Nathan Baschez - How AI is revolutionizing the way you write on the Internet!

I think that there are some people who start businesses from a very like zoomed out standpoint. Like they're almost like put on their hedge fund manager hat and they're like, what are the macro trends that are changing? And then like they triangulate from there into some like industry and then they sort of research it a whole lot and they like talk to people and like that probably can work. The downside of just solving problems for yourself is like maybe it's actually like a small market, you know, or maybe, maybe it's like a hard business that you might get in.

Welcome to the Undefeated Underdogs Podcast, where I unpack and narrate stories of ambitious people who turn obstacles into opportunities. My goal for this podcast is to create a platform to narrate underdog stories and maybe play a small, teeny tiny role in inspiring you. I intend to highlight the underdog mentality and make authentic conversations with people who play the long game, take action with a chip on their shoulder, and convert obstacles into opportunities. Buckle up as I'll be bringing some authentic founders, VCs, community builders, and content creators who got underestimated their whole lives, and yet they beat all the odds to become insanely successful. Now today I want to tell you a little bit about our awesome sponsor acquire.com. Selling a business is as tough as building a business. As someone who went through this process once, selling my own startup, I know the pain it takes to get to the end zone. This is where our sponsor shines. Imagine this. You're a founder who's built a solid SaaS product, acquired customers, and generating consistent monthly revenue. The problem is you're not growing forever for whatever reason, lack of focus, lack of skill, or just plain lack of interest. And you feel stuck. What should you do? The story I'd like to hear is you buckle down, somehow reignited the fire, get past yourself. And the clichés and start working on your business rather than just in the business. You start building an audience, move out of your comfort zone to do sales and marketing, and in 6 months you triple your revenue. The reality isn't as simple. Situations may be different for every founder facing this crossroads, but too many times the story ends up being one of inaction and stagnation until the become business— the business becomes less valuable or worse, worthless. If you find yourself here or your story is likely headed down a similar road, I offer you a third option. Consider selling your business on acquire.com. Capitalizing on the value of your time is a smart move. Acquire.com is free to list and they've helped hundreds of founders already. Go to try.acquire.com/sharath and see for yourself if this is the right option for you. Now let's get into today's episode. All right. We all agree on one thing. Writing is hard. It's freaking hard. My next guest is actually solving something about writing. Ever faced like writer's block or is it like feeling like, okay, I can't write anymore. I need some help. My next guest, Nathan Bachez, is building Lex.page, which helps you accelerate writing process. Like wherever you feel like difficulty, you know, be it writing on the internet, Is building a tool like, you know, it's like a genie that pops up and wishes you and eliminates all the hurdles for you. He's also like a, you know, he's a product manager. He's like a developer. He's the first employee of Substack. I can go on and on and on about this guy and I'm pumped. And he's the first person to actually write code for Product Hunt. That's one thing I want to definitely cover. But all things aside, I love this guy so much. You know, he's very authentic on the internet, especially on Twitter, and I followed him so many times. His advice is like priceless. I'm pumped. Nathan, welcome to the show. How are you feeling?

I'm feeling great after that intro, man. I need to listen to that recording like as my alarm clock in the mornings or something like that. Wow. I mean, yeah, thank you so much for having me. It's great. I'm excited to chat.

Yeah, you're an all-rounder, so you did a lot. So I think if I didn't do a good job introducing you in the proper way, I'm doing a bad job as a host. So I hope I did like a good job, at least like, you know, bringing some of the things you did in the past. But again, appreciate you coming on.

Maybe we'll see if we can pay it off via content. But yeah, no, I'm excited.

All right. So right off the bat, did you buy Apple's Vision Pro?

Did I buy? No. Oh man, I thought about it. I was very, very tempted. And I think for me, if I had more money, probably I would. But, you know, I just have had this feeling of like, you know, I want, I want to like stay focused. I'm like trying to just really focus on Lex. And the conclusion I came to after hearing some of the early reviews and talking to folks is like, it's really, really cool. It's unlikely that there's gonna be significant numbers of people anytime in the next year or 2 who are like using it as a primary device to do writing, which would be the thing that would push it over the edge of like, maybe we should think about something with it. And even if I tried it, it would be like, I, the bet was, okay, obviously like Lex as is needs to work fine in there, but I mean, websites seem to basically work fine, you know? Like it's not like there's a bunch of quirks or whatever. And I can just test it with, I'm sure I can find someone who has one that I could like use if there was any like real issues. But, um, I, I will say though, eventually I 100% believe that people will be writing in some sort of device like that. And it'll be an amazing experience to be able to have, you know, your notes, your writing, maybe chat, maybe different layers of the experience that are visible in cool ways that, that don't make as much sense when you're not in an immersive environment with, you know, very limited screen real estate. Like, Mm-hmm. Long-term, 100%. So like, will I buy it a year from now? Will I buy the next version? Like, probably, probably. It's just sort of, we've got a lot of other stuff that's just higher priority to focus on. And I'm like, I would waste way too much time playing in this thing if I got it. So yeah, and it was hard to justify as like a Netflix device for airplanes and stuff like that. Although that, it does seem amazing. But yeah.

Yeah, I 100% agree. I feel they're early, you know, I think VR, AR, it's not new, but I think I believe that the market is not mature enough to adapt as quickly as we all adapted smartphones, you know, for example.

Yeah.

It's not like too immediate, like, okay, it's like something that I use, you know, pick up a phone, call someone. That's like a basic use case. That's why I think smartphones are, you know, when Steve Jobs launched iPhone, everybody went bonkers because they can do beyond just calling another person, you know, on a really, you know, touchscreen, which is like without any buttons, right? I think I do agree. I think at this point it's like buying your own personal theater.

Yeah.

I think that's it, right? I do see, I saw Casey Neistat's review. I saw like, you know, Marquise Brownlee. I think they're all amazing and true. I feel, I think what you said is absolutely true. I think the common use case is writing, on the internet can happen on Vision Pro probably like 3, 4 years down the line.

Totally.

Even like coding, coding your side project or something.

Designing. There's so many things that feel like they don't quite work that well yet, but for sure they will at some point. And so it'll be super important for anybody building any sort of creative tool to be ready for that. Um, I just think probably right now the state Lex is in, there's like just making the AI more useful is like the most important thing. And then we can get to like how, how great it is in an immersive interface versus a flat interface. But, um, you know, I, yeah, no doubt in my mind that, um, it's where it's headed. It's just a matter of, you know, refinement and time and, you know, letting, letting improvement like incrementally happen across a whole bunch of different dimensions, like processing power, screen quality, like battery, all that kind of stuff. But it seems like it'll probably get there. Yeah. It's like the Apple Watch, right? Like, remember when the Apple Watch first came out and it's like, cool, what can it really do? Like, you know, it was like kind of unclear and then people started to figure it out and now like it's, it's a pretty successful product. But at the time it was kind of like the first year is sort of weird. Like you didn't see that many people have it. It felt kind of like an optional thing. And then they really found their niche in health. And I think that there's going to be different niches for Vision Pro. There's some glimpses of what those are going to be, but it just takes time for it to all.

Yeah.

I think the future happens in, in a different way than a lot of people sort of expect it to or think it to sometimes. The naive interpretation is like, there's a brilliant invention, it comes out and it changes everything. That's like the storybook version of how innovation happens. And I think the real version is there's some like long-running thread of stuff that's like not really that surprising to anyone. It doesn't come outta left field. Like everybody kind of knows it's a thing. And there's a long period of like, it doesn't really work, but people are kind of trying. And then there's some period of like, there's a version that's out that kind of works, but it's not that great at a lot of stuff. And so then people are like, well, this thing kind of sucks and it's not useful for that much. And then they're like, the hype dies down and it just sort of exists for a little while. And then there's these like waves where it's like, oh, there's this thing and this one, and it kind of keeps going. And eventually you're like, oh wow, everybody seems to have these things now. Kind of. And maybe there's a moment when it's really like you're riding the S-curve steeply. of adoption. But, um, you know, I think, um, I think people are probably putting, you know, too much pressure on this iteration of the Vision Pro. Um, if they're expecting it to be like, you know, immediately everyone's, it's competing with just, you know, whatever laptop usage generally or something like that, but it's fine. I think that's the necessary order of things, you know, and it's not a bad thing.

Yeah, it is. It's definitely, I think, 100% with you on that. I think people buy hype so fast that they kind of forget about validating the substance, right? I think that's what's happened with like, you know, even Apple TV when they launched like 10, 15 years ago. I was talking to one of the engineers who worked on Apple TV on this podcast, like, you know, last week, and she was telling the same thing. So one thing, one amazing thing Apple does really well is they watch other people, other players play the game, and they take, probably take notes and kind of like launch a thing that is so advanced in the future that people are either, people are not ready and they build a lot of amazing marketing, right? Like we all know they're meant for their storytelling marketing.

Yeah.

So she was selling the same thing about Vision Pro. it's not there. Like, it's so advanced. People are like, you know, freaking out about the hype, but actually, like, kind of the real-to-real use case is going to like mature over time. But yeah, I saw you post about the FOMO. I was like, okay, did Nathan finally recover?

Did I do it? No. Yeah,

yeah, no,

I stayed strong.

You stayed strong. Yeah, there you go. I'm too late.

That's the story I'm telling myself.

We never know, right? Love that. Let's go back to 2013. I know probably, you know, on every interview you've been on, people must have asked the same question. Hey, how did you end up working on one of the best products in the world right now? One of the best startups in the world, Product Hunt, right? So I'm also, I fall into that category for sure, you know, since I am kind of part of the ecosystem, you know, for a while. And I'm forever a Product Hunt, you know, fan, follower, whatever you want to call it, because I think Product Hunt gave me everything.

Yeah.

Every freaking thing. Like the whole career I made so far is because of Product Hunt, the community, the network, you know, the love and everything. So yeah, so talk to me, take me through the story of what really happened. How did you end up writing the first lines of code for Product Hunt back in 2013?

Yeah, I was friends with Ryan Hoover. We had done like some other little side projects at the time and like met up a couple times and just, you know, we're kind of hanging out in San Francisco. And he launched the sort of prototype version of it, which was basically just an email list. And I was one of the people on the email list. And maybe a couple weeks later, when it was kind of like, ah, people really like this thing. He emailed me and he was like, hey, I'm thinking about turning into like a real website, not just an email list. Like, what do you think? And my response to the email was, I think that's great. I think it makes a ton of sense. And if you want, I could like hop on board and just build the first version with you. And he was like, cool, that sounds good. So we just started working together. It was honestly as simple as that. And, you know, there was some like back and forth on the design and like some of the key elements. I think some of the things that were sort of novel about Product Hunt were separating out the like name of the product from the tagline and having that structured data versus just like a general forum, right? Where anyone can write anything in the headline. Mm-hmm. Having dates be like the main headings, and then within that day, having all the different products that were submitted that day and then sorted by upvote within the day. And then, you know, on Ryan's side, the like community aspect of it was everything, right? Like he, it's been so fascinating. 2 of the things that I've been associated with that have been the most successful also were the most fanatical about early kind of like community and user acquisition. And, and so like on the Product Hunt side, Ryan, you know, was just had an incredible network in tech and like really like got everyone on board and got everyone excited. And was, you know, constantly DMing people, emailing people, having coffee with people to just stoke the early community there. Because he knew, like, Product Hunt is an extreme— there's the tech and the design side of things, which is like, it helps, it amplifies everything. But it's a totally different site if it's like, you know, a couple dozen, like, random folks who maybe aren't that into Silicon Valley or that excited about these things, or it's sort of, you know, It doesn't feel like the center of the world of tech, you know, versus like if you can get prominent people in tech to like in the VCs and stuff like that, like paying attention to what's on Product Hunt every day, then it gets a totally different, you know, mystique about it and value to it, frankly. And so that was huge. And the other thing that I've been a part of that was pretty successful that I saw the same thing was Substack. Hamish McKenzie, one of the co-founders, he is a writer and he, you know, formerly was a journalist, you know, full-time and obviously then went on to found Substack, but just spent all of his time talking to writers. And there's been so many other newsletter platforms that were more created by software people and they just kind of have this like build it and they'll come attitude. But really the ones that had a fanatical co-founder about talking to the target user and building the early community were the ones that stood out a lot. So maybe there's something generalizable from that. It's certainly something I'm trying to apply with Lex as much as I can. But yeah, that's, it's an interesting observation for sure.

Yeah. Again, you hold number 1 badge at Product Hunt, which is like, you know, which I'm so fascinated about because you will hold that forever for the rest of your life and rest of the history. So Kudos on that. So you've touched upon one thing which is really fascinating to me. And I want to also talk about Lex, how you're kind of implementing some of the lessons you learned working at these companies with people like Ryan. You talked about there is like a hustle, founder hustle, right? Like you just take no for an answer. You just go DM people, cold outreach. That's what you said, talking to users. Yeah. What is the secret there? I'm just curious. I know kind of the answer for it, but I would love to know from your perspective, when you start a company or a startup and you have here and there network here and there, like, you know, you have some sort of name, what is your approach of or advice to people? How do you, how do they go about, what did you experience with product hunt? Like, you know, is that, there is a secret for that? Like, what is the playbook there?

It's a, it's to play the long game. I think if there's one thing that differentiates successful kind of like hustle where you're reaching out to a lot of people to try and find investors or find customers or build an early community, the long game is so important to play. So I'll explain what I mean by that. First of all, I'll show you what the long game is not. So I'm sure if you go to Twitter and you open up your DMs, you'll see that tab message requests, you know, 12 people you may know, 100 people you may know, whatever. There's a million people in there, you know, like, hey bro, like reaching out with a quick idea, like, can I help, you know, double your conversion rate in like the next month or whatever? Like, it's just, you get a ton of those. If you said yes to every one of those, you would spend all your time just in meetings talking about, you know, bringing on potential like coaches or consultants of various kinds. Totally saturated, and it wasn't necessarily always that way. Probably something about the way that Twitter is working now with their what you can pay for and whatever else is like maybe that's a thing. But so that doesn't work, and I think people know when they're basically spamming people versus when they're not. I think a lot of people though maybe don't understand how to get past that. They just think it's I gotta like email more people, DM more people. Spam more people basically, and it's just a numbers game, right?

Mm-hmm.

If you're getting like less than a 50% response rate, like not necessarily conversion rate overall, but like response rate, you're probably doing it wrong, I think. Because at least the kind of outreach you should be doing is like plausible enough for someone to get back to you.

Hmm.

Right? And so if that's the case, then here's, here's how you play the long game first. Build relationships in a longer-term way, and start with those relationships, and develop early product-market fit with the people who already you've got some social capital with, right? Because you've helped them in some way, or you've just like been Twitter friends for a little while, right? Like that's why I'm here with you, right? We have the social capital. And so the I think I think just being a genuine person showing up. You know, like making friends basically is like step one, right? And like being useful, being helpful.

Mm-hmm.

And people wanna know how, how can I be helpful or useful? Literally just everyone has something they're working on, you know, especially in tech if you're like building a new product or whatever. Do you know how many people a day send me thoughtful feedback about Lex? Like very few. I do get it sometimes, but it's like, it's gold for anyone who's like trying to build something to just get useful, honest feedback from someone who's like legitimately in the target market of like, hey, here's what I loved. Here's what I ran into, some issues. Here's what I wish it did. You know, that kind of thing. Like record a Loom, just sign up for someone's thing and record a Loom and send it to them and be like, hey, huge fan of what you're doing. I signed up and had a few thoughts. Hopefully this is helpful. You know, let me know if there's anything else I can do, like rooting for you or whatever. Like that's so different giving versus asking, right? It's so simple, but it's like, How many people do it, right? Not, not very many. And I think also there's something to be said for bootstrapping relationships. So like, like it with Ryan, the reason why he emailed me to be a part of the initial Product Hunt list was, and the reason why he emailed me when he was thinking about turning into a website is because like we both lived in San Francisco and like knew each other through mutual connections. Mm-hmm. And so when you develop a relationship with someone and then like more relationships springboard off of that, then that's a totally different type of like, that's where you get into the, like, if you have something new and you're telling someone about it, you get up to the like 50% and above response rates because it's like in the context of a real actual relationship, you're a part of the actual human social network, not just sending someone spam in a DM or whatever. And so to me, that's like the biggest mistake that most people make when they're trying to do the kind of hustle thing is just basically thinking that the only way of doing it is by spamming and trying to spam as hard as they can and, and seeing if, you know, the law of large numbers, eventually something will work. The other big thing is once you get beyond that initial phase, to really genuinely focus on improving your value proposition. And again, that's the thing that it's so simple, but a lot of people don't focus enough about it. They think, You know, I need to like send messages to more people rather than I need to make my message more compelling. And not just my message, but the actual thing that's behind the message, the value. Because the real source of growth is when someone tells someone else, hey, you should use this thing. I use it. It's awesome. Like that's how most things actually grow in the world. It's not from the person themselves saying, I think you should use my thing, you know?

Right, right. I love that. I think you kind of knowingly or not knowingly opened up a playbook, which I kind of follow personally. Even with you, the reason you're on this show right now as we are talking is I've kind of followed you for a while. Number 2, I've kind of like, you know, I provided Lex feedback for Lex, if I remember it correctly. You know, I DM'd you, we were like going back and forth. I think this was in 2023. So it took 1 year like to kind of give— I think value is something that people abuse. Yeah. They kind of like Have this misconception that if I just provide my two cents and just stay back, right? That's it, right? It's a consistent, constant thing that you need to do with certain people that you really admire. You want to be you know make relationship with, and I feel again that's that's the playbook. And like you said, how many people will do that? Very less. And how much length they can go with? Like you know going following someone like. providing 2 cents repeatedly again and again, again without asking. So it requires a lot of patience. I think long game and being very patient with their ask is, I think, like a deadly combo. I think I would say, I feel even I've kind of landed everything what I did in the past by playing that playbook. So I love that. Let's talk about You're a product manager, you're a product person, you're like, you're a developer, you code things, you're a founder, like so many different hats you wore. You sold a company to GA, Dash. It is like a, it's like you're like an all-rounder, I would say. So what are some lessons you learned from each persona, like as a developer? Yeah. as a writer, as a founder, you know, I think you invest as well, like advice as well. So talk to me about your persona-related lessons that you kind of like, you know, match over with each other. Okay, now I have to bring my writer hat, even I'm a founder, right? So talk to me about that.

It's all been pretty natural and just a solution to immediate problems at hand. Just pragmatic, right? So like, when I was in college, I got interested in startups. I was like, you know, reading Paul Graham essays as we do. And, you know, the first thing was like, I thought of ideas for like, oh, this would be cool to build. And then, you know, I thought, well, okay, like, what's the next step? Like, there's sort of a fork in the road, either find someone who can like design it and build it, or one person design it, another person build it. Or like try myself, you know, see if I can learn the skills to do it. And so I kind of did both, right? Like I still was trying to find developers and designers to like, you know, work with potentially.

Mm-hmm.

But also I didn't want to just like only do that. I felt like I should have, you know, at least sketches, right? And so then you make the sketches look a little nicer and then you like read it, how to make, you know, UI look nice and like Then you sort of like, you pick up some design skills and then like, you know, I got in, I got, I was interested in the programming part always. I think I was for a really long time had this feeling of like, oh, I'll just learn like a little bit. I don't really wanna be like a programmer, you know, but I'll, I'll just see if I can pick up like some HTML and CSS. And then when I picked that up, like maybe a little bit of JavaScript, you know, and then you just, it's like the next incremental thing, right? And it was always very, just sort of like, there was no, um, big vision to become a programmer, but I just sort of backed into it because it was practical to do so. It was useful to do so. And then, you know, ended up like a lot of the, most of my time, like 80% when I was at Substack was writing code. You know, I also did stuff like met with writers and helped host the podcast and like answered support emails cuz it's, you know, small team. But like a lot of the time I was like writing code and, you know, the CTO of Substack is like an amazing programmer. So, um, learned a lot while I was there. Um, and, and same thing now, like spend a lot of my time writing code and reviewing code and at work with some really amazing engineers at Lex. Mm-hmm. Um, so, uh, yeah, you just kind of pick stuff up as it's practical too. And I think maybe the other thing that's been helpful is to not, not feel like, oh, if I don't want to even start because it's so hard to like learn to do the entire thing. It's like, well, if there's like a little thing you could pick up that's helpful, just do it and don't worry too much about If you have to change your whole identity to be a programmer or a designer or whatever. I think people place too much on those identities. And at

the end

of the day, there's so many common principles from all the things. Like if you think about, okay, the principles of like what makes a good essay or what makes a good interface. Good. It's like, does it conform to people's expectations? Does it offer something new that's interesting? Does it solve a problem for them? How to, you know, you just, I think it really helps to be, to learn to get good at multiple things and to sort of carry lessons from one to the other. It just kind of develops this general, uh, muscle maybe around anticipating how people will respond given something in front of them. Right. Um, Mm-hmm. 'Cause that's what it is ultimately. You're creating something that should help people in some way and, and how it helps them and how they'll react and what would be obvious and what will be confusing is all Very similar, I think. And of course there's stuff that doesn't exactly translate, but everybody's gotta write anyway, you know, like no matter what job you have, like you're gonna have to do some writing, even if it's only Slack or Discord messages to your colleagues. It's helpful. It's helpful to structure a piece of communication well, even if it's a short piece of communication.

That's brilliant. So I think what you're saying is very interesting. I think this is the first time I'm hearing, don't tie yourself to an identity, which is kind of like a very counterintuitive advice. It's like, Be contextual, whatever the thing that serves the purpose, put that hat, serve it, move on. Right? Like, don't spend too much time worrying about like, hey, questioning about, am I really a programmer? If you want to program, am I really a writer? If you want to write. As a founder, I think probably you have to do pretty much everything, right? So just serve the context and just move on. I love that advice. Probably that's something I want to implement as well. Also, Nathan, there is like a pattern in your life of what you're doing right now, which is Lex, right? What you're solving. So you were part of Substack, you built this amazing ecosystem called Every. It's also like a writer-based newsletter covering tech and everything. Now you're doing Lex. I want to definitely ask about Every, you know, the story behind why did you do that? you know, and all that. So I think there is like some sort of innate passion towards writing, writers, or some empathy towards them. That's why you're building Lex. So talk to me about why Lex, like why now when in the world of— the first thing I think when AI blew up, everybody focused on writing. I don't know, for some reason everybody's like, okay, let's actually build a writing tool. There is Copy.ai, Jasper, like so many other tools that are competing with each other, doing, solving writing problems. So why Lex and what made you attracted to that problem?

Yeah, well, it's interesting because, uh, I feel like a lot of the stuff I've worked on in my career kind of ties back to how to help people use computers and the internet to just be smarter, to be able to do bigger things than they could have done before. And so With Substack, it was, you know, helping people who maybe they had a job as a writer somewhere, or maybe they're beholden to like a publisher of some sort, go direct, own the relationship with their audience, and publish the kind of writing that if they had total control, they would really want to publish. And that led into Every because I was like, oh, this model really works. But it would be really cool if there was some way to help writers who maybe didn't have enough audience from day one to start on their own, reap the benefits from this kind of model, but do it in a context where they, you know, have access to audience already, you know, they don't have to start from zero and we can pay them day one, even if, you know, they don't have subscribers, you know, that are paying specifically just for their content. So, you know, kind of like a magazine basically, but like built on this new, this, this new model of like more, more revenue share and more independence for writers. So, and then Um, kind of what came out of that was I was writing and editing every day, you know, publishing my column Divinations or, um, you know, editing pieces that other people wrote. And Google Docs was the primary tool we're using. And I just got so frustrated with it. I was just kind of like, on a pragmatic level, it, the origin of Lex initially was something more like how Linear probably came outta frustration with Jira, right? Or like Superhuman came outta frustration with Gmail. It was just like, can't we have a better tool? You know, like,

mm-hmm.

If I right-click an image, it doesn't show me like save image as, or like copy URL or copy image or whatever. If I hit the keyboard shortcut to bring up the emoji picker, it doesn't do that for whatever reason. I don't know why it just doesn't work in Google. If I, there's all sorts of little examples like that. Like if I send you a link, 99% of the time I probably forgot to like share you on it. And so you have to request access and then go back and forth. Right. You know, the whole stuff with pagination, it doesn't work on mobile web. The mobile app, if you use it and you've got like track changes on, you type a letter on your keyboard and then you have to wait like 10 seconds for it to show up. And it's just basically, it's functionally like unusable if you have like suggested changes in a document. So all those things kind of added up and I'm like, it would be great if there was just a better tool. And I just, I, I would like to use it personally. And so like, I don't know, just nights and weekends, like I'll start hacking something together and seeing how far I can get. Right.

Hmm.

And then this whole separate train of thought was there around AI where I was kind of like curious what it could do. Right. And I had no idea. I was like, okay, it's clear that You know, GPT-3 is like really good in some ways, but it's unclear what you'd use it for. It, it creates this really impressive sounding output, but, um, you know, how would it actually solve problems for writers?

Mm-hmm.

And the sort of dot that I connected early on was like, well, okay, one of the most common problems and frustrations I have as a writer is I get stuck. I'm writing a draft, I don't know what I wanna say next. It's a painful experience. You know, you like, Beat your head against the wall. Maybe you just like go to Twitter or something, which is what I would normally do.

Mm-hmm. Absolutely.

With, with Lex, you just type plus, plus, plus, and the AI generates a paragraph. And, and the idea is not necessarily that you would use that paragraph all the time or that it's perfect. It's more just that it gives you something to do to sort of unstick you. It gets the juices flowing. It's like, oh, that's one way I could take this. Interesting. And a lot of times, even the things that it's wrong about, the, the reason why you don't wanna use it is useful cuz that gives you something to react against and you're like, ah, well no, I know what I actually do wanna say because of this. Right? Um, and so that was kind of like the first feature set, um, of Lex. And, uh, you know, did way, way, way better than I ever could have hoped for or expected. And then from there it was kind of like the big decision is, okay, well, what do we do with this? Right. And we came to the conclusion that it makes the most sense for it to be its own company. Like you, you don't wanna like bury a venture-backable tech startup inside a media company. And they're, they're basically just 2 different businesses. Right. But I never would've started working on it honestly, if I thought that inevitably it would be its own company. It just started as like a side project for Every, and then it just kind of like took over a lot of stuff. Right. In a way that was awesome and I'm super happy to be working on it now, but it was unexpected, you know? So yeah, that's kind of, I guess, the progression, the initial inspiration for Lex. The other thing I'll say that's interesting, maybe also speaks to the previous question you're asking, is the kind of ping pong in my career between being on the tech side and being on the writing side. Mm-hmm. Um, and Every was really the main and only time that I've like, you know, 90% of my time I was like working on a piece of writing, either my own or helping someone else with theirs.

Mm-hmm.

And when I was at Substack, I was, you know, writing code like for the settings screen or whatever and thinking, God, it would be so much cooler if I was like working on a newsletter that I had a cool audience and I was like making cool points and creating, you know, That, that just felt like it would be way more fun.

Hmm.

And then I did that for a while and it surprised me how I liked it, but I didn't feel like I could write forever. It felt like I had sort of like a timer on it where it's like, if the premise of this is I'm like writing about tech and strategy and startups and product and all this kind of stuff, like basically I want to be doing that. and write about it sometimes rather than writing about it all the time. And I just like felt like I couldn't do it forever. So it was a big, I think, background thing of like, I had to, I really had to just do it to learn that lesson. I didn't know beforehand that that's how I would feel, but then when I was in it, I was like, oh, this is becoming more and more clear over time. And now that I'm back to building software, I'm definitely like, Yeah, this is the thing that I'm probably gonna do my whole career is like build software and write. Like it's a big part of what I do, but it's great when it's like 10, 20%. For, for me right now, I guess maybe it, it could always change, but it feels like I see no end date for this. There's no like looming, I don't know how many things I have to say about this, you know, or how many things I, I could do that. So, um, that's an interesting thing where like. Yeah. Maybe everybody's got different preferences around that, you know, cuz then there's some people who I worked with at Every who like had been doing like, you know, tech and marketing and stuff like that. And then like it was a dream to be able to write and now it's like, thank God I get to actually just write all the time. That's what I really wanna be doing. And so it's a very personal thing and I think it helps to experiment with different stuff and try different things. And if you have this feeling of, you know, if you're really being honest with yourself, you definitely can't be doing this thing forever. Like, take that seriously because it's very possible to have like just your baseline happiness increase, you know, by like a couple, like 100, 200% just because you're doing work that's more suited to you on a day-to-day basis that you don't feel like the time is running out or something like that.

And yeah, I think that's how you would, I feel that's how people should, should discover what's play for them and differentiate what's work for them. Right. I feel you have to try different, different things. I love the fact that you mentioned that because I, for, for some reason I sense that even I think it's so inspirational when you say some things like that because it can relate to so much to it. I do community building, I do content writing, I build products. When it comes to building products, I don't know, for some reason there is this, the baseline happiness immediately, you know, appears like, you know, oh wow, like something I want to do. Okay. I want to like, you know, hack how to talk to people, how to like market it, like, you know, so on and so forth. There's the mind triggers immediately, like, okay, these are like 100 different things you can do, right? So I feel finding that itself is so— I think there is a gift. A lot of people struggle like just figuring that out. So I love that. 2 things I know I just want to talk about a lot about Lex, but I want to ask these 2 things before we dive a little deeper into, you know, The feature set, like the differentiations, all that. A, it seems like you at Substack, you were coding and you wanted to write, so you built Every. At Every, you saw, you know, how terrible, like, you know, the tool is working, not working for you. And now you wanted to, like, you know, build Lex. So kind of like you're doing some things for yourself. Yeah. Fundamentally, so there is a lot of upside for it. I do things for myself. I see I can empathize, talk about ours. You know the upsides, but I want to understand what is the downside of doing something for yourself?

Oh, good question. Um, I mean, I think that there are some people who start businesses. From a very like zoomed out standpoint, like they're almost like put on their hedge fund manager hat and they're like, what are the macro trends that are changing? And then like they triangulate from there into some like industry and then they sort of research it a whole lot and they like talk to people and like that probably can work. You know what I mean? Like, I don't, I'm not here to say, I don't, I don't know what the best way is to do things. All I know is anytime I've tried to do that, I come up with the dumbest ideas. And so, you know, I guess the downside of just solving problems for yourself is like, maybe it's actually like a small market, you know, or maybe, maybe it's like a hard business that you might get in. Like media is a hard business, right? So like, I didn't set out to be like, what's the most amazing business I could start and then do that? I just set out to like, I want to do the thing that I feel like doing for whatever combination of reasons, right? And then you learn like, Oh, maybe it's kind of a hard business. Maybe I like don't feel like having to have an opinion every week on tech things. Like you, you start to run into those things. Mm-hmm. But, um, it's hard to know before you start. And I feel like learning by doing and then just tou— and like touching the hot plate and kind of like, it's, that's how you like really learn. If you're coming at it from this like abstract macro, like zoomed out perspective, and then you like triangulate in on something that could be an area and then you find, try and find a problem or whatever and all that stuff, like For me, it for me it just doesn't doesn't work. I'm much better when I'm like, I just want to do the thing that feels useful and pragmatic to do now. You know, but yeah, I think the downside is probably I could, and I think I actually have added these layers over my career more and more. Add the sort of like well, if it worked, what would it be? And is that the most exciting thing versus some other thing or whatever? You know, like there, I think there are some. Some layers I'm adding, but I feel like I'm I've I wouldn't I don't regret the order that I've done things because like I had like one thing that we haven't talked about is I am first company hardbound. So I learned some really really important lessons from that. I was building this; it was a mobile app with like a new storytelling format that you know we were the only ones who were willing to really make stories in our format because it took a lot of work, and it was a cool kind of like a comic. Book meets an Instagram story type of a deal. And like, it was hard to make. So we made them. And when we managed to write a story about a topic that someone was interested in, and, you know, get it in front of them, and they download the app, and they try it, they'd be like, whoa, this is really cool. And I thought, well, awesome, we've done that. So we can't go wrong. Well, take a step back. What are all the assumptions of things that had to happen there? And then how do you build a system to like make that happen at scale and like build distribution into this app? And like, how are you going to build a platform? that gets other people to create things? Well, maybe you need to make creating less hard. Oh, well then who's gonna do it? And you just start to open up all these wormholes. And so like, you know, I learned from there kind of like every piece of this sort of system, every step in the equation needs to basically work and you can like refine it, but there should be no steps where you're like, nah, let's just not worry about that at all for now. You know, we'll get to that later. First, let's make this piece really awesome. And not control for these other things. Let's just focus purely on this. So it's like in Hardbound's case, that didn't work. Right. And so everything I've done since has been, has not had that problem. And then, you know, I've, I've had other problems, but I keep kind of like triangulating, I guess, and, and learning. But yeah, I do think, um, I do think there are downsides to the approach. Maybe I could have triangulated a little bit, uh, faster on some things, but you know, um, I agree with you.

Yeah, learning by doing is the best thing, right? I think I do agree. Also, I think to add on top of what you said, I feel like doing some things for yourself means you're in your own bubble. Like, often you kind of go into the same loop again and again and again. You make progress, but you kind of like move away from reality in a way because you're doing it so passionately for yourself. I think there is, there should be like another persona inside you that could, that should question, hey, is it like really worth the effort and time that you put in yourself like while building this? So yeah, I'm not saying that you need to play like a devil's advocate. That's not being like, or being a pessimist about the thing you really love to do. It's just that you need to like, you know, swing a little bit on the outside and see like, you know, from a third perspective. third type, whatever it is, right? So, yeah, another question. I feel like you seem to be very self-aware, you know, you kind of like pick those signals that you want to do, you don't want to do. So talk to me about, I don't know, it's really fundamentally an important skill for a founder to develop self-awareness, right? But in your opinion, how should one develop that skill? When they, especially in startups.

Yeah, it's super hard, honestly. I mean, in some ways I'm probably pretty good at it. In other ways I'm definitely not good at being self-aware. Like the whole thing that I realized about how I probably had like some sort of expiration date for how long I could have an opinion about a tech thing every week, you know, and be doing that versus building software. Um, at some level I probably knew that that was true, but I was really it was hard to look at it because the implication of what that would mean is, well, what's my role here at this company that I am a co-founder of? You know, that's kind of like what we do is publish writing. So like the main thing that needs doing is writing and finding writers and editing the writing. And like, you know, I also knew that I didn't want to just like edit other people's writing all the time. So I think to have courage to look at things that, well, this feels like it's probably true. And even though the consequences of that are like maybe big or scary potentially, like it's worth looking at it, you know? And so I think maybe the, um, the thing I would encourage is to really listen to yourself and your motivation. And when it feels like you're kind of kicking the can down the road on something that's like not quite working and you're trying to live with it. Just facing it directly, thinking through like what's the worst that could happen, talking to people about it, all that kind of stuff. I could have done a lot more of that a lot sooner, but I was very myopically focused on well, I just need to make this work. I can worry about that other stuff later. Like we need to we need to grow this thing more, you know? And and also what it meant for you know, like if you co-found a business with someone, like it's tough to be like I'm not sure that. The thing that this business does is something that I would be happy doing indefinitely, right? It's hard to know that before you do it because when I started it, I was like, "This seems great. I would love to do this indefinitely." And then about a year and a half in, it was kind of like, "Oh, you know, I don't know if it actually is for me indefinitely." And so, yeah, I think I think there but there's a lot of different meanings of self-aware. I mean, another version of self-aware is kind of like. The thing we talked about at the very beginning of the conversation, like, are you just spamming people? Like, versus, you know what I mean? Versus like actually building kind of some sense of playing a long game and building real relationships and trying to help and not just have it be totally, uh, one-way transactional kind of a thing. Um, and so I, you know, I think it's kind of, um, it's a multi— it's a multifaceted question, uh, as ChatGPT would say. But, um, you know, it's, uh, it's a good one for sure. And I don't know if there's one simple answer other than, um, Probably listen to the things that you kind of feel but might be a little scary to look at. Like, just try and look at those, make them explicit, talk to someone about it, and really, just as a thought experiment, think through what if. What if I actually went with this feeling that I have, and how scary and bad would that be? And what's the right way to go about doing it? Because a lot of times what people do is they ignore it, and then it comes out in like a worse way later on. That's something that I'm for sure guilty of, where you just, you don't. kind of courageously look and say, well, what's really going on here and what should really be done about it? And you just kind of like suppress it, you know? But it always comes out, you know? So that I would, I would say that's probably my best advice if I could give it to myself like a couple years ago.

I love that. I think one thing I find of, I kind of practice being self-awareness is taking a lot of falls. You know, intentionally, you know, just to add your point, right? You basically have to shoot many shots. Not many shots will go into the basket. You should be aware of that, but you have to do it again, again, and again and again. You know, some things work, some things doesn't. I think by falling in that metaphor, like not going through the basket, is okay, maybe I should adjust, you know, maybe I should— you know, take a leap or whatever it is, like, you know, according to basketball reference. That's a terrible way of explaining. But the point is, I think you basically take a lot of shots. I think you become more self-aware by taking a lot of shots. I think that's one thing I kind of practice myself is to take a lot of shots, you know. What is the— so you seem to be like an idea guy. What is the craziest idea you ever had?

Oh man, craziest. I've had a lot of dumb ideas. I don't know if it would qualify as like crazy, but it's just kind of like funny. Like, this is an example. And actually some places are starting to do this now. It's like the, it's like a funny thing where you like think of something and then places do it. So this is so stupid. This is one of the first ideas I had when I first was thinking about like building some sort of business at some point. So ice cream, right?

Mm-hmm.

The best part is like when you get the right combo of ice cream and cone. But

yes,

a cone is like you get all ice cream and then you get basically all cone. So like this is solved in other areas. It's called a dip, right? So like you dip a chip in salsa or like whatever, you know? So like why don't you make ice cream that's like just soft enough to support like a chip, like a cone chip basically, right? Like, that would be great. I, to me, I still think it's brilliant. There are some places, like there's this place, um, in LA called Ginny's where when you get a bowl, they'll give you like a little cone chip, but it's like just one. Like, I want a little baggie, you know? Like, I want, like, it's like when you, when you get a burrito and you get the little baggie of chips, like I want that, you know? Um, so, uh, yeah, nothing to do with tech. Um, but yeah, they're, they're throwing it out for you.

I was, I was more expecting about like some sort of tech idea that, that probably will inspire me and the listeners. But yeah, I'll take that. I'll take that. Probably

if you want tech ideas, I got them. I got a startup ideas document we can like go through or whatever.

Yeah, yeah, I mean, you know, drop one or two. Like, that's like, if you— I think the question more I want to explore here is you seem more of an idea guy. If you weren't doing Lex, what would you do? You know, if given a chance, for example, hypothetically in an alternative world, what would Nathan do or focus on? If we were—

that's a good question.

Um, Lex,

I'm looking at my startup ideas document right now. Um, it's so funny. There's some, there's some ideas in here that are like, what's that?

How many, how many of their, uh, ideas are there in the document? I'm just curious.

Uh, it's a bulleted list. Let me see. Make it a numbered list. Uh, only 8 right now. Cuz I don't, I don't add things to it very often. Like it's not like an active thought in my head. I'm really pretty, I'm pretty focused on Lex and I'm pretty happy with Lex. Okay. So there's 2, I have 2 ideas that if I weren't working on Lex, I would probably do. I don't know which one I would pick. It kind of depends on, maybe I'd do a little bit of both for a little while and see which one felt like it was gaining more steam. One is more of a layup and the other one is like a bigger, a bigger shot. Yeah. So the layup one, um, in the process of building Lex, there are so many basic things that we need to set up in order to create a good system for communicating with and learning from users, I feel like you could kind of bring them under one umbrella and make a much better product. So basically combine FAQ help desk type stuff with changelog updates with the like chat, chatbot type experience of just like, you know, how allow the user to ask questions and then have AI answer it for them about like what our product does.

Hmm.

And like. Making a better experience for teams to more efficiently keep their things up to date. Like, could you connect with GitHub, for example? And like, when stuff gets merged into main, it like tries to figure out what it is, maybe asks some questions of the person who merged it if it needs to, and then like auto-populate changelog entries based on that, and then auto-update your FAQs, which are more evergreen. Mm-hmm. And there's like, Stock and flow, right? Like your FAQs are kind of like your stock and then your changelog is your flow, right? So like, but they, they're separate. You need both of them. And no one's really built like a great thing for this. We use Canny and I love Canny for collecting user feedback, but it doesn't do all the things that I just described. And I feel like with AI now, you could, you could probably make something that's awesome that would save teams a lot of time. So there's kind of like a save time, save money angle on it, but also I think Help people get to product market fit faster or maintain product market fit. Like no matter what kind of stage your business is at, you always need to be reacting, understanding and reacting to like how people are behaving. So also if you can connect it to analytics data, subscription data, all the other stuff, like here's what your most valuable users are saying, here's what they're asking, that kind of stuff. Like here's what features people are not using. Here's what features people are using, kind of like some sort of smart AI layer around. the information flow to and from customers is like, that's one idea. And it's funny cuz I, to me it's sort of like a layup, but it's, it's, that's, I think not really an easy or straightforward thing to build. But I don't mean a layup in the sense of, um, it would be easy to build. I mean more like, I think that's a good SaaS business. I don't know if it's a venture SaaS business, like maybe it is, it's just, it's very crowded market, right? So that part is kind of tough.

Yeah. Yeah.

The other idea is, um, Instead of focusing around a writing creation tool, focusing around a reading experience that's mediated by AI. So, you know, there's a lot of companies that have sort of attempted to do whatever, like article curation and stuff like that. And I think all of it's pretty much wrong. And the thing that I would do is build like, okay, so people go to ChatGPT for like Google. What's the AI over— well, if ChatGPT is to Google as question mark is to like Twitter, what's the place you go out of habit to like learn new things about the world? So it's like news and more is what I would call it. Like, and it, but it's mediated by AI. And, um, like I, the, you know, one version of this is like, you could imagine like getting updates about the news and you can ask questions and all that kind of stuff. But like just the news is sort of weird, right? Like what? I don't think people actually care that much about like, quote, the news. Like you have to have a better angle than that. So this one's extremely fuzzy and it's way less of a layup because it's consumer. And it's one of those things where if you can build it in the kind of way that gets it a network effect, maybe it could be a really huge promising thing.

Sure.

But it's super high risk. It's like, you know, very, very speculative and it's way— I feel like Lex is sort of in a better, to me it's like the best of both worlds of like, you have this really high-level creative, interesting question of like, how do you help AI help people write better? And there's a lot of people that wanna write. Right. And I think there's always going to be people that, a lot of people that wanna write and need to write effectively. And it's just a thing that you need to learn or, and somehow there's gonna be a way to teach you. And I think built into the text editor is like the best way to help make your writing better. And, um, you know, That's kind of in this nice middle ground of like, it's not just a B2B SaaS tool. Uh, cuz for me, I like things that have something that's a little more like, uh, it's like it makes people smarter somehow, or like helps upgrade you as just like a human being. Right. Uh, but it's not so aspirational and out there as like the, what's the like infinite article that is like the perfect thing that's, that, that talks back with me and it's like my, my tutor or teacher type person that tells me things about the world. Um, Right. That's like more out there and aspirational. Um, maybe one other idea that's more on this like B2B SaaS side of things that I've just, personal pain that we have is something around, um, helping companies building these kind of AI products with evaluating if their, mm-hmm, like prompts are doing well, the performance of these things. There's some companies that do this, but I don't think anyone does it in the way that I would want it to yet.

Mm-hmm.

So. Can you build an interface that helps humans create data? Can you use LLMs to create more data? And, um, can you import data that you already have? And then how do you use that data to flow into prompts to sort of like test, hey, which version of the prompt performs better? How do you evaluate the results? Can humans evaluate results? Can AI evaluate results? Like this whole problem is a big one. Mm-hmm. Um, cuz anytime you get past very simple testing on vibes or whatever, you need to do something. And it seems like right now most people just kind of roll their own thing and it's software engineers who have the paradigm in their head of like unit testing. Right. Or it's machine learning engineers who have the paradigm in their head of like evals, which is more right. But I just think the tooling around it for software companies is not what it could be, cuz it comes from this more of like open, open researchy, you know, AI research world, which is not as practical for like just software companies that wanna build useful features. Um, and wanna spend a minimum time and have a really nice workflow. So yeah.

Whoa, you just dropped some of the awesome ideas. I hope, like, you know, I'm inspired. I think probably, like, I don't know if I can build it, but I love to, you know, see people bring these ideas to life. Let's talk a little bit about AI. Yeah. I know, I think the dumbest question would be like, hey, what's the future, right? But I would say, do you think AI is like the wild, wild west? You think people hit the gold rush? Where do you think people are right now in terms of the current reality check with AI?

Yeah, I think we're kind of— I think we're in the sort of trough of sorrow of like, we had this huge thing and it doesn't feel like it, right? But I think compared to what it will feel like in a couple of years, that's how we'll look back. And this is, I think, what this period of time will be known as. So And specifically, I'm kind of talking about like large language models, right? Because AI is a really huge— it's like, um, I'd be like saying, what do you think of like, um, software or whatever? You know what I mean? It's like a big thing. Um, but so specifically with large language models, I think where we're at is it became clear, you know, about a year and a little and change ago, year and a half maybe almost, that large language models were capable of reasoning. And that, that would be very useful in a general purpose way across basically all software. And that every category of software is gonna be, if not transformed by this, then at least significantly improved and upgraded. It's like a new material just dropped, you know, like all of a sudden we were living in a world where we had, you know, whatever, bronze, and now we have steel kind of a thing. And so like it, it, it flows through, or like plastics or whatever, you know what I mean? Like it just flows through. Mm-hmm. All software. And so that was kind of euphoric because it didn't, but it didn't quite work yet. And it's more like it basically works and it's clear with like a bit more improvement, it would like really, really work. And so everyone's like trying to figure out how do we build the systems? What is the right way to architect these things? Like all this kind of stuff. And now we're in this phase where there's this whole wave of companies that got created and funded and, you know, are growing. based on this. And then there's always the existing companies that adopted it, built it into their technology. And now it's kind of just like we're figuring it out. It's getting deployed. People are figuring out what patterns work. People are refining things and it just, it takes longer than it might seem to do that. Right. Or like, Mm-hmm. 'Cause like usually the thing that kind of feels like, oh, it's a wow demo or whatever, like it takes a lot more work to actually make it A growing sustainable business, and so I think what we'll see over the next couple years is some companies that were started during this period will emerge as like native new things that are awesome and will grow on to be amazing companies. Of course, a lot it won't work out. You know, maybe there'll be some nice acquisitions, and there'll just be a bunch of shutdowns. And then, you know, the technology will keep progressing. It'll get better, and all the things that like work kind of okay today will start to work really amazing. And also, it'll unlock new use cases that. It doesn't work at all today, but because people aren't really trying, it doesn't seem like possible. And so, um, I think that, and also the other big thing is the cost is gonna go down, right? Um, mm-hmm. Because as the economies of scale get built around like, you know, the data centers and the manufacturing of the chips, and there's just like this whole bottleneck right now around GPUs that, um,

yeah,

I think inevitably at some point it will be a little bit more chill than it is now. Um, and so the cost will go, also open source will become a thing and just like in general, the world kind of economic patterns, like take some time to sort of sort themselves out, but eventually it'll kind of sort itself out, which leads to efficiency, which leads to cost cuts, which leads to people being able to deploy it in more places for more kinds of tasks. Right? Because right now you can only really deploy it in areas where you can support the cost of it. Sure. Um, which is not as, not as much obviously as, um, as it could be. So, um, anyway, yeah, I think, um, I, I think it's an exciting time for sure, but I, I think it's a little bit of a lull if you ask me, which sounds kind of crazy cuz there's so much stuff still happening, but compared to like a year ago, it was sort of a lull. But I think over the next year or 2, we'll start to see these things kind of establish themselves as like, oh wow, they made it, you know, it worked.

So what is— absolutely, I think it's— yeah, I keep hearing the same thing, which what you said is we're still early. There's a lot much to do. You know, the maturity is not there yet. The midwit theme, right? Like, you know, we're still not there yet. You know, it takes some time. What is it, what, in your opinion, what is a use case that's impossible right now becomes possible in the future using AI?

Oh, that's a, that's a tough one. I think basically there's something about continuous learning that is fascinating to me that I think will unlock a lot. So right now you pre-train a model, it costs a lot of money to build a competitive, big large language model. And then you can do reinforcement learning from human feedback on it, and that kind of gives it a particular flavor, right? Mm-hmm. And then kind of it just gets put in developers' hands, and then there's not a lot that can happen. You can sort of fine-tune it, but you can't really teach it new things through fine-tuning. So you basically have to inject context in the prompt in some sort of smart way and just have it apply its reasoning to that context that you injected. But the whole process of actually figuring out what's the right context to inject is like, I don't know what's like kind of near here in vector space, or can we hallucinate an answer and then find what stuff is actually near that in vector space? If that's like a better method, there's like some, you know, different tricks that people are doing. But, um, It's not how the brain works, right?

Mm-hmm.

The brain is, you're, I'm both being changed by this conversation in terms of like, I'm learning, I'm creating new memories. It's like adding new stuff into my brain. And also you're running inference at the exact same time on like the process of running inference changes the output basically. Right. Mm-hmm. Um, so I'm, you're running inference on the Nathan neural net and I'm spitting out these words and these movements and whatever my neurons are like doing things. And it's also changing my neural network all the time. It's this continuous process. And so it feels inevitable that the best models of the future will kind of feel, will behave more like that probably. Um,

and

we won't have these like frozen in time, oh, here's where the training dates cut off and anything else you need to just throw in the prompt somehow. We'll see. But I think that, that, um, that would have huge ramifications. And then the next thing that would make it have even bigger ramifications is if that could be personalized per user. Mm-hmm. Or per organization or whatever, but like easy to make lots of copies of that. Right now it's not easy to do, to sort of copy paste a model and have it learn for a person. But if we have all these kind of like personalized, really smart systems that continually learn, that's, that, that would be sort of like a crazy wild next thing for it. This is highly speculative though. I'm not probably the best person to answer this question. There's a lot of people who are, have much more knowledge than me about the actual underlying technology. I know the basics of like how they work, but you know, we're not creating our own large language models, we're using them. It's an interesting question though, for sure.

Yeah, yeah, no, I appreciate your self-awareness. Let's talk about, let's, you know, a couple questions. I know we are kind of over time, but one of the, you know, one of our audience asked this. I want to cover this as well. Given there is so much competition, you know, first, A, do you treat other players as competitors? One, number one. Number 2, Why will people switch to Lex when there is like an existing tool that is kind of offering a similar thing? What do you see? How do you differentiate Lex versus, you know, all these players?

Yeah, I think if you look at it at the level of like AI and writing tool, then there's like obviously lots of competition. The way that I look at it, I don't know of any other company that's as focused on the very specific thing that we're focused on. And, you know, we need to do a better job obviously of like communicating what we're focused on and developing the product around it. It's one of those things that like, there's a lot of little things that make the difference rather than like just one simple thing you can do. But the thing that we're focused on, which is both the reason why to switch to Lex and what makes us, what makes us different, is having AI that increases your odds of success with your writing, whatever success means to you. Most AI writing tools are like, can I quickly generate something? Like, just do it for me, right? So like, you know, Jasper, the goal of it is you fill out a wizard, it generates the piece of writing that you wanted it to, and then you like put it in your CMS and you just do that a whole bunch of times. And it's like cheaper than hiring someone off of Upwork to write an SEO blog post for you, right? Or to write copy for like a Google ad or a Facebook ad or whatever. You know, and there's, there's other tools that are, that are kind of, most of the tools in the space are basically some variant on that. It's like, have the AI do the thing for me that previously a human needed to do to write.

Mm-hmm.

Lex's theory is there's a certain type of content for which that's good enough and it's worth it, but there's a lot of other type of content where actually it's not good enough and it won't be good enough for a long time. Mm-hmm. And, um, always there's actually gonna be demand for human beings to sort of be in the driver's seat if they wanna freeze their thoughts in time and put them down in permanent form. That's just useful, right? Like, that's it. Writing is a useful technology cuz as long as humans are in charge of like deciding about things, which seems like hopefully for a really long time that'll be the case, then you'll need some way to freeze in time your thoughts, which is, that's what writing is. Mm-hmm.

Um,

and almost everything starts with writing, right? Lex started with piece of writing about like, what should Lex be? Um, if you wanna make a podcast or a YouTube video, you wrote down some stuff before we had this chat, right? Yeah, absolutely. Um, like there's, everything starts with writing. And so can the AI guide you towards success is the big question. Mm-hmm. And how should it guide you and what should the interface feel like? We are like 0.1% of the way towards that, right?

Mm-hmm.

Um, I even think given current model capabilities, we are probably only like 10% of the way towards what it could be. Right? It just takes a lot of time to sort of build a system that helps people both from the early stages of conceptualizing what it even is you're trying to do here and giving you feedback at a high level to the like middle phase where you're kind of, you're, you're, you're creating the first draft, you're creating the outline, whatever, whatever it is that you're working on to the final phase of like polish. If, if it needs polish, if it's like a final product that you're gonna ship. Right. Um, there's just so many opportunities for things to go wrong and I truly like deeply believe that AI can help. with every stage of the process and just make people a lot more successful. And not in a kind of like, this is my thing about most AI writing tools is like, they're kind of like Auto-Tune. Like you can make a bad singer sound good with Auto-Tune, but it sounds like Auto-Tune basically, right? Yeah. So they all kind of, it's, it flattens it. Everybody sounds the same. And like, I'd rather have Lex be, instead of like, like maybe there's, you could turn on Auto-Tune if you want, but ultimately that's not that valuable. The real thing is for Lex to be more like, A Rick Rubin, where it's like, how do you uncover what you're actually trying to do and say? Like, that is the much more valuable thing. And like, as far as I know, not only is no one really doing it, I don't think any other writing tool is even really thinking about it that way. It's just like, literally, there's a competitor called Wordtune, right? So if it's like, you know what I mean? Like, I think people, I think people think about it mostly like, it's, it's do it for me, automate it for me, and like kind of the Auto-Tune thing. Which, look, there is value to that. And like, I think for bringing up the floor of writing, that's great. But like, and, and we should have that and we do have that as a part of Lex. But I think raising the ceiling and helping you think more deeply about what you're trying to do and, oh, at a high level, how you guide people towards success. Um, that is, I think that's the reason why people are gonna choose what tool they write in the future. And I hope it's Lex. Mm-hmm. But if you think about it from like an economic standpoint, To write, there's sort of a cost and an anticipated or desired outcome. And so most writing tools now are like totally just focused on the cost side of things, but the outcome is what really matters, right? And so we're focused on the outcome side of things.

Man, what an amazing way to describe or metaphorically say what you're doing, right? I love the comparison about Auto-Tune versus Rick Rubin being a producer.

Nice.

Maybe you should use that in one of the landing pages. I don't know, like you've been testing copy on the landing page, right? Probably that's, I think, the fantastic way to differentiate Lex from other players. Like there are players who do Auto-Tune, we are the Rick Rubin of music or writing, for example. I love that analogy. And I also like kind of kind of learning from you is the fact that AI is not going to be, or deliver, it's not going to deliver a perfect outcome. It's going to guide you to the perfect outcome. I feel that realization is something that I'm kind of also, you know, adapting to myself. I'm also kind of building like a side project, which is AI-driven. It's basically what it does is it's called GuestLab. Sorry for the plug, shameless plug.

But yeah, I'm excited to hear about it.

It's basically, I'm a podcaster. I do research. I spent like 3, 4 hours on you talking about doing like a little bit about what you did in the past and all that. It's human work, right? So the thing I'm building is going to kind of speed up that research by, for example, what we do, we're right now solving is to take your LinkedIn, plug it there. It's going to, AI is going to give you 10 topics to discuss, 20 questions to talk about. And some of these questions are generated by AI. I just literally printed out, I wrote some notes, but I just printed out some of the questions from AI. So, they're thoughtful, they're insightful, you can kind of like mix and match. So, it's called Guest Lab, but that's what we do, like, you know, kind of accelerate. Cool. And same thing, in the beginning, I thought like the AI is going to give you, give me the perfect research. But as I'm using the tool, it's no, it actually helped me 6 hours to 3 hours.

Yeah.

Hopefully 3 hours to like 1 hour. Like that's the goal so that I as a host become better doing research in the most fast way. And AI is going to guide me, hey, that's not the right question. Probably this might be the right question, right? So

totally. The thing to me that would make it even more incredible is if you had I mean, maybe you're already doing this, but like you studied what makes successful interviews, right? And like figure out the principles behind it and apply it as like a— Yeah. Yep.

Yep. Manually. Yes. Yes. Manually. Yeah. I know kind of like, you know, what makes an interview successful. Yeah. Go ahead. Go ahead. I didn't mean to stop you.

But yeah, just like if you could help people. So like, this is my theory is like, um, for a lot of these types of things, um, there's actually some pretty fairly simple principles. And you just actually need to remember to apply them consistently. And that's the hard part is people just forget, or they apply them in too straightforward of a way. And it seems it comes off as very cliché. But like, to help people sort of creatively with their own voice and saying what they want to say, package it in a way that fits the kind of things that human minds like can interpret and understand and it feels clear and it feels interesting is like, I don't think it's this magic mysterious thing that it takes special, you know, charisma from, from God or something like that. It's like, I think it's more like the checklist manifesto of just like, there's some basic rules and you actually just need to remember them. Yeah. Um, and that'll take you 80% of the way, maybe not 100%, but like pretty far. Um, so, uh, that's the goal of a lot of what we're doing with Lex is to, apply some of that idea. That if I had to say maybe there's like one, like sort of Peter Thiel, like what's your secret or whatever, that I think maybe that would be it is, um, it's a lot more like you just need to remember to wash your hands, which is the famous example from The Checklist Manifesto is like doctors just like washing their hands. If they just remembered to wash their hands, they would like have a much better ratio of like patients that make it through complex surgeries versus that don't.

Sure.

Um, I think for, for writing, it's The same thing.

I love that. I probably like, I might steal it from you as well. Like, what is the fundamentals of interviewing? And probably apply that over time in the tool. Like, same thing what you did with writing.

Yeah.

Nathan, I think we can go on and on, but I also want to respect your time. And, you know, this has been a fun, fun, fun conversation. Like, there were predictions, there were lessons, There was philosophy. There is like some innate personal experience that you shared with us today. Genuinely appreciate you being on the show. Any closing thoughts before we wrap up?

Oh, just, you know, thank you so much for having me. It's been— I love chatting. And if anyone who's listening, you know, wants to chat or connect, Twitter's probably the best way. I'd love to hear if there's any like questions or whatever. I'm always happy to hang out on Twitter. So just @mention me.

Yeah. And do check out Lex. lex.page is the site. You know, it's amazing. I love the Lex logo, by the way. Whoever did that phenomenal job.

Hamish. Different Hamish, not the Hamish from Substack. Hamish Smith. But yeah, Hamish Smith. He's incredible.

Awesome. Yeah, we'll put all the links in the show notes. And folks, that's a wrap. Hopefully you learned one or two from the conversation and do Do follow Nathan. What he's doing is amazing. I think, you know, we need more writers, period. And if AI is going to guide us, that's the best way to attack the problem.

Good writing is good thinking. Good thinking is, I don't know, everything. It's the key to the future, isn't it? So yeah.

Awesome. I think that's it. That's the formula. Till then, you know, cheers. Have a good day and stay tuned for more episodes. Cheers.