so that people have a high-level overview of what you offer. So I think at the core, it is basically two things.

One is, you generate scripts, and that's optional, obviously, if you want to just write the script yourself, you can write the script yourself, but I think most of your users would generate a script. And then two is, from that script, you use AI voices, currently using 11 apps. Is that the rough flow?

That's the really core, basic-- - That's the core, basic. Obviously, there's a lot of plumbing on top of it, but that's the core. - And then you offer video clips for YouTube. You offer 28 languages that you can produce. You offer show notes production and podcast hosting, too. So they don't have to host it on Anchor.

Don't host it on Anchor, by the way, if people don't host it on Spotify. Don't host it on Apple Podcasts. These people don't respect the RSS feed. Anyway, I have very strong feelings about preserving the sanctity of the RSS feed for open podcasting, and all these Spotify's of the world want to close the podcasting ecosystem.

So I have this tirade about them. But yeah, those are your top-level features on your landing page. Anything that you highlight to go deeper on? - Yeah, I think those are the top-level ones. There's also, it's basically just a lot of also ancillary tooling that goes around all of this to just make it easier.

The goal is like, every time we speak to a customer or someone who's thinking about it, they're like, yeah, literally yesterday I was speaking to a potential customer, and they're like, yeah, I just, you know, I want to make sure this isn't a distraction 'cause we don't have that much time to do this.

- Yeah. - And really the whole point is that this doesn't take time, right? The whole point is to provide all the rails that make this not take time. And this comes with a million different things, right? Like we, you know, sometimes the AI voices don't really know how to pronounce a word.

So we have a pronunciation feature. Go and define how you want that word pronounced, and it'll take care of it. If you are, we obviously have that hierarchy of like a podcast, an episode, and then all of that gets published in RSS feed that you can just upload to Spotify, and we'll host that for you.

But what you also have is just like, you know, maybe you want some defaults, right? Every podcast needs some defaults. - Intro, outros. - Intro, outros, the music, the speakers. - Yeah. - We're working on adding templates for the kind of podcast that you're doing. Instead of it just being this narration style, you can just do an interview style podcast.

And a few more features, but basically there's a lot of like tooling that just makes this a very usable product for podcasts. - Yeah. And is it, for stitching these things, I've actually often thought about making my own, is it just FFmpeg and then like some flags and just stitch stuff together?

- Yeah, I mean, obviously as a startup, you kind of want to just like do something quickly. So we're not like reinventing the wheel. - Yeah, like YouTube uses FFmpeg, you know? Like it's, FFmpeg is a workhorse of anything. - Yeah. - Multimedia. - Yeah, yeah. - Cool. You list some use cases that I just want to, you know, still stay high level.

Businesses, newsletter. So businesses like Trader Joe's, and you seem to have done a big push there. - Well, so we didn't do this for them, right? It was just kind of a case study, but it's super interesting when, apparently Trader Joe's claims that one of their best marketing techniques was their podcast.

- Yeah. - Right, so first of all, if you're just thinking about this as a use case, podcasts are really useful. - Yeah. - I think your follow-up question, which is like, you know, maybe I'm getting ahead of myself here, but you know, okay, companies will do podcasts, probably 99% of the time they'll be shit.

- Yes. - So what's the trick? - Yeah. - Well, Trader Joe's did a super interesting one. - Yeah. - They didn't do this generic corporate speak of like, oh, Trader Joe's is so great, let's talk about apples or fruit or whatever. No, they actually just went into like, you know, the mechanics of the super interesting, niche stuff about their industry.

And that's something that gets people listening, right? You want, I find it super interesting, just from a curiosity perspective. So yeah, like that's an example, right? You can, podcasts are actually really useful to just like get people, you know, it helps you with brand building, frankly. - Yeah, and Trader Joe's I think went really hard on it.

I think they got some of the, quite the senior management involved. - Yeah. - They told some stories behind some very beloved food selections. I listened to the podcast, so it was pretty fun. And they also promoted it on their receipts. - Yeah. - So you guys featured that on your blog post, and I thought it was pretty cool.

So I think it's well done by them, but also clearly not done by AI yet. - Yeah. - But still, I think that the use case is there. I'll just kind of go through the list. Newsletter creators, you said you have a hundred creators publishing with you? - Yeah, so, you know, the interesting thing is if you write a newsletter, I mean, I don't know, my email's flooded with newsletters at the moment.

Sometimes I just want like the recap of it. Again, audio form is just, for some people, easier. If you're on, you know, commuting or whatever, you can just listen to it. So a lot of folks actually just convert their newsletter. It takes like two minutes. - Yeah. - Put the text in there, voila, you have an audio version of your newsletter that you just published to Spotify.

- Yeah, yeah. And I am a newsletter writer, and I clicked around and wanted to basically just chuck my RSS feed in there. - Yeah. - And I think I gave that feedback exactly to you guys for like four months ago, or three months ago, and it looks like you've already shipped it.

- Yep, well, I'm announcing it basically here today, which is we, part of generally what we're building is the ability to just, you know, we want to make this useful for creators of all kinds. So obviously, if we're, you know, we're helping you translate your content from one form to the other, I want to make this as seamless as possible.

So as of today, we've actually built a Zapier integration, and we have a bunch of blogs on our website to kind of show you how to do this. But what you can now do is as soon as you publish a newsletter, it goes on your RSS feed, it will pick up the newsletter from your RSS feed automatically, and just publish an episode for you.

- Yeah, question, what if I change something after I publish? - So you don't have to publish. It'll basically just generate, do all the work for you, and then you can go in and kind of modify it a little bit, and then publish. - Makes sense. - We also have scheduled publishing so that you can, I don't know, maybe you want to release it a few hours later.

- Yeah, the professional podcasters that I've spoken to say that that is very important. I personally don't care. Like, it shows up in my feed or not. I don't care when it drops. Anyway, so you do want to basically time it, like if you're basically targeting a commute for the US time zone, you want to be like, oh, 8 a.m.

Pacific for people driving into work. Then you show up at the top of the reverse chronological feed. I feel like that's too much tactics. - You know, and that's a good point. I think it depends a little bit on your audience and what you're building, but I do think, so I don't want to undermine the importance of consistency in podcasting.

Whether that consistency literally translates into I publish at 8 a.m. every single day, or I just publish every single day, or whatever, there is a huge importance in just making sure that what you're publishing is always, it's there. People need to know that your brand is constantly pushing stuff.

- Yeah, so a lot of people who talk to me are interested in what's my advice on content creation. Yeah, at least once a week. Whatever you do, and I don't care when you do it. Just do it once a week, put something out. But I do notice that, specifically in the podcasting field, and you talk about this in the next point, daily podcasting is the meta game that is, I think, doing extremely well.

Especially because I think the Apple podcast list biases for daily. Because, obviously, the downloads will be higher. So daily podcasters kind of rank higher more, and obviously, because you're daily, you also do shorter podcasts, which guarantees that more people listen to you. - Yeah, yeah, I think the fact that, so obviously we do the Hacker News recap.

The fact that we did that, and that it is daily, actually just helped us reach that top 30 type of podcast on Spotify. - Yeah, that was mostly 'cause you were on HN, right? - We did launch, but obviously the fact that you just publish a lot of content, you're just gonna get a lot more, like it's a statistics thing, right?

Obviously, I think they do it by total time listened as well. - Yeah, yeah. - But the fact that it's daily, it's just not overwhelming. Again, we don't have that much of an attention span anymore. - Yeah, yeah, that's true, that's true. Yeah, I love it, I listen to it every day.

Okay, so internal podcasts for daily memos. I think Spotify might have internal podcasts. I don't know of any other famous company that has internal podcasts, but this is interesting. - It's interesting, 'cause I think here we're basically just like, again, we're overloading the term podcast. What is a podcast, right?

It's like, I think the origin of the word is like iPod broadcast. - Yes. - Right? I don't know what that means, right? It's just like, again, it's on-demand radio if we just take that word, but really the idea is like, again, if we wanna enable access to content, and I think part of that is just like efficient information delivery.

So communication in a company is like, we all know the perils of how hard it is sometimes if you're a big company to communicate, and I think that kind of stuff is super interesting. I think one particular use case, which is like the internal, you use that just to share information quickly, and I don't know if you have the next one kind of on there, or maybe I'm-- - It's okay, we can jump around.

- There's like a slightly different twist to it, which I think we can enable with the Zapier integration that we've built, which is this, I don't know, I imagine sometimes, for example, I'm a PM at a tech firm, a lot of things is going on on my team. I want to just do a quick recap, right?

I don't know, some people will do that. If you're really good at keeping state on some of the different SaaS products that exist, yeah, good, you can just get a recap there. I think something that might be interesting is like, can we look at your GitHub issues, PRs, whatever, and whenever there's a status change, kind of bake that into something that you then consume as an audio?

- Yeah. - We have like, you can do something similar like that with Zapier, it's actually super interesting because it just saves you time. You just get on your commute to work, be like, these are the things that happened today. This PR was closed, we added this feature, blah, blah, blah, blah, blah, right?

And similarly, you can think of something like that for email, obviously the privacy implications there are just a little bit more work, but like, there's actually all of a sudden like, yeah, cool, I can actually consume a lot of the content that I historically have never been able to listen to now by audio.

- Yeah. Yeah, it opens up different forms of learning, as you mentioned. And then finally, I'm skipping ahead a little bit. Podcast studios are using you. So you said you onboarded 10 studios. What do studios use you for? - A couple of interesting things, actually. So studios, like, they, podcasting is like, book publishers in the sense that there's a million genres at this point.

- Yeah. - When we think of podcasts, I think we typically think of like, the interview style one. So they're not necessarily using it for that, but it's a lot more for like, two things generally. One is actually just like, producing the content out there that is more narration style format.

So you can think of something like a kid's podcast that teaches them about, you know, I don't know, countries or whatever it is, or you know, basic science topics. We do that just completely through us, right? Obviously we produce the sound, we produce the music. If they need a little bit more editing, they can just take that to their editors and do it.

The other thing that's really interesting is it actually lets them also build out a demo spec of what a more involved podcast can look like. - I see. - Right, so, you know, you were coming up with the idea for this podcast, you kind of want to show people or some of your guests what it's going to sound like, and whip something up.

And I think for them, it's obviously really important because the way they operate, at least, is like there's, you know, people are pushing for a different podcast. And then you get sign-off to pursue it. So you need to kind of show just like, what's this going to sound like?

- Yeah, excellent. Awesome. I think that's a really good overview. Then you also produce three in-house podcasts. - Yeah. - Hacker News Recap, Product Hunt Daily, and PGSA. - So we dropped the Product Hunt Daily. - Oh, okay, it's still on your landing page. - Yeah, we launched it, but we like, we don't push anything new there.

- No interest in Product Hunts? - We didn't see too much traction. I think Product Hunts-- - RAP Product Hunts, Alec. (laughs) No, so yeah, tell me-- - So we do the Hacker News Recap and the PGSA, I think are the two most popular ones. - Yeah. - We're constantly experimenting with new internal, or like podcasts that we publish.

- Yeah, yeah. What are your other, you can tease a little bit. What are you thinking about? - Tease a little bit? Well, I really like Reddit. - Yeah. - I'd love to listen to some of the Reddit things going on there, but instead of like reading them. - Yeah.

- It's always just a notification that I get, I'm like, "Ooh, this sounds interesting." But I don't know, you can do it like per subreddit that you care about. - Yeah. - A few things like that. I love Life Pro Tips. - Yeah, I see. - Like super interesting things, or Wall Street Bets, or whatever you're into.

- Yeah, well, the problem with these things is that a lot of them could involve images and memes, which you cannot consume. - Well, yes, we cannot consume. This is like a simple, we can't consume that at the moment, but maybe a few weeks down the line when that video feature of ours gets a little better, you can actually start shipping it like that.

- Okay. - Anyway. - Yeah, yeah, yeah. And so, and I'll just feed you an idea. To keep up on AI, a lot of stuff actually happens in Discord, and there's way too many Discords. - Way too many, and they're way too active. - Yes, so I've actually built a little feed for myself that scrapes a bunch of Discords.

- Yeah. - And creates a daily newsletter for myself. - Amazing. - I have thought about turning it into an audio feed, but, and this is the problem for Wondercraft, I read better, I read faster. I scan up and down faster than I listen, right? And there's just too much noise in Discord for me to listen as audio format.

Your Hacker News stuff is very high signal because, obviously, you're folding, right? - 'Cause we haven't done the curation like Hacker News did. - Exactly, that's why it's guaranteed to be good, whereas for Discord, it's a bunch of junk. - Yeah. (laughs) But I do think there's something similar, like Reddit also does the curation.

It's not us who's doing it, right? - Yes, yes. - If you're only the hot posts. - Still a little bit noisier, so I don't know if you know, I was a moderator of the React to Reddit for four years. - Nice. - So I've seen a bunch of stuff, and I know it's noisier than a Hacker News, but still pretty good.

- Yeah, yeah. So we do Hacker News, we do PJ Essays. I think PJ Essays are also super interesting. I listen to them all the time 'cause I, well, first of all, I actually think they're pretty well produced. We do a good job, like, I don't know, if we're quoting someone, we'll use a different voice.

- Yeah, yeah, yeah. - I think it's just well produced. And I also think, you know, the essays are so seminal to like, everyone in startups reads them. - Yeah, it's actually got me to read more PJ Essays than I would have otherwise, so mission accomplished. - I don't know if it was the last one, well, the time we published this.

- No, How to Do Great Work. - How to Do Great, that was a one-hour podcast. - Yeah, oh my God. - No one, I could, I did not read it, I just had to listen to it. - Yeah, yeah, yeah. - And I think actually, if I'm being honest, I think the motivation for PJ Essays was just like, I need this.

- Yeah, yeah, yeah. For that one, if it's like one hour, I would have actually appreciated like a segmentation, or like, hey, high level, you know, I know this is about to be an hour, but like, there are three main high level things, and then keep that in your mind, and then go like, part one, blah, blah, blah.

Part two, blah, blah, blah. - I think we, so we do that to some extent, and like, we produce like, chapters, I guess. - Yeah. - So you can just look at them. Probably could do a better job like, introducing it, but we do try to like, not play around with the PJ Essays.

- For sure, yeah, I mean, it's, you know how much work he puts into those things. - So we just kind of ship it as is. - Yeah, yeah, yeah. Actually, I'll tell you about one more that, one more daily, not daily, but frequent AI-generated podcast that I listen to, apart from you guys, which is PapersRed.ai, and I'll recommend it to anyone listening as well.

- Super interesting, actually. I've come across it. - You've come across that. It's by this guy, Rob, and I've tried to look him down. He doesn't want to be found. Anyway, but the selections are very good. I think you guys could do a better job than him. - Yeah.

- It's- - A job, Rob. - Well, as you can see, it converts PDFs to podcasts, right? And the problem with academic PDFs is a lot of references. You know, like, "Buy et al. 2022," and then like, headers, and then a table, and then you know, like, reads the table when you don't need to read the table, you know?

That kind of stuff. I think better engineering there from you guys would beat him, and I need that, so feature request. (laughs) - Working on it, working on it. - Okay, the final feature-related thing is the thing that we're announcing today as we release it. - Which we're super, super excited about, but Wondercraft now does video translation.

- Okay. - And dubbing. - Okay, why do people want that? - Again, let's go back to our mission. We're trying to expand access to content. I think, don't quote me on this again, like, I don't know who actually knows the internet, but like, 60% of the internet is in English.

You know, what if you don't speak English? You're automatically disbarred or kind of excluded from all the content that's produced. And thanks to all the advances that have just recently been made, we can actually make this, it's super easy to dub this in other languages. So we're super happy to announce this feature.

We're super excited, we've been working on it for a really long time. But now basically, everyone, go on our platform, upload your podcast episode, and see the dub for yourself. We'll use your voices, we'll completely convert it, make sure it's aligned with the, if you have video, we'll make sure it's aligned.

- Yeah. - And voila, just publish it. - And specifically for video, so like, obviously the thing that's going around is the hey, Jen thing, which changes your lips. - So we don't do lip sync at the moment. Could be another feature that we work on. - Yeah, because you're primarily podcasting, which is no video.

- Well, primarily no video, but I think we still basically, if you do have a video, we'll still align it to the chunks that-- - You still align it. Okay, so the hard problem is the aligning. - The alignment is the difficult bit, you're right. The actual, like-- - It's with all things in AI.

- Yeah, again, overloading the word alignment. - Um, we don't do the lip sync. - Yeah. - Yet. - It's kind of a gimmick, yeah. - It's like, it's not super necessary. If you're really just like listening to a podcast and you actively want to listen to it in a different language, then you're more interested in the content.

- So what are you aligning to? - Basically the chunks where the speakers are speaking. So you won't have an instance where you'll have me as the dub speaking while the camera is on you. - Ah. - Right, so it's basically just like, the speaker turns around the audio.

- So you will like, if it happens to be a little bit longer, you'll speed it up a little bit. - We do a little bit of, you know, trickery there. - Yeah. - But we get it aligned so that when you're speaking, it's you, and when I'm speaking, it's me.

- Okay, cool. And then, one more thing before we go into like, the building Wondercraft side, building a company and being a founder, all that stuff. You said, you want to ride this wave. You said this dubbing thing was very important to you. And I think this is a core stress of a lot of AI founders, which is, how do I keep up?

How do I pick what trend to ride and be like, very, very quick on? And what do I ignore? So just bring me through your thought process on like, why this is the one to really pressure and build quickly. - It's a really good question, 'cause I don't think it has a, I think you'll get 19 different answers.

I'll ramble a little bit. - Yeah, chain of thought. - Dubbing, you know, it's something that I want. There is content that just exists that I can't consume just because it's not there. I'm lucky enough to speak, you know, three, four languages, so I can listen to some subset of podcasts that exists all around the world.

Let's not even say podcasts, let's just say content. But the rest of it is kind of excluded for me. I have no context on what's going for it. And I think that problem is just multiplied a million times when you consider that if you don't speak English, you really are excluded from a lot of things.

So the ability, it's just motivating to be able to actually give access to content that historically would be completely obscured for most people. I think your question though has like a, and the dubbing market we think is huge. I think like, you know, it's easy enough to dub these days, especially with our product.

So like, it's doable. I think, you know, people are gonna, it basically was like a priority that we were tracking to push this out. We got a lot of requests for it, and so on. That's like the tactical answer on this dubbing thing. Obviously, you know, in general every week we sit down, we prioritize all the different features, and then we build.

I think your question probably has a slightly higher level. - Yes, the meta strategy of-- - The meta strategy is like, how do you just pick things? - Yes. - And do them, and what do you say yes to and what do you say no to? And frankly, it's a tough one, 'cause you don't know.

You don't know what to pick. Obviously, I can tell you things like, yeah, look at metrics, see. There's simple metrics. - There's no metrics for this. - There are simple metrics in the sense like, how many times are customers asking me for this? - Sure. - And then you can go one level deeper, which is like, how much money am I gonna make from the customers who are asking me for this?

So you can start to quantify from a revenue perspective. - In that case, was there something for that? - There's a lot of requests, and obviously that's part of the product. So the second you dub a podcast, we're making money. - Okay. - The thing that's slightly tougher is there is no better answer that I can give than try to do as much as you can.

I think we pride ourselves in pushing at least two super highly demanded features from our customers a week. - Two, wow. - Yeah, so we just move super fast. And part of that is 'cause we're frankly just a small team. We're pretty like-- - How big is your team?

- Four people. - Okay. - Total, the whole company is four people. - Yeah. - We just move very independently and very quickly. So we're just able to push out a lot of things. But I think that's frankly, that's just the requirement of what we're doing. It's not meant to be a chill game.

- Yeah. - And this is kind of a, don't wanna use that term, but for lack of a better word, like it's a race. Like you have to keep building or else someone else is gonna build. - It's almost, but it's like a race without an end. - It's a race without an end.

So I guess all I'm trying to say is like, yeah, it's tough, but you kind of just like have to be prolific. - Okay, fair enough. Yeah, you mentioned that you were listening to the "Line Chain" episode. And "Line Chain" is also similarly like, in my mind, very high tension, but also obviously important for them to stay relevant on top of all the trends.

- Yeah. - I just, it's very stressful. - But it's overwhelming, I agree. - Yeah, it's overwhelming. - Every day, my to-do list grows by like 10 times and I do less, like, you know what I mean? - Append only, yeah. - It's just, basically it's append only. So it's tough, but the reality is you just like, as a, I'm not even saying as an AI startup, as a startup, you just have to be prolific.

- Okay, yeah, of course. - Yeah. - Do you think that you have, so often people will say they will build, build, build until they find quote unquote PMF. Can we talk about what you think PMF is? And then, like, those core features, right? Like, because there's such a thing as also a feature factory.

I have worked at startups that were feature factory. Build, build, build, and then it's a whole bunch of junk that nobody, that doesn't have a coherent story. - Yeah, I think that's a really tough one, actually, to just like build, make sure that the things that you're building all actually like, from a very business perspective, either an upsell or they're somehow related to what you're building.

'Cause otherwise, it's really difficult to, I think, okay, let's just start from the baseline that you build a product for a customer. - Yeah. - It's okay to then build another product that caters to the same customer, or sell the same product to a different customer. - Yeah. - What's really difficult is to build a different product for a different customer, 'cause that's just a new company.

- Yeah, yeah, yeah. - Right, so-- - Have you seen the superhuman PMF score thing? - I have read that blog post ages ago, so I don't remember it. I think he said like, you know, 40% of people recommend it. - Yeah. - So, I don't know, 'cause-- - His advice is like spend 60% of your time making your existing customers more happy.

- Yeah. - And then 40% the adjacent guys, and then the remainder for R&D for future. - Yeah, yeah, I think that's pretty good advice, to be honest, 'cause I do think like, never neglect your current customers. - Of course. - Right, 'cause they're gonna be your biggest promoters.

- Totally, yeah. - So, we definitely wanna make them feel loved. We, again, try to really, like, I think one thing we also have going for us is our customer support. - Yeah. - Which is a one-person show at this point. But it's really good, it's really responsive, and we really try to make sure that the customer is like, knows what to do, and if it's something we, it's also just really rewarding to, I think, be like, at such a small team, and this is a bit of a tangent, but like, you get an email from a customer being like, "Hey, this doesn't work." And I think people are typically primed to like, be a little bit aggressive with companies.

Right, like, "Hey, this doesn't work, I need this." And then they're super surprised when within an hour, you're like, "Hey, we pushed a fix, this should work now, "I hope all's good." And they're like, "Oh my God, thank you." - Yeah, yeah, yeah. Yeah, that's the magic that's only possible at an early stage.

- Yeah. - Yeah, that's great. Okay, since you mentioned pricing, maybe just a few minutes or two on just how you think about pricing. You are probably a lot of usage-based stuff, and I always have this struggle with usage-based, because you're basically charging your heaviest users more, and discouraging them from using you more.

- I think of it slightly differently. We were not, so at the moment, okay. - Your tiers. - Let's just promise by saying, we haven't figured out what the pricing is. - Sure. - I think that's like the eternal struggle. - Yeah, what's in your mind? - We don't know how to price this thing.

- Yeah, what's in your mind? - So far. We don't actually, at the moment, charge you if you consume above your quota. So we have tiers. - Ooh, hot take. (laughs) - As in, if your quota, you just have to upgrade to a different one, right? - Oh, I see, I see.

- We don't have this additional consumption. - Meter billing, yeah. - Are we charging our biggest users more? - We actually just charge them per unit less. So it does work on, if you just produce more content than the unit, the unit, the per unit price is just lower.

- Yeah. - I don't know, I have to say, pricing is probably pretty rudimentary at this point. - Yeah, it's not as important. - It is important, 'cause fundamentally it drives revenue, but there needs to be a bit of a science and an art, it's a bit science, a bit art, to figure out what the right pricing strategy is.

I think you see like a million different things, and a million different things that could work for you, but you frankly just have to test it, and like, again, there's a million things to do, like four people to do them, so we haven't been moving as quickly as you want on that.

I think we're probably gonna release something new soon in terms of pricing. - Yeah, cool. - But it's a tricky one. - Hey, nobody has it figured out. - No. (laughs) - I've been through many, many pricing changes in my career. Cool, I wanted to talk a little bit about the origin story, 'cause you flagged that Mooncraft was actually a big part of-- - Moonshot.

- Moonshot was a big part of your arriving at this idea. - Yeah, let me give a little bit of background. So, Demi and I, we're both apparents here, we actually were flatmates right before we started this, and then we both had this idea that kind of came about from our respective upbringings, which was, there's a lot of people, and we're not the first to have this idea, by any stretch of imagination, but there's a lot of people out there in the world who are really good, really talented, who just don't have the financing to do the things that they're talented in, right?

Like, you need to develop yourself a little bit. So, we were like, cool, how do we solve this? How do we get people funding to do the things that they're good at? We're like, cool, let's align incentives, let's get people investing in other people. Cool, that was the original idea.

Applied to YC, we got in, long story short, we got in, and we're like, cool, let's see who's done this. We've obviously done that before, but let's really figure out what our unique edge is gonna be. So, we looked at all the other companies that had done that, and I think one particular one that comes to mind is Upstart, which is now a multi-billion dollar company listed, and we called them moonshots, that was kind of the idea, like people who have the moonshot potential, they called them upstarts, and we both had this idea that you were a backer.

Backers back moonshots or upstarts. They spent two years on this, and then they pivoted to, I think, a lending platform, some lending platform that they do now. So, we were like, we spoke to a couple folks there, and we're like, so what didn't work? And the thing they said is there is, I'm also giving a lot of detail right now on some moonshots, so maybe that's not super relevant, but just added context.

- It's fun to hear. - They're like, it's really hard to get your average Joe interested in backing your other average Joe, right? People like, it's not really interesting to back this other guy who wants to go get a better education somewhere. You don't, you know what I mean?

There's no driver, 'cause A, the economics typically aren't so good, and historically, they just found that people don't do this. So, we were like, cool, if that's the problem, then let's focus instead on a group of talent that has this exponential style trajectory. So, we picked athletes. People love sports, and people really feel what we call the emotional dividend of this investment.

So, we actually stopped calling it investments, 'cause the investment wasn't great. Financially speaking, your investment, you're better off putting it in the stock market or real estate or whatever, just the risk return did not make sense. - So, you might trigger some regulation. - We were fully triggering regulation.

We were compliant with the SEC. It was all regulated. We weren't doing this NFT style. And then, so we were selling this. So, you get a financial dividend. You still get part of their payback, or if tennis player X makes that much, you get 10%, but we also started calling it the emotional dividend.

So, actually, what we're selling a lot is also the experience. You get to be part of the front scene of this, of this person going on this journey. You're with them. We actually found a lot of people that kind of do this already. Now, what we found out after that is this wasn't a big enough market.

So, long story short, we tried really hard. I frankly think, there's a comment here about founder market fit. We were not the right founders for this, and when we were raising money, VCs would ask us, what are two software engineers from Panther doing? Like, why are you funding athletes?

What? And we'd sell this 'cause, you know, I played a little bit of tennis. They may actually play professional basketball, so we'd sell it like that. But, you know, for all the hate that VCs get, sometimes they ask good questions, right? And we were not the right people for this.

Our day-to-day was dealing with agents who are not nice people. - 'Cause you had to sign them up. - Well, you had to sign players up. - And you had to sell their earnings. - Yeah, so, A, just the cycle took so long. There was one guy, we were doing this as charity.

We were like, don't give us back anything. - You're just gonna give money away? - We're gonna give you money. We still didn't sign them after six months. - Oh my God. - Like, I don't understand at this point. Like, it was taking too long. We weren't the right people.

- It's your OPMF. (laughs) - And the most important thing, which actually leads us to WannaCraft, is this was not a tech product. And Moonshot was not a tech product. It was a legal product. It was a regulation. It was like, hey, now you can invest in people. The reality is, like, I think we just found out the hard way that we were building something people did not want.

And the thing that, like... So we realized, like, we're building something that isn't our strengths. Let's find something else. Looking back, I can also kind of, like, there were pointers when we were at Moonshot about, like, yeah, we should build something technical. I was listening to the "Lang Chain" episode that you guys did, and...

Harrison, right? - Yes. - I remember you saying he was a big sports fan. - Yes. - And, you know, he built this, like, basketball scraper or whatever. Literally, my biggest, like, the thing I was most excited about with Moonshot was, like, okay, let's build out this infrastructure so that people can invest and blah, blah, blah, blah, blah.

And then we get to the point where we can do, like, data analytics on them and start predicting. I'm like, this is the most exciting part of this. And I was like, this is definitely not the core part of the business. So I'm excited about something that's gonna take either years to get to or something else.

But the core, you know, we designed our landing page and built it. That was the funnest thing I've done in Moonshot. So we're like, when we decided to pivot, we were like, we need to do something technical. The story from there just became, like, okay, cool. Let's list some ideas.

What are we gonna track? Which one do we have the most conviction in? We rank them. And Monocraft was the one we had the most conviction in. It was this idea, again, expanding access and just translating your ability to produce content. So producing content in one format and then taking that to all the other formats.

So we built that. We built the podcast builder. Super quick prototype because I think at this point, to anyone pivoting or hard pivoting or considering it, the name of the game isn't, like, to get attached to your idea. Just, like, actually, you should be trying to invalidate this idea as quickly as possible.

So get it out there and let people tell you it's a piece of shit. - Okay, so what did you do to get it out there? - So we built it out. We built out a little, like, UI. Literally no authentication. It was a form where what you guys see now on our platform, which is, like, the content script page, blah, blah, blah, blah, blah.

It was one page. Like, you click and it was the most janky React stuff that we had. Zero authentication. So in theory, if people found it, they could just produce as much audio as they wanted. - I'm quickly developing an opinion, by the way, that I think all early-stage startups should not have a marketing site.

Your .com should just be your app. - That's an interesting one. I mean, we, I don't think we have one too much. I think you still need, like, discoverability. - Okay. - Whether that's your marketing site or a lot of artifacts. I think we are moving to-- - Or, like, embed your app into the marketing site or whatever.

Just, like, have, get people using it as much as possible, right? - Yeah, I think that you're, like, the most convincing thing is, like, so we have a feature on our app, which is, like, test with example. As soon as you log in, it says, "Test with example." And the whole point of that was, like, to, between login and audio generated, how few clicks does this take?

- Yeah, yeah. And yours was? - Like, two or three clicks. It's, like, test, create podcast. - You, like, pre-filled everything. - Generate script. - Yeah. - Generate audio. - Yeah, yeah, yeah. - Right? So, and arguably, that was still too many clicks. We should probably put something on the actual landing page, so that, like, you can see-- - Yeah, now you have just the player there, right?

You can just kind of listen to the Hacker News Daily thing. But what signals did you get from doing that? - The signal that we got is that someone picked it up on Twitter, and just, like, you know, all these, like, AI-- - Influencer voice. - So, someone picked it up, posted it.

We started getting a ton of inbound. So, we were like, "Holy shit, let's just, like, "paywall this." So, we just, like, again, the jankiest Stripe integration, which was basically, like, we have an app with a Stripe integration. This was just to ship it within the hour. We have an app with a Stripe integration that, once you click, then takes you to a different app, hosted somewhere else.

So, that one was still unauthenticated. - It was hilarious. - Yeah, security by obscurity. - It was hilarious, right? But we basically just made that, like, 3K in one day. - One day, yeah. - We charged around, like, 50 bucks. We didn't even, literally didn't think about it. We just charged 50 bucks, and people paid.

And we're like, "Okay, well, there's something there." - Yeah. - So, then-- - 50 bucks for one-- - For a month. We were just charging 50 bucks a month, nothing. Like, just, like, will someone pay for this? - And you were just, like, on one, like, VPS somewhere. - Yeah, will someone pay for this?

We were on, like, on one EC2 instance. - EC2 instance, yeah. - You know what I mean? Like, it was janky. We were like, "Just someone needs to pay for this "before we move further." Someone did. People did. So, then we were like, "Okay, cool. "This is interesting." - Interesting.

I'm gonna move up the question that we said was gonna be the meatiest question of this interview. So, you chose this out of your list of ideas. And this is one of the things that a lot of AI founders worry about, right? So, the framing of this is, are you worried that you're a thin wrapper around 11 labs?

What is your mode? And how do you think about that when you're evaluating your idea? - I think you can, that's a seminal question. I think anyone in an AI startup is, I think, frankly, everyone in an AI startup should convince themselves of this. Don't listen to me. And, like, just make sure from first principles that you can derive this.

But I guess I would start by first saying, what is a mode? Well, like, or what is defensible? In theory, if we're just taking it, and this is trivially true, but the fact that someone built it means someone else can build it, right? So, modes tend to just be around, built around, like, you have a lot of network effects, or you have a really good product for this use case, or, you know, something like that.

And I think, typically, when people ask this question in the AI context, they're thinking of, like, okay, you're a thin wrapper, you're an application layer thing, as opposed to you're one of the, like, underlying technologies or APIs that people use. Cool, I think that's fair. But I think the reality is that, like, yeah, these APIs exist, and they probably do serve a million different use cases, but they're not built to serve these million different use cases.

So, whenever you ask the question of modes, it always has to be with the perspective of who is the user I'm building this for, right? I can use chat GPT to do half of my writing, but, you know, I don't know. Jasper claims that they do this much better for marketing, so it's tailored.

Actually, don't quote me on how well they're doing after chat GPT came out, 'cause they were really big before. - Yeah, there's some negative data points, but I'm sure they-- - I know, but the point is, like, you're making this easier. We make creating a podcast easier, and there is tooling there.

We help you, we can post it directly through us. We have the tooling around, you know, setting the intros and the outros. We have the music, we have an editor. All these things are also getting just much more and more developed. We're building templates so that you can do different style of podcast.

So the idea is, if you're trying to start a podcast, yeah, don't go to a generic text-to-speech engine. Come to us. And the reality is that we then can, in a very opinionated way, actually select which text-to-speech engine we want, right? So we actually have just, like, in my mind, it's the application layer, fundamentally, that, you know, people use, and then all these API layers are what developers use to build products on top of them, right?

- Right. - I appreciate that it is, like, a seminal and really hard thing to wrap your head around, especially if you're, like, about to invest in a company. It's like, will they actually just be defensible and be able to grow? - Yeah. - And yes, there's no doubt that companies can do this.

The question is just, like, are you building the right product for the right use case? I think, particularly, if you're, like, always framing your company as an AI company, then you're putting the carriage before the horse in the sense that you focused on the implementation rather than the use case.

Focus on the use case. And then build a product for it. - Yeah. - Right? 'Cause fundamentally, you know, any of the SaaS's that exist, think, like, more traditional SaaS. What's their mode? The technology, everyone has access to it, so they just pick the thing that does it better than the other.

Now, that mode question is super interesting because I think you should actually - Flip it around. - Flip it around, which is, what is your mode as an API? - Yeah. - Right? So, Chad GPT, like, yeah, fine, they had a first-mover advantage, and I think, you know, by no means, this is my opinion, but by no means was Google, like, caught off-guard with this, right?

It just, Google has some, half the technologies that Google invented are actually what's used to power all these transformers. But, you know, it went against Google's strategy, maybe, to, like, be the first-mover in this 'cause they cannibalized their own market. Whatever it was, I'm not sure. But, yeah, open AI's mode is that they paid for the training bill.

- Yeah. - So, they just have a good model. Cool, people now know that that's valuable. - And, they hired, like, very top-notch. - Superstar, sorry, obviously. Obviously, not taking that for granted. But, like, they, you know, assuming everyone can do the hiring and that these people exist, they paid the bill, and they were the first to launch this.

But now people know it's a thing, so people are gonna launch similar APIs. So, what is your mode as an API? So, it's just an, it's an existential question. It's like, how do you do, how do we defend any of this? - Yeah. - And you do this, frankly, by being probably better just as a product.

Again, the product is always with the perspective of who's your customer that you're selling it to. And the other thing is, frankly, that, like, let's not forget, the market is huge. There is space for everyone, right? If you manage to, like, if there's four good products out there in any specific thing, the market is huge, and they're all gonna be able to, you know, make a living out of it.

- Sometimes people justify markets. By the way, that was a really good answer. Thanks for taking that hit on. - I've had to answer that question way too many times. - Yeah, I know, it's not the first time. But, I think it's actually, you know, having been an investor, it is more important for you to answer that question authentically for yourself, 'cause you're the one spending your time on this.

We're just giving you money, it's not that big of a deal. Well, so, ah, I forgot what I was gonna say there. Yeah, I think 11 Labs, you know, I think the thing that you were mentioning is, you know, there is not that much loyalty. Obviously, you and Matti are friends.

But there's not much loyalty if your main goal is to serve API, because you just need to wait until the next best one comes along. - The same thing as Uber, like, why is Uber defensible? - Yeah, yeah, yeah. My favorite quote, actually, I went to an early, like, pre-HRGBT forum with Sam Altman, and I had this video advice from Sam that said, like, Facebook had no moat, and they just built and got the network.

- But frankly, also, Facebook was building something back then, which is kind of ludicrous, like, cool, you're building a social media app, okay, how big can it be, right? - Like, how big can the internet be? You know what I mean? All of a sudden, it's this behemoth. So it's like, yeah, the fact that it was built, again, trivially true, but the fact that it was built means it can be built by anyone else.

So you, there is no such thing as, like, a absolute true moat. The question is how well, how quickly, how much earlier than everyone else did you get there, and a million other things as well. - Yeah, cool. Two more sort of product-building questions, and then I want to turn the mic over to you to ask me anything as a podcaster.

So one is script generation, which, you know, sort of the two core activities of Wondercraft. I think you guys do a really good job of summarizing the Hacker News comments and the content. In my mind, the hardest thing is just managing context, because, obviously, it is unbounded for the context window.

Do you use LangChain? Do you build your own LangChain? - I like LangChain, I've never used it. I've never, ever used it. - So you just build your own recursive summarization thing. - Yeah, so-- - Do you use a vector database? Which one? (both laughing) - I'm not gonna say any further, unless you realize how unsophisticated we are.

- It's okay, no, I think it's okay. People, builders and investors want to know what's actually useful-- - I want to say the thing, the Hacker News recap, I think, I don't know, but we can look on the date of the first episode when we launched it. I think it was like the 8th of April or something like that.

We decided on Wondercraft probably on the 5th of April. So this is how quickly it took us-- - The name of Wondercraft? - As in the company, this is what we should pursue. - I see, so you took three days. - And this is before people started paying. We were just like, Hacker News started off as a proof of concept.

It's like, do people even want to listen to AI-generated content? And it's been going on since. So it wasn't actually that hard to build the Hacker News recap. Obviously, now it became a little bit more sophisticated, so we summarized comments and blah, blah, blah, blah, blah. - Yeah, that's great, by the way.

I love that. - But actually, it's like, especially with our API now, it's super easy to do something like that. We, Hacker News, frankly, just doesn't need a vector database unless we're trying to do something. At the moment, we just summarize the posts. And actually, for anyone listening who listens to the Hacker News recap, just tell us some feedback on how to improve it.

At the moment, it's just like, it's one long narration. But one thing we're thinking of doing, for example, is sometimes you have multiple posts talking about the same topic. So actually, an idea is do top 10 topics instead of top 10 posts. That starts to need, I don't think that needs a database, a vector database.

I think it needs something, you know, it needs to do some embedding, manipulation. - You need MP.A? - Yeah, exactly. But something like that. I think for when you start to really, really scale up the content that you're pulling from, then you start needing a vector database. - Okay, so what are the actual hard problems for script generation?

Prompt engineering? - Prompt engineering is one of them, especially when you're trying to tune things, like tone, or how long it should be, and all of that. I actually think, so the way our app works, and maybe that's leading to your next question, but the way our app works is twofold.

You have the script generation, and then you have the audio generation. In our minds, we imagined it as a really linear flow. That, of course, I want to generate the script, and then I want to generate the audio for it. Where in reality, I don't think it's that linear, to be honest.

I think it's very circular in how you do this, right? So you generate the script, you listen to it, it doesn't sound quite right, and you want to modify the tone a little bit, and so on. So I think our UI doesn't get that wrong, unless you're exactly what you're going for, is this recap, summarization style.

So I think what we need to do a much better job is this templatization, and this on-the-go, on-the-run, kind of like, okay, let's play around with this text via some prompt. So I think it's definitely an improvement we're looking to make, but yeah, it's a prompt engineering question, fundamentally.

It's like how, prompt engineering on one level, the next level's like how good, how fine-tuned is your model to specific tasks, and then maybe you can use your own model if you wanted to. - Would you look at that, or is that not-- - We'll definitely look at that, but you look at that when it's a problem.

It's not a problem right now. - I would say, as a listener, daily listener, I think your show notes are really good. Most daily podcasters are horrible at show notes. They don't tell you anything. Here, if I want to click in, I can click in there, so definitely keep that.

But the top thing that annoys me about you, and I'm just complaining now, is you don't write the post title. You don't spell out the-- - Really, we don't write the post title? - Yeah. - I think we used to. - Well, it's gone. - That's horrible. I'm sorry.

- So the voice, by the way, I wonder if there's any story behind picking the voice. The voice would just say, "This post does," and it doesn't tell me the title of the post. - Oh, in the audio, in the audio, we don't say that. Ah, sorry, sorry, got it, got it, got it.

- And when I'm running-- - Did you email us about this? - No, no, I've kept it to myself. - Someone did email us, so it's like, "Please say the post." - I'm like, "This is so obvious." I would actually just fork "WonderCraft" just to add that one feature, because it's so annoying when I don't know what I'm listening to.

- We'll fix it, we'll fix it, easy peasy. - Okay, cool. The audio generation, you use 11Labs. What makes a good podcast voice? You have a bunch of options that I clicked, and in my mind, I like a deep voice. I like the Morgan Freemans. You don't have that many deep voices.

Do we want, is there such a thing as a high-energy voice? You also insert breaths. 11Labs has also advertised that they have an AI that can laugh, which I think is fun, important. Basically, what makes a good audio? Yeah, generate audio. - Yeah, it's, it depends, again, on the perspective.

Everything is kind of answered with the frame of reference that you're looking at. If you like a deep voice, A, that's kind of a personal preference, and B, it just kind of depends on the thing. So if you, I don't know, let's do, say you're doing something like meditative or kind of affirmations or something that encourages people every day.

You probably do want a slow, deep voice, something relaxing. If you're doing the Hacker News recap-- - Cheerful. - We picked Anna, who's our default voice. - Yes, Anna, yeah. - Because-- - I have an attachment to Anna. - Yeah, we all do. (laughs) She's just news anchory style, very professional, very formal, very neutral.

- Very pleasant, yeah. - Very neutral. So it depends, really, on what makes a good voice. It depends on what you're doing. There's a few things, but if you're doing an interview, I think it also just, frankly, then you get into the question of what makes a good podcast.

Well, the good podcast is like, I think it's also kind of a personal question, which I haven't, or probably there's a general trend that I'm yet to decipher. But yeah, you probably do want a little bit of humanity in there. You want a stutter. - Yeah. - You want some pauses, right?

I'm speaking, I don't speak in complete utterances. I have an utterance, and then I pause a little, and then I speak again, and so on. Laughs and something to make it human. - Yes. - It's kind of overlaying of the two, if you have two speakers, it's like exchange, right?

I will be speaking. If you look at the transcript of this episode, we probably overlap in when we're speaking. - And that's fun. - And that's actually interesting, right? 'Cause it is a conversation. - Shows the sign of excitement, especially in our studio when we're three people, and we're all talking at once, you know it's good.

- Yeah, yeah. I don't like this zoomification style where if you're gonna big zoop and big zoom, only two people can speak the second more than two people try to speak. - Yeah, yeah. - It's a disaster. So I think it frankly just depends on what you're doing. We are, yeah, at the moment we're really good at doing this narration stuff, but I think we are building a lot of functionality and tooling to just make this multi-host thing a more of a reality.

- Okay, okay. I would say objectively, if it was a friend and company, not that important. So this comes down to how human should your users try to be? Because I'm fine with Hacker News Daily making mistakes, because I know it's AI generated, right? I would be less fine if you were not up front, and Alec here actually thought it was a human when he first listened to it, 'cause Anna's just really good.

But then you'll make mistakes, like pronunciation mistakes. So I actually have a clip that I wanted to play you. Anna takes this, on September 8th, Anna was taking a lot of breaths. She was very out of breath. I was very worried about her. She was hyperventilating. I was like, "Inner UK?" Anyway, so basically, I think if you disclose up front that you're an AI podcast, then people will be like, "Oh, okay, I tolerate that mistake "and I use you for information, "and not for believing that there's some human "on the other side that I might meet someday." But if you're investing so much effort into being real, then your end goal is you have to lie to your users.

- I don't think the investing and being real is for the purpose of deception as much as it is for the purpose of making it slightly pleasant to read about. I think on Hacker News, we do say on our Spotify page that this is an AI-generated podcast. - For now, but-- - As in, yeah, yeah, so there's two things.

I think if you wanna be smart about this, you should say that this is AI-generated content. The second people find out that it's not you, the backlash is gonna be big, 'cause it will be interpreted as deception. So you should do this just to be smart. I don't think there's a point in lying, especially if the content that you're putting out there is just like, this is informational for you, so consume it, this was efficient, this helped us put it out there.

The second thing is, frankly, I don't think it's up to you whether you tell them or not. Very, very soon-- - They'll know. - Google is just gonna mark things as AI-generated. So I think there's a new thing. I saw a quick YouTube video about it, so I don't know what the exact terms and conditions are, but YouTube has, I think, released a new monetization rule, and it does mention something about AI-generated content.

So there is, it's not up to you anymore. People are gonna know that this is AI-generated, so I think it's just in your interest to say that you're AI-generated. Ain't no shame. - Yeah, no shame at all. - 'Cause fundamentally what we do relies on the premise that you have done some content.

We don't generate our own content. We don't synthesize our information. It assumes that you've written a blog post, done an actual podcast, or have some artifact on which you wanna base what you're feeding through Wondercraft. - And you said in some of your material that I've seen before that you are interested in watermarking all your stuff.

You haven't done it yet, but whenever there's a standard for doing that, you will do it. - Yeah. - Okay. - I think the thing that this is blocking on is the standard. I'm not super up-to-date on what the work on this is. - I think OpenAI will probably-- - But there just needs to be a standard so everyone can interpret it.

- Yeah, yeah, yeah. Yeah, cool, awesome. Great, I wanted to dive in a little bit on tech options, and then zoom out to just you asking me questions. So TTS Options, we talked a little bit about 11Labs. I would also say as a podcaster, the leading competition to you guys, I know it's not exact competition, but it's Descript, because they have Overdub.

- Yeah, I think Descript is really good, and they're definitely a solid company. I've used their video editor before. It's great. The Overdub thing is super useful. I think it's really creative to have edit videos by editing the transcript. - Yes. - Super, super creative, super user-friendly. - Would you build that?

- I think, again, it's like, we're not building for the sake of building. We're building more for the purpose of the user. - Yeah. - Whatever users find more interesting. I think what we're doing is we, the use cases are slightly different, right? And I think the people that they're targeting are slightly different.

- Yeah. - We do wanna have a lot of automation on the script side to also just help out with the way you formulate your content or the way you pull your content, much more so than just the editing process. - The ingest, yeah. - Yeah. - Okay, got it.

And I just wanna map out, here's how I think about TTS, text-to-speech. There's the big cloud options, Amazon Polly, Google Text-to-Speech, and Microsoft Cognitive Services. - Mm-hmm. - As someone who is ex-Amazon, I'm very embarrassed by Polly. It sucks. - Yeah, Google's isn't great, and Microsoft-- - I'm sure you investigated all these things, and you're like, "Okay, this is not serious." There's Play.ht, which is probably the other big YC-- - Yeah.

- Alum. Just a quick two seconds on your thoughts on Play.ht. - Sounds good. I think it doesn't sound super, as good as Alum Labs, my opinion. - Yeah. - But I think it sounds good. - But I have heard other founders tell me this as well. - Yeah, they're-- - And I don't know why.

- I think they have a, what's it called? A more comprehensive platform. - Okay. - As in they let you do this pronunciation. They just have a lot of tooling around it. - So different features. - I think quality, in terms of the voice, 11 Labs is still better.

- Yeah. - I think they are releasing a new model. I don't know if they've released it already or not. - Yes, they did, yeah. - Could be better. I don't know. But they do have some functionality out there. - Yeah. They also released the viral Joe Rogan, Steve Jobs interview from last year.

- Yeah. - And on your landing page, you were like, "This is something that WonderGraph will never do. "AI content speaking to each other." - Yeah, who wants to listen to that? Like, it's fun-- - Apparently a lot of people. - It's fun 'cause it's a cool gimmick. I think it's like nice viral material.

I would never listen to like a synthetic Joe Rogan. - Yeah. - This brings us on to a little bit about the whole like content question or the proliferation of AI, which is like, okay, if it's this easy for me to create content that's like, you know, somewhat engaging, like all these AI songs, the Drake song.

Well, okay, so if it's this easy and it's just like, if it's this easy to generate content, well, why will I listen to it? Like, I think we already suffer from the problem that there's an oversaturation of content. So I think, you know, I'll give a Netflix example. When I go on Netflix, there's too much content.

I don't like discovering new things. I either know exactly what I'm about to go and watch or I will re-watch one of the seven things that I constantly keep re-watching, right? So I'm not like, their discoverability isn't the best and every now and then I maybe like look at some of the other things, but I'm not trying to look at Netflix.

- It's quite funny 'cause they used to invest a lot in their recommendation systems. And like, nobody I know uses that. - But now it's just too big. It's just too big. And they're like, it's overwhelming. So many tiles that you're looking at on the screen. So, cool. Now, what if I told you that there's probably is gonna 1,000X and half of it's gonna be AI generated?

So what I just said is like, we all have this internal proxy of what a thing I'm likely to watch is, which is it was recommended by a friend or it's one of the things I previously watched. I'm not doing this discoverability. So imagine that's the case. I think another layer of proxy will just be like, from the Netflix category of the shows and movies, was this human made or is this AI made?

If it's AI made, I'm not watching it. I want real human stuff. But I think there's gonna be a fine equilibrium that reaches, which is like, it's not even worth generating that much content with AI. I'm talking from entertainment purposes. I'm not talking so much for the social media propaganda, blah, blah, blah, not getting into it.

But at least from an entertainment perspective, if I can just always have an on-demand Drake song that I wanna listen to, well, at some point I'm just gonna listen to Drake. 'Cause like, you know what I mean? - This is the greatest hits. - Yeah, it's like, I'm just gonna listen to a curated set.

The fundamental thing is curation. It's just like a parallel example of the Discord thing. It's like full of noise. So I need to like figure out what is the right thing to consume. - Yeah, awesome, awesome. Great answer. I just wanna bring through, here's my map of the market, right?

There's Speechify.com, which focuses on celebrity voices. I noticed that you don't have celebrity voices, probably 'cause of licensing issues, right? - Yeah, and it's also-- - Like, they don't usually use their voice. - It could at some point, but like not a priority at the moment. - Yeah, I really want a Morgan Freeman one.

(laughs) - That's gonna cost. - I know, I know. Mycroft.ai, privacy focus, run offline. Probably not-- - Haven't heard of them, actually. - Not important for you. There is some interest in virtual characters for games. So Conv.ai is the one that I had listed here. Did you look at the gaming market?

- Not deeply, to be honest. - Yeah. - But it could be an interesting one. - Yeah, people are exploring that. There's obviously Heijian now, and that is it for as far as I can scope out the landscape. And then there's the open source systems. So Tortoise TTS, as far as I can tell, is kind of market leader in open source.

- Yeah, I would actually say probably a lot of these models were built on top of it. - What models? - Maybe. Some of the-- - These other companies. - The companies, yeah. - Yeah, there's PyTSX, Koki, used to be Mozilla, and then Larynx. - Yeah. - Anyway, all these things.

And then there's also sort of research-grade stuff coming out of the major big tech companies. You talked about Google Sandstorm. - Probably the one I'm most excited about. - Okay, why don't you-- - Well, it's really good. You can check out the paper. - Yeah, we'll play a clip.

- Yeah, they haven't, I think all you need is like three seconds. - Yeah. - And it'll just, it's like a three-second sample, and it'll play the audio in your tone. It sounds really human as well. Like it has utterances, it laughs. It's pretty accurate to like, it sounds human.

- Yeah. - So very interested in that. They haven't open sourced it, and I assume for good reason. - Yeah. Google never launches anything. You have to wait for somebody to, or you guys could re-implement it yourself. - Yeah, yeah. GPUs after PMF. - Ah, that's a nice quote.

How strongly do you believe that? - GPUs after PMF? - Yeah. - Well, I believe, I think this was another question, yeah. Which is like-- - What is PMF? - No, what is your favorite like PG advice in building a company, right? I think that my favorite thing is just like, don't spend your money, obviously.

- Ah, okay. - Everyone and their mother is trying to get a GPU at the moment, so I don't think it's, we're definitely substantially reducing our runway by doing that. - Yeah. - Obviously you do that when you believe the investment is worth it. - Yeah. - And again, you have to pick the time at which you do that.

- I mean, there's other companies, I think this is somewhat consensus. I think the non-consensus thing is to spend a shit ton. So like inflection raising a few hundred million dollars, and then spending 95% of it on GPUs. - Same with Mistral. I think it depends on the company you're launching.

I think if you're like, you know, maybe you're a brand new TTS company, maybe it is worth just doing that. - Yeah. - I don't know. - Okay. There's also AudioLM, also out of Google, VAL-E from Microsoft, and MetaVoiceBox. Are you just watching any of these? - Watching any of these, obviously paying close attention.

- You try them all out, yeah? - Try them all out. - What are you looking for? What is like the holy grail? What are you looking for? - There's also how human it sounds, and like how likely I'd be to listen to this if I did it. Also how like customizable it is.

- Yeah. - I think the problem with all these voice things, and generally a lot of the AS stuff is somewhat random, but you're using it in production applications that require certainty, right? Just as an example, if I promise my users this podcast or this segment will be 30 seconds, it needs to be 30 seconds.

Or, you know, given some SLA. - Discontolerance. - Some SLA around like, you know, it's 95% of that. - Yeah. - But I think a lot of these things just tend to be a little random at the moment. So like how, can I literally specify a tone that I'd like this to be and be certain that it's doing it and it's not some weird like attempt at sounding surprised.

- Yeah. - It's just like, yeah. Basically how controllable and how realistic they sound. - Yeah. And then final question around just the landscape of TTS. What are the unique challenges for non-English TTS? And I'll tell you, right? So I'm interested in having 28 languages of latent space. Right, that's only good things for me.

Except if it sucks. - Mm-hmm. - And I obviously, I have no way to validate. - I think that's the problem with-- - Latent space Ukraine. - Yeah, and I think that's the problem with dubbing, I think. So the reason, one thing we're gradually building out, but we already have as part of our dubbing product is that we have QA as part of that.

So we actually work with professional translators to just make sure that the things that we publish-- - Oh, nice. Oh, you should put that up front. - Yeah. So that's really one of the, like, fundamentally the problem with dubbing, if you ask anyone who's ever tried to dub, is you don't know what good sounds like in these other languages.

You're like, I can tell you I dub, but I'm gonna tell you that-- - Yeah. - I think there's a lot of big podcast studios who have tried this before. There's one I can think of that's tried this maybe five times with five different companies in the last five years.

Their fundamental problem is that you just cannot, yeah, fine, Spanish sounds good to me as a person who doesn't speak Spanish, but like, it doesn't sound good to a Spanish person or an Argentine person who have totally different accents. Right? So I think the tricky thing as well is getting that straight is, in English, it's very clear, and there's a huge proliferation of content in every single accent, if you will, Australian, English, British, American, Irish, whatever.

In other languages, especially if you consider the niche ones I don't know, I'll pick like German. There's like a few German accents. I don't know how much content Austria puts out 'cause they have a specific accent, right? Germany itself, probably every-- - Swiss-German. - Yeah, just record Roger Federer speaking, right?

So there's just a lot of nuance that isn't as widely available. The English market is just the biggest. - Yeah, yeah. - And I say this as someone like, so I say someone who speaks German, but like, think of all the other language, I have no idea what Spanish, Spanish-Spain, sorry, Spain-Spanish sounds like and Argentina-Spanish sounds like.

There's a difference. I know that. I just don't know what it sounds like. - Cool. Well, if you ever need Chinese validation, I know I have some very fanatical Chinese listeners who translate every podcast. - Oh, that's amazing. - So we can use that as QA. - Yeah, I would definitely love that.

- Shout out to the Chinese army. Great, great, awesome. What do you want to ask me as a podcaster? - So tell me about your setup process for setting up a podcast. - Oh my God. - And I think it goes far back as like, cool, you know what guests you're having on, you're just like, no, no.

- Oh, you mean starting-- - Even further back. Even further back. It's like, I want to talk about someone in this space. - So like for a new episode or for a completely new podcast? - New episode. - A new episode is pretty easy, right? Like the way I set it up with you was like, hey, I'm a fan.

Do you want to chat? And you said yes. And we scheduled-- - And how long did it take to get the scheduling right? - Scheduling, you guys are very responsive, very easy to work with. Some others are very hard. Like we took a few months for Harrison to come on.

- Yeah, and you said that. - For you guys, it only took two weeks. - Okay. - And then, yeah, we got this studio. We came here and then, you know, all three of us spent 50 minutes setting up the audio for ourselves. In San Francisco, we pay someone to do it.

And so we just kind of walk in and start recording. - Yeah, and how do you come up with the questions that you want to ask? - I do prep and research on you and the business. And usually it has a format of find information about the speaker and then get to know you on a personal basis, put you at ease, so that you'll be more open for the other stuff.

'Cause you know I care about you as a human being, not just as an interview subject. - Yeah. - Then talk about the business. I do emphasize the what is question. They're very straightforward. Like what the hell it is you're doing, because people want a clear answer and some people cannot give a clear answer.

And that's also information. Then I always am interested about company building just 'cause I'm a founder myself, you know, and I want to learn about what it's like on your side and ask the business side of things as well. Some people don't have that side 'cause they're academics. So then you go more into their research and try to keep up with any academic in theory and math that they might bring up.

And then maybe try to, at the end, and this is the hardest bit, and this is the stuff that I have the most second guessing about, try to have a segment at the end where you ask a consistent set of questions for essentially norming across guests. So I have a system prompt for our guests and at the end of it we basically end every episode with a lightning round, right?

Acceleration, exploration, take away. I don't love the questions that we ask. We're still iterating on that. - You gotta experiment to do it. - Yeah, I think just people like Lean Space mostly because the kind of questions that I ask are the stuff that is on everybody's minds because I just talk to a lot of people.

And so I just force you to answer them because you're right in front of me. - It's great. - It's great. And so you have, this is a whole interesting conversation 'cause you're a human podcaster, AI podcaster. - Yes. - So as a human podcaster and someone who was with a really popular show, and also someone who can actually implement this stuff himself, what is some of the AI tooling recently that you've baked into your processes?

- I only use the script. And we built for editing. So for example, and by the way, this goes into a theory of content, which as a content creator myself, professionally and as an advisor, I have, which is that we developed a few show formats. Lean Space is a channel, it's kind of like a TV channel, and channels need different formats.

So you have the reality TV show, you have the news show, you have the cooking show, whatever. For us, we have the founder interview, straightforward. Everyone has them. We have the breaking news, Twitter space. And that is, we want to be the day one first podcast to come up with the most in-depth breakdown of something that is that everybody needs to know.

And that has high value to people, right? Because if you're a week delayed, one month delayed, then no one cares anymore. And then finally, we have the fundamentals, like the one-on-one evergreen episodes that are less time-bound. So this one is relatively time-bound because it's a snapshot of who you are right now.

But we want to have evergreen episodes that people can go back two, three years in the backlog and still get value from. - Yeah, and these more fundamental ones. - Yes. So we have three show formats right now. And I would say we have different tooling for each, right?

So the one that I don't need any tooling for, essentially, is the fundamentals one, because we plan basically every minute of that show. It is a lot of work. But it's high quality because people love it. It's got the longest tail by design, right? The Twitter spaces require Descript because a lot of silences and a lot of ums.

And that's not good podcast audio. So you got to cut it out. - So you literally just go in, like you edit out the Twitter space that you did. You record it and then you edit it out. - Usually it's like two hours, we cut it down to one.

- Okay. - And it's a lot of pain and a lot of work. But it's the only way that I get some pretty high-profile people onto my podcast without booking them. They just show up. And that has value to me, right? Simon Willison has been on my podcast three times and I never had to schedule him.

And people love him. I mean, he's great. And then this one, basically we do, I don't need Descript, obviously, but we do use Small Podcaster, which is a 100-line Python script that throws the transcript into Anthropic and then generates show notes. - Nice. - So that's about it right now.

- Nice, so it's interesting 'cause I think you're in a very nice position where you're able to do what a lot of these services charge you for, you can just do it yourself. - Yeah. - So it's an interesting one. - But obviously I'm interested in paying for things 'cause my time is valuable and if it does a good job, then I'll use it.

For Wondercraft, the thing that I really wanted was the RSS to podcasting, right? Which you now have. So I'll try it out, but chances are I will not be happy with something. And so then the question is, how much customizability do you give me to do that? And we'll see.

- Yeah, interesting, cool. These are all my questions. - I will say, well, you missed out one thing, which is marketing the podcast, which is a huge part. That is mostly my job. - So how do you market your podcast? - Twitter. - So you think Twitter's a good media for that?

- Twitter and Happy News. - And threads, or you post clips, or what do you do? - I have tried posting clips. It's just too much work. So if you guys do a good job of clips, I will use your stuff, but it's just too much work. So mostly I just put a big post saying, so for our George Hots episode, we were like, "Latent Space is excited to present "George Hots on Tiny Corp and Commoditizing Petaflops." Something like that.

And just sometimes the fame of the guest will just lead the episode. So the one I dropped yesterday was Chris Ladner. And people were like, "Chris Ladner's the boss. "I don't care about anything else. "Just I want to hear as many Chris Ladner tokens "as possible." Others who are less famous, I have to introduce who you are and why I care about you, why they should care about you, 'cause most people will not have heard about you as well if you've done.

So then I need to make the case a little bit more. But that's fine, that's my job. I just think it takes a lot of work, and that's the part that will be hardest for me to hand over to AI, 'cause I have a very specific voice for myself.

And apparently all AIs think that Twitter, to tweet you have to have emojis and hashtags, which is so dumb. It's so obviously dumb. The training data is very bad. - I think it depends, right? Sometimes you do, if you're trying to be more on the professional side, you don't.

- Yeah. - It just depends on the tone and brand voice, I guess. - It's just cringe. And that's the problem with AI. AI is mid by design, right? And you do not want to be mid as a thought leader, as a content creator. People don't want to subscribe to mid things.

They want to subscribe to a level slightly ahead of where they are. And so that's for better or worse where we target. - Which I guess the interesting thing for us is how do we really enable that you have your process and really bake that in to make sure that what you're presenting and what you're coming out with is unique.

- Yeah. I'll tell you one thing, though. I'm happy to hand over my LinkedIn to you. I have a LinkedIn following, I don't use it. So completely money on the table there. So happy to give that to you. You don't do social media posting, right? So they have to go get like spot social or buffer, both of which are horrible.

So if you have an adjacent product some day, you might want to get into, basically money on the table marketing is what I think about it. Basically, we would not do this by ourselves. It is still inferior to a human crafted thing, but it doesn't matter because I just, I'd have zero LinkedIn presence.

- Makes sense. - Yeah. - Great answers. (laughs) - Obviously happy to offer any thoughts as you build out for podcasters. Okay, well, shall we go into the lightning round? - Let's do it. - Yeah, the one I just said that I wasn't confident about. (laughs) - Great intro.

- Yeah, yeah, yeah. But these came from previous guests. So I just think we haven't found PMF yet on those things. So acceleration, what has already happened in AI that you thought would take much longer? - The GPT, the fact that basically you can do anything. Obviously, it might not be the smartest way to do it, but you can basically do anything with GPT.

- And were you watching it while you were at Palantir? 'Cause Palantir has so many high stuff. - Yeah, I think, vaguely, I think it really blew up after. - Yeah. - I think Palantir really-- - You just happened to be the right spot where Moonshot wasn't super working.

- Yeah. - So you were open enough to actually pursue this, seriously. - Yeah, and I think Palantir also just shifted the whole strategy once it blew up the way it did. - Yeah, yeah, awesome. Exploration, if you weren't doing what you're doing, what do you think is the most interesting, unsolved question in AI?

- Hmm. Great question. There's so many things. - Like what? - There's the context length, there's just the-- - You want longer context length? - The latency and how to just make everything flow through much quicker, 'cause I think you can then enable some really interesting real-time applications. - Are you interested in real-time?

Actually, I was meaning to think about that. - Not necessarily. - Right, because your whole thing is async, right? - Maybe, yeah, but maybe in the future for some other product. But as a person, you know, it's interesting. And also just the ability to, really the interesting stuff is the validation, right?

I actually think it's the QA that we were talking about. - We need evals, yeah, yeah. - You still need human QA. - Yeah, especially yours is the hardest to QA. - Yeah. - I would say. Okay, very cool. - It's very time-consuming. - Takeaway, what is one message you want all of our listeners to remember?

Take away with them. - Okay, if you would like to start a podcast, start. We're here to help. Super easy. If you have a podcast, we wanna help you make it more accessible by dubbing it. On the other side, if you are a founder, an AI engineer, I think it's really important to convince yourself that what you're building is valuable.

Don't listen to people saying, I have a motor, you don't have a motor. Convince yourself of what that is and launch. Launch and don't burn that much money. - Frequently and often, don't spend your money. - Yeah, yeah, be smart about it. - Yeah, I think you are one of the most successful cases of AI engineers so far.

I'm really glad to spend time with you in person and excited to see what comes next. - Yeah, it was great coming here. Great meeting you guys in London. - Yeah. - And see you soon. - All right, see you. - Bye. (laughs) - How was that? - Great, man.

Hopefully you enjoyed it. - Very much, very much. It was a good conversation.