Transcript: 'He Got Thousands of Users Before His AI App Even Launched'

The transcript of AI & I with Monologue general manager Naveen Naidu and Every COO Brandon Gell is below. Watch on X or YouTube, or listen on Spotify or Apple Podcasts.

Timestamps

1. Introduction: 00:01:27
2. A live demo of Monologue: 00:03:51
3. Hard lessons from Naveen’s years in the wilderness: 00:06:27
4. Building a muscle to ship fast: 00:12:29
5. The spark that became Monologue: 00:21:11
6. Dogfooding your way to a killer feature: 00:26:09
7. Why the harshest product feedback is the most valuable: 00:29:45
8. Every’s strategy for launching an app in a crowded space: 00:31:47
9. Giving Monologue the Every “smell”: 00:40:08
10. Naveen’s one-person AI stack to build beautiful apps: 00:45:09

Transcript

(00:00:00)

Dan Shipper

Brandon, Naveen, welcome to the show.

Naveen Naidu

Hey, Dan. Hey, Brandon.

Dan Shipper

It's really good to have you.

Naveen Naidu

I'm really happy. This is the first time recording a podcast for me and to be doing it with Dan, who I've been following for the past five years, it’s a dream come true here for me.

Dan Shipper

Aw, that makes me so happy. I'm glad this could be your first podcast for a lot of different reasons. And just so happy to have been working with you for the last year or so. So for people who don't know, you are an EIR at Every and you're the GM of Monologue, which is a smart dictation app, which we are launching publicly very soon.

By the time this episode is out, I believe, Monologue will be fully publicly released. It's currently in beta and it's such a cool product. It's so well made. It's a product that everyone uses internally. You've done an incredible job with it. You've built it completely by yourself, and it's already over $1,000 a month in recurring revenue without us even launching it. It's just growing. It's so cool. How does that feel?

Naveen Naidu

It feels—I'm first of all really grateful to work with Every, because the $1,000 thing—I think if you're a normal indie hacker, just reaching that milestone takes you years upon years. For me, having the Every audience, having you and the Every team behind me, is just like a dream team. And obviously the product needs to work, everything needs to—you need to solve a proper problem. But yeah, I'm really excited, nervous—everything is happening at the same time.

Dan Shipper

I love that. And we also have Brandon here. Brandon is now the COO of Every—the last time you were on this podcast, you were only the Head of Studio. How does it feel?

Brandon Gell

It feels great. I'm echoing what Naveen said—dream come true with the dream team. Building dream products.

Dan Shipper

I love it. So I want to start—I want to talk about a little bit about your journey to Monologue. But before we do that, Naveen, can you just demo Monologue for us so people who are listening and watching can get a sense for the thing that you built?

Naveen Naidu

Yes, let me share my screen. So first I'll talk about writing an email, and we can get into Monologue's visuals and other stuff as well.

So let's say I want to write an email to Kieran. Kieran is a GM of Cora and he's a heavy user of Monologue. So Monologue sits on your laptop, on your screen here. And I'm using my hotkey, which is right side option, and I'm double tapping it, which is going to put me in hands-free mode.

“Hey Kieran, how is it going? Can we meet up tomorrow around 8:00 a.m.? Actually make it 9:00 AM. And if that doesn't work, maybe you can check my availability here. Thank you, Naveen."

So I just tapped the right side option key again. And here's the cool thing—the Kieran name, it understands I'm talking to Kieran. It got that right. And then here it fixed—I mentioned saying 8:00 a.m. but I later self-corrected and made it 9:00 a.m. And then here there's something interesting happening—how did this link get attached here?

That's because in Monologue we have instructions, where I mentioned my calendar link, so it just understands you really well. So you can just write with—this is just one example, but you can take this and use it when you're talking to customers, when you're talking to other use cases as well. Yeah, it's really cool.

Dan Shipper

So yeah, for people who are listening, basically all he did was talk into his computer mic, and what Monologue did is it transcribed it into a really well-formatted email that included his calendar link and took out the ums and ahs and self-corrected certain things that he corrected.

And it's honestly become the default way that pretty much all of us do work at Every very quickly, which is really cool. It's similar in a lot of ways to something like Wispr Flow or Superwhisper, which is something we can talk about as part of this. I think the thing that—the way that we've been talking about it, which I really love—is it types what you meant to say. And that whole thing has been really, really cool.

What I want to ask though is like, you have this really polished product. It's growing week over week. People are using it all the time. I think there's someone who used it like over 1,000 times last week or something like that. It's pretty crazy how good it is.

But you were actually in the wilderness for a long time, even though you were working with us for a while. You became an EIR—I think we had our first conversation about a year ago, almost exactly a year ago. And it took you from then until now to find this and have it be ready to launch. So what was that like? What was it like coming into Every, and what was it like kind of going through the wilderness and how did you get here?

Naveen Naidu

Yeah, it took me some time to get here. Maybe a quick intro—quick journey would be I quit my job around 2023, right after I think ChatGPT got released. So there was this feeling—you know, something big is happening in the industry, kind of a feeling. You can see ChatGPT got released in December 2022, and they released the API in March 2023. I vividly remember that because that's one of those moments for me. So yeah, I quit my job.

Brandon Gell

How did you get the conviction to quit your job in that moment?

Naveen Naidu

Not really. So what happened is in my previous company, I didn't get the conviction—I was really nervous. I didn't know what I was doing. So at my previous job, I had this deal with my boss saying, okay, I'm going to work three days a week with you. The rest of the four days, let me figure out what I want to do with my life or what I want to do next. And he was more than happy to give me—he was paying me full-time salary. He paid me full-time salary during that three-month period, but I still only worked three days and the rest of the four days I just kept thinking about what to do.

And only then I think ChatGPT API got released, right? So I started hacking products around the same time. I released this app called FridayGPT. That was my first time having some random stranger give me money to buy an app. So that made a big difference and I'm like, okay, I have to do this.

Dan Shipper

How did that feel?

Naveen Naidu

That was like—you know, I never met that guy. I don't even know who he is, right? But that's a really great feeling that, okay, I can earn something on my own just from this pure—the idea was just in my mind, I thought about it, I made it, executed it, and someone paid for it. That was eye-opening for me. And I thought, okay, I have to do this full-time, right?

So I just quit and then started my own company, came back to India actually—before I used to live in Japan. So I made pretty big different changes around 2023. After that, I started my company and I started working on multiple products. And one of the biggest mistakes I made during that time is I just worked on one product for six months or something and didn't even release it. So it's just sitting on my laptop. That's called Send Wisely.

Dan Shipper

What product was it again?

Naveen Naidu

It's an email marketing tool, similar to Kit.com.

Brandon Gell

I've never even shown this product that you built.

Naveen Naidu

I didn't even talk to anyone about it because I was really ashamed that I built it—just sitting every day, going to a coffee shop or something, building it. And I never showed anyone and didn't do anything at all with it.

I just kept on building, kept on adding features, and that made me—you know, technically I was able to just play around with tech stuff, but never really actually went into business. Okay, I have to sell it, I have to market this. I never thought about all that stuff. But later on, okay, whatever is done is done. Let me just start again from scratch.

So I again went back to the drawing board. I made a Mac app, and this time I had a goal—okay, I have to release this Mac app in one week. So I built it, I posted it on Reddit, and a lot of people had a similar kind of problem, and then they also wanted it. It's kind of like a wrapper—a ChatGPT wrapper. But the thing is you can select text and hit a hotkey, and you can run a prompt on top of it. So the idea is very simple.

A lot of people are going back and forth to ChatGPT doing the same thing, right? Maybe fix grammar or some other thing. So select text, hit—the funny thing is that's such a good idea.

Brandon Gell

There are—I mean, Edmar, a different EIR on the Every team, messaged me yesterday saying that there's a company that wants to talk to us. They do $30 million out of Brazil and it's just a ChatGPT wrapper. You can ask the same exact questions that you ask ChatGPT, but also in this tool, and they do $30 million ARR. So that idea very well could have been your idea. I know people bought it, so what caused you to sort of lose steam on it?

(00:10:00)

Naveen Naidu

Yeah, so it did a really great job for me, at least with the confidence. At one point I even reached $5,000 per month. That month that Mac app was making $5,000—that was during peak ChatGPT wrapper season.

So I was making a ton of money and getting good eyeballs on it. But one mistake I made is users have to bring their own API key. So it's a very niche product and it's a one-time fee. Users had to pay at that time $39 or $29—I don't exactly remember. So that's a niche product and after that, I don't know what to do.

So I started thinking, okay, let me build another Mac app so that I can bring in more revenue. So I built another Mac app called Screen Time Plus. And then slowly I started hacking a lot of side projects and stuff, but nothing really like, okay, this is going to be my next five years. It's just I'm making some random apps.

And the only thing that I got really good at, or compounding-wise, is I'm able to build apps really fast. And then I build that one single core feature and post it on Reddit, post it on X, get eyeballs, convert them to sales. But I don't really have any solid business plan of, okay, how can I take this to the next level? And that's when I met you guys and joined as EIR.

Dan Shipper

That's really interesting. You're saying so many things that feel very familiar from my own journey, and I'm just like, this is why I love working with you. This is why you're awesome. Just things that—lessons that I learned too along the way.

For example, one of the things you said is that you just negotiated with your boss to take four days—only work three days a week—rather than having to quit your job all at once. I think there's this sort of feeling that, oh, you have to do startups in a certain way and you have to be all in from the very beginning. But there's a lot to be said for putting your toe in the water and learning to ship apps while you still have a job, so you have more runway and less pressure.

Another thing you said that I really resonate with is—well, one is working on something for a long period of time and then not shipping it and then realizing that was bad, and then really forcing yourself to be like, okay, I'm going to ship something in one week. I'm going to build all these different apps. And I think that's a really good part of the learning experience—just like, go from nothing to an app that you're trying to sell, instead of being off in your cave hacking away on something that no one actually wants. Yeah, so I definitely vibe with that.

Naveen Naidu

Yeah, that was a big lightbulb turning-on moment for me, which is, oh, I can just do things. Just—I have a small idea, just code it up, release it. It doesn't have to be like, oh, a lot of features have to be there, or billing has to be there. I just figured it out on the way. So that was a big mental shift for me during that time.

Brandon Gell

I was thinking about how—I wonder if people, when they work on something that they haven't shared with anybody, a lot of times it's for a product that they themselves can't immediately use and don't actually use themselves. Versus if it's something that they actually use themselves, they're excited to share it with people. And that's kind of a good indicator.

And it's honestly the way that we build products—we won't build anything that we don't use all the time. It's just sort of a rule, an unwritten rule that we now have, because it'll never launch, I guess is what we're learning.

Dan Shipper

Yeah. And it's the best way to know that the product you're building is good, because we just—if we use it and then we get the rest of the team to use it, we're pretty sure that the audience is going to like it. And that's just a really cool loop.

And I think we're in this unique time where the game board has been reset because of AI. So there's all these new opportunities to do new stuff that you can just look for problems that you have personally. And if you have that problem, there's a good chance that other people have it and that it hasn't been solved yet, which is really, really cool.

That's kind of a good segue to how Monologue came about too, because Naveen, correct me if I'm wrong, but my understanding is we were launching TLDR—

Naveen Naidu

No. Okay, let me give you a quick timeline maybe, just to refresh. So I joined last year, October, so that's the last quarter of last year. And I started working on TLDR. It's kind of like synthetic podcast generation from your meetings, Slack, Discord. It just churns everything and converts into a beautiful podcast that you can listen to.

Dan Shipper

TLDR. Those were the days. Yep. I still get pitches for people to do that. I'm like, we tried it, bro. Doesn't work. We launched it right before Christmas on this crazy chaotic schedule where we launched like four things in a row. And it actually did fairly well, but there was—I don't know, it was... Anyway, interesting product. Yeah, maybe we positioned it the wrong way. I don't want to accuse anybody of positioning it the wrong way. But me and Brandon had a big fight before it came out, and he was—I'll say for the record—he was correct about how we should position it. I went a little galaxy brain, but... Anyway, it would've worked either way. Yeah. Sorry, finish what you were saying.

Naveen Naidu

So for me, from my side, what happened really is I think it didn't really sync well with our Every audience.

No one, to be honest—I think I worked on that for 10 weeks internally. No one really used it. No one was listening to those podcasts. No one was reading those meeting notes and stuff. And even when we released it, I think—we have this Every audience, right? So they were also like, okay, this is cool, but I'm not going to use this, obviously. Yeah, that was kind of the vibe we got.

So after that I took some time off, like two weeks. I didn't touch my laptop. I'm like, you know, I continuously worked 10 weeks on this product, and I did the same mistake like I previously did one year ago, which is just working on something which I didn't like. Put it out fast, really fast. So coming back after the holidays, I made myself—okay, I'm going to release one experiment per week. I even remember Brandon having a KPI as well for that quarter where experiments from studios—like number of experiments from studios—and that really helped me also.

Okay, I'm not going to worry about creating products and stuff. I'm going to just think about creating as many experiments as possible. So after that, I did Kairos. A lot of Every audience might remember Kairos. They love it. People love that.

Naveen Naidu

You have to actually sell the books within the app to make any kind of money. If not, how are you going to make money? People are not going to pay you $20 on top of already buying books. So that didn't work out, but people loved it. People loved the article. I still get emails and messages saying they like it and they kind of try to include that into their habits.

After that I started working on a Grammarly alternative. That's because I'm from India and—oh my God, I forgot all of these, you guys. You remember it now?

Dan Shipper

Oh yeah, now I do. It's all coming back. Yeah, I blacked that out from my brain, but—

Naveen Naidu

Yeah. So after that I did that experiment where—you know, my English is not that great when I'm writing, so I use Grammarly a lot, but Grammarly is really slow. It's shitty, to be honest. Someone has to make a better product here. So I asked around the team—Kieran is paying for it, Dan is paying for it, so I'm paying for it. Why not? Why can't I take this up?

So I took that up. I actually first started with an iOS app, so within one week I released the iOS app. We even emailed some early adopters. We put up a form. It's going well. But the energy of that app—the overall usage is good. It's going in the right direction. It's just that energy around this idea is not that great.

Dan Shipper

Okay. This is already—you are solving a 20-years-old thing. People are not really excited to—your energy or the user's energy. Which one?

Naveen Naidu

Everyone's energy, I would say. It didn't feel like a creative act. It just felt like we were solving very specific problems.

Dan Shipper

Yeah, I wasn't sure how we were going to make something mind-blowing there.

Naveen Naidu

Yeah. So that's the same thing, but I remember thinking this is such a good idea. Yeah, I think we released an iOS app. I think at one point we got to around 200 users using it, using the keyboard, actually changing their own default keyboard. And we also had a Mac app where it highlights the corrections and you can correct it. So technically we invested a lot of stuff in there, but it just didn't feel like people were adopting it that highly.

So during that time, what happened is I was pretty low—I wasn't motivated to work on that anymore. I just wanted some new, fresh idea. So on the side, I didn't tell anyone—on the side I just made Jole, which at that time was the working name of Monologue. So I just made that app for myself and I just started using it.

(00:20:00)

Dan Shipper

Why did you make it? What was the thing where you're like, oh, I just want to mess around with this and I won't tell anyone, but this is kind of fun?

Naveen Naidu

Yeah, when I joined as EIR, one of the ideas was a Whisper Flow alternative, but better. That's one of the things—it was on my list.

Dan Shipper

Too bad you didn't put that higher on the list.

Naveen Naidu

I don't think we were ready for it.

Dan Shipper

I have comments on this, but keep going with the timeline.

Naveen Naidu

Yeah. So what happened is I had that idea. It was itching me, you know? And another thing—Eder, one of the users of Underwrite, and he mentioned, I want this voice-to-text in keyboard. Can you do that? And I got a similar kind of request from other users and I also want that.

And another thing is one of the previous apps I talked about, EGPT, has that feature. I already knew users were actually using voice and how it was helping them a lot. So I talked with a couple of users and the thing is, it's not a great product. Two years ago that I built, it's not really a great product.

So I just thought, okay, let me make a really great product with all the things I learned because now I feel like I'm really confident with my technical skills, product skills. So I just threw away everything and started the code from scratch. I'm looking at my commit history. Are you interested in looking at the commit history?

Dan Shipper

Sure. Yeah, absolutely.

Naveen Naidu

Alright, let me—so officially it's called One Look Mac app now, but I started it in April I think.

Dan Shipper

So for people who are listening, basically Naveen is paging through his commit history. And what's really interesting about this is—okay, so you started working on this app in April. It's now September. So it really has not been a long time, and you are one person, and this is like a beautiful production-ready app that you've done everything for that hundreds or maybe thousands of people are using every day. It's pretty cool.

Naveen Naidu

Yeah, so I think it is just the compounding, right? If anyone out there who is looking to start or create a startup or create a product and stuff, one thing I would say is just keep on building, keep on creating as many experiments as possible because once you hit that one great idea, you can just release it in two days.

You can see that commit history, right? So describe what we're seeing on the commit history. Okay, so I started—my initial commit here is April 15th. And I did a bunch of changes, bunch of things. And on April 17th, I have v1.0.1. So in two days I released it.

Brandon Gell

And there's something that I just saw that's interesting in there. You have a commit in there for auto enter, which I think is one of people's favorite features of Monologue, where you can speak into Monologue and it automatically will process what you say, fix it, and then automatically—paste it and click enter, which if you are a developer, allows you to really stay in flow because you can just speak and you don't need to use your keyboard at all. I use it for Discord, I use it in messages, I use it in Warp. And that was—I think I saw that on the 11th or is that the 17th that you committed that?

Naveen Naidu

Yeah, 17th here.

Brandon Gell

Okay, 17th. So it's one of the most important differentiating features and it's one of the first things you built. To me, it just tells me you just need one killer feature to really convince yourself and see if people really want to use what you're building. And then a lot of the work that you've done is like polish and fixing bugs and adding things here or there, but a lot of the product was there the day that you launched it. So you can have a simple product that works super well. But that last 20 percent is so hard.

Naveen Naidu

A lot. Yeah, 100 percent. So at that time, what I'm doing is on the 15th I built that and I started using Jole at that time to build Jole. You know, I've been dogfooding. The auto enter thing came because—ah, why am I always hitting enter? Why can't it do that for me? So I just added that. So kind of all the features that first two days that you see is around that.

Dan Shipper

That's the beautiful thing about building for yourself—you know that that's what you want. And also it's the beautiful thing about building inside of a company like Every, because you've just been flooded with people saying like, I need this, or this isn't working, or whatever. And so you have this really, really high fidelity feedback loop all the time of good feedback.

Naveen Naidu

Yes, 100 percent. So after that, what happened? Like two days, I pushed it on the 18th. We have the show and tell on Fridays. So I think that's one of the fun meetings we usually have—we get high signal feedback there. So I did the demo at that time. Pretty, you know, not feeling that great about the progress I'm making with Underwrite.

Then at the end, like maybe five minutes or 10 minutes remaining, I demo this app that I've been working on for the past two days. And Kieran got really excited about this. I don't remember everyone else's reaction—I just remember Kieran's reaction because Kieran also has his own crappy version of it. He mentioned he uses his local thing, so he got excited. I got excited, so I just released the app. I think right there, I just sent him the build directly.

And the next feature he asked for—"I need local models because I like local AI models." So I just added that feature for Kieran. And slowly Nih started using it, Alex started using it, Brandon—like that. Everyone started using it, everyone started giving me feedback, and then it just—yeah, off to the races. That's it.

Dan Shipper

Yeah, this was really exciting time and I'm laughing to myself because I remember feeling—I'm going to go back to sort of why we felt like we were going to do this and this is like an easy decision, but I wanted to share a quick anecdote.

Brandon Gell

I remember we created a Discord channel and Kieran hit a bug where it wasn't working for him and he sent you a message that I honestly will never forget. He said immediate churn with a screenshot. Classic Kieran. Yeah, I love him. He's such a dick in the most loving way. And yeah, he really—he just said "immediate churn," and I just looked at that and I was like, Holy shit, he's right. This is immediate churn. And just saying it so bluntly, I think—and that's sort of a version of the feedback that you get from everybody here I think has made it like a pretty bulletproof product. But yeah, I'll never forget "immediate churn." That's funny. I want it on a hat.

Naveen Naidu

Yeah, go for it. I mean, no, I take Kieran's feedback really seriously, because I think he's like the top 1 percent of users where, you know, if they hit a bug, you get immediate churn. If I keep Kieran happy, I'm going to keep a ton of people happy.

Yeah, I just submitted a PR on Cora, which he denied. So I know the bar for Kieran is pretty high. He was like, this is vibe coded.

Dan Shipper

You're right. Which is crazy because he is the number one vibe coder. He's just wielded it so well that now he thinks he's above vibe coding.

I think this is an interesting product for me too, because it is the first product that we've incubated inside of Every that I have not either built the first version of or been intimately involved in the building of. It was the first product where I was just off doing other stuff and I just saw a bunch of people talking on Discord about this thing that at the time was called Jole, and then we renamed it Monologue.

And I was like, I guess I'll try it. And I tried it and I was like, Whoa, this is good. And that was the coolest feeling ever, I have to say, to just feel like we're making awesome stuff and I just had nothing to do with it. So it was just really fun to watch that happen.

I want to talk real quick about the strategy behind it. So there was obviously—Naveen, you built the first version of it and I remember in that show and tell I had sort of an unlock and was like, We need to do this.

(00:30:00)

Brandon Gell

But the reason that I felt like we needed to do it, even though it felt kind of wrong because there were other companies that already existed that were doing this—and typically all the products that we've built have been net new products—it was very hard to get over the idea that we weren't going to build something new. We were just going to try to take a concept that already existed and make it better.

But the reason that I got so comfortable with that and actually reached out to Dan and was like, dude, we have to do this immediately—and you were like, maybe, and then you used it and you were like, I'm totally in—was I realized that people subscribe to Every because they want to be super up to speed on AI and they want to use our individual products. But what we want for them is that it's their last subscription that they have outside of a ChatGPT or Claude. So they have a ChatGPT subscription and an Every subscription.

And I realized voice—very, very good voice dictation—is one of the main tools that people are using, and we're actually doing subscribers a disservice if we don't include this as a part of their bundle. We just need to figure out a way to do it that feels right to us and sort of levels up the whole industry.

Which, you know, you started this in April, we're recording this September 10th. I think we've done that over the past many months. It's taken a long time to get it to be really, really good.

Dan Shipper

Yeah. I think that was my hesitation too, is like I wanted to only release this if we felt like it felt like an Every product. It felt like ours instead of just, you know, we just vibe coded something and it's kind of a shitty thing. And I really do feel that way. You're an absolute craftsman, Naveen. It's a beautiful product.

Naveen Naidu

Thank you. Really appreciate that. So yeah, I too had that in mind when even when I was presenting that, I remember—I know this is like already an existing product, like I just made it for myself, like don't—you know, this is my baby. Don't be harsh, kind of when I'm presenting it.

But once like Brandon talked about this—don't worry about competing with other apps, because we already have a ton of Every audience. If no one else is subscribing, you are going to anyway get these Every subscribers who are going to use it. So just focus on Every subscribers. Don't worry too much about external—like you being competing with others in that initial period.

So that actually made me a little bit—you know, okay, I can actually make this work because I can just focus on Every audience. Let's try to make a really good app. So that put me at ease because I don't want people to think like I'm just copying something.

Dan Shipper

Totally. And this is the first time that I felt like the bundle strategy that we've been pushing for a long time is starting to really work. So for people who don't know, with Every, you pay one price, you get access to all the writing that we do and all the software that we make.

And it makes so much more sense to build Monologue in the context of a bundle because, you know, when you're comparing—okay, I can buy one app or I can buy a subscription and get a bunch of apps plus really good writing about AI—that value calculus makes a lot more sense. And it's just been really interesting to watch that happen, because I didn't even think that we would be building something like Monologue as part of this strategy, but it just seems to work.

Naveen Naidu

Yeah. And that strategy really helped the way Monologue is right now, because what happened is April is done. Maybe I can show the initial periods because of the uncertainty—

Dan Shipper

Yeah, sure. Show us the usage.

Naveen Naidu

So the usage here is that I keep track of two things, which is the number of words that are getting dictated and number of times that Monologue got activated. So initially, I just walked you through that initial period, right? 17th, 18th, like April, and then I gave access. I started adding analytics the following week because people had been using it a lot internally.

And here you can see this is not my usage—I removed my usage because I do a bunch of testing, right? Initial period. So this is somewhere like we got 130 users per day. Slowly it went to 150, 200, like 300, 350.

Dan Shipper

And this is the number of times that somebody clicked the button to start dictating?

Naveen Naidu

Yes.

Dan Shipper

Okay, cool. Yeah. And for people who are listening, it's just a smoothly going up graph. And this is a usage graph, which is very hard to do.

I mean obviously there's jags—it's a little jagged for weekends and stuff—but I think it's really important to say, we work on a lot of products. We've all built a lot of products over our lives and seeing something that just smoothly goes up in usage is extremely rare.

Naveen Naidu

Yeah. Here you can see the week over week graph, which is like—initial week we got 230, but slowly end of May, we are hitting around 1,500 users per week.

Dan Shipper

Usages per week?

Naveen Naidu

Yes. Usage.

Dan Shipper

Okay. Got it, got it. Yeah, I think we are hitting around 2,700. These are all internal users, I assume?

Naveen Naidu

Yeah. This is all Every internal. If you want, maybe we might have to blur out the emails here. This is only to June, right?

Dan Shipper

Show us the full graph.

Naveen Naidu

Full graph. Okay. We might have to blur out emails.

Dan Shipper

Oh wow. So look at that. This is—look at this. This is even better. Yeah. We're hitting 7,000 usages per day. I mean, this is crazy. It's per day. This isn't even linear. There's no landing page for this product. There's no—Naveen, show us the landing page.

There's the landing page. So for people who are listening, it literally just says Monologue and then download and then it says, please don't use this because it's a public beta. And it has a little Made by Every sticker at the bottom. So that's crazy. And it makes me very confident in this product.

And people are—okay, so the other thing that's crazy too is there are a bunch of competitors. A lot of them have raised millions of dollars. Like we have a competitor Wispr Flow that's raised like $50 million. We have invested in this product less than $20,000. And one person—Naveen is building this product alone.

Brandon Gell

Yeah. We should talk a little bit about how we have supported you with a bunch of other people to really make this super artful, but I just think that that's crazy and it's one of the most obvious proof points to me that we just live in a different world and raising $50 million to do something is not only unnecessary—it puts you in such a corner.

We have already built something that is successful. Period. Because we can build really quick. Naveen's really talented and we haven't invested a stupid amount of money into it, and it doesn't need to be a billion dollar outcome. It could be like—I'm at this point, this will be a multimillion dollar outcome. And I feel like we have that in the bag.

And that's really special to build from that place because you're building from your front foot instead of trying to dig yourself out of a hole. Totally. It means we get to make art, which is worth talking about because this is a very, very artful kind of teenage engineering style product.

Dan Shipper

Can you demo the welcome screen? Do you have that set up?

Naveen Naidu

Yes.

Dan Shipper

I know Naveen is so proud of the onboarding because he has a button that's easily accessible to reset the onboarding. Oh, I just want people to—yeah, just go through the onboarding again. So just to set some context around—

Naveen Naidu

Okay. So what happened is, I think end of June, Brandon and I sat down in one of our calls and were like, okay, this is working. People are actually using it, but UI looks shit. UX is shit. We are just not happy with the way things are. So we immediately brought in Lucas. So Lucas is our head of—creative lead.

So I think having Lucas on a project like this just from the start made a ton of changes—like, you know, whatever additions that you are seeing within the product. So Lucas brought that Every synergy into Monologue from the start. That's why the way it's working right now. So let me show you quick things.

Brandon Gell

I remember when Naveen and I had this one-on-one, I said, This app needs to do three things. It needs to work really well. It needs to look like teenage engineering or something like that. For people who don’t know, Teenage Engineering is a hardware company that makes beautiful recording devices, music making devices, computer cases—they just have a very cool aesthetic. It needs to look kind of like teenage engineering. It needs to look different and it needs to be a part of the Every bundle. And if we do those three things, this product will work. And that seems to be happening right now. And the way that you execute it on the design side is just out of control to me.

Naveen Naidu

100 percent. This is one of our inspirations.

And Lucas worked—he gave us this monophone idea. We never thought of monophone at all. Initially I remember Jole being just a blob of text, just radiating. And Lucas came up with this monophone idea where you have to feel like you're talking to something, so that's—yeah, show you.

Awesome. Yeah, so this is the onboarding where you see the sound and animation. Maybe Brandon can talk more about the sound and everything that we—

Brandon Gell

Yeah, I just gotta say I love it and we just have so many people reach out being like, "That was awesome," just from downloading the app and starting it.

Dan Shipper

I had someone who has a fairly big Twitter audience message me being like, this is the first time I've gotten emotional using software in many years. And I was like, fuck yeah, we did it. We did it. I think—I don't know, we just had such a strong perspective of like, we want the—well, we didn't even know. Lucas actually came up with the idea, the word for monophone.

(00:40:00)

Brandon Gell

So he was like—this monophone design was really special. We knew we wanted sound to be really cool. He also came up with the name Monologue. Yeah, he did. He did come up with the name too. We wanted sounds to be a part of it. So we have this really cool intro sound that my friend Jake Sheriff made.

We brought in a designer to do the app design and the landing page design, which is the first time we've ever done that. We've always done app and landing page design internally. This time we brought somebody else from the outside and he's great, and he is amazing. Lucas did creative direction, he did actual design and it was like a very special combination. And then we brought on a guy to do animations, so it was just like—

Also in the middle of this process, I learned that Naveen is like top 0.0002 percent in India at math and science for whatever cohort you are of that test. And I learned in this process, he's also very good at managing other people, which is a very rare combination.

Naveen Naidu

Thank you. I don't know what to say, so I think it's just like—I feel like I'm in a video game, you know, collecting all these skills so that I can build this Monologue. So I think in my previous job I was managing people. So yeah, I collected that skill and later, when I'm working—it just like everything came together nicely here.

Dan Shipper

For people who are watching, I think there's probably a lot of people who are like, "Wow, this is awesome. I want to start doing this too." Can you tell us a little bit about your workflow, what tools are you using, how are you using AI in your workflow? All that kind of stuff.

Naveen Naidu

Yes. So for the initial stages—that zero to one—it's really, really important for you to just keep it to one single feature and use AI obviously. And one of the key things I used initially at that time—I remember Windsurf, it just feels like ages ago. So I was using—what is that again? Yeah, it even feels like a year ago when I've said Windsurf. But I was using Windsurf a lot during that time, and I think Windsurf is one of the first IDEs to introduce this agentic coding as well, I remember. So my codebase has been getting a lot from Windsurf.

And right now my tech stack looks—Claude Code a lot. Most of the code that's in Monologue is done by Claude Code. And then recently I've been using Codex a lot. So let me show you Codex.

Dan Shipper

Yeah, show us. We've never actually shown Codex on this show, except for—we've shown the web version because we had the guy who launched it inside of OpenAI on the show for the launch. But we have not gone there in a while. So yeah, tell us about how you're using Codex.

Naveen Naidu

Yeah. Let me show you Codex. I think around May time, May 18th, they released it and I initially got some errors, but on the 18th, you can see how many tasks I'm giving to Codex. Basically every day you have a task—

Dan Shipper

You're just scrolling through a list of tasks for people who are listening and yeah, every day you have a little task that you're kicking off to Codex or maybe looks like multiple tasks a day.

Naveen Naidu

A lot. There's a lot. Yeah. It all depends on my mood as well. If I'm in that mood of—okay, I will be in manager mode. And let me just give a bunch of tasks to all these AIs and review it. But if I'm in active development mode, I'm going to just use Codex. Let me think—

Dan Shipper

But now it looks like you skipped a month. So why did you not use it for a month?

Naveen Naidu

Claude Code made a big difference for me. I think now it's—yeah. I think what happened is I was using this with—I don't remember the models. I think o1. I think it's o1. It's a fine-tuned—

Dan Shipper

o1, yeah. Yeah, o1. Yes.

Naveen Naidu

I think during that time it was really great. And after Claude Opus 4.1 got released and I just completely moved to Claude Code and now I'm back to again Codex CLI. So actually—yeah, here you can see it. I'm not—you can see I'm using Codex actually. Today I just developed this—I'm working on this Monologue speed test website and I just made a plan and within the plan we have these phases—phase one, phase two, phase three—and I'm just implementing these phases.

Dan Shipper

Phase three. Interesting. Yeah, so I'm just using Codex a lot where GPT-5, or the high-end medium, is so good. Interesting. Yeah. Tell us about why—why'd you switch? Why do you think it's better?

Naveen Naidu

One, I think 4.1—it just like, I think I got like, with Claude Code, it just says like, "You're absolutely correct. Absolutely right." And it doesn't push back on my—whenever I'm trying to get some proper feedback on my implementation. For example, I actually used Monologue to dictate this here.

Dan Shipper

And he's just highlighting a bunch of text where he's asking questions about his codebase.

Naveen Naidu

Yeah. And it gave me a really great plan. I think planning is really, really good compared to Claude Code. Here what happens with—yeah. Planning and even with the code itself, it's—actually I have a video as well recorded where I compared Claude Code versus Codex, or GPT high medium.

One thing I clearly see is Codex works on what's asked. It doesn't go into some like its own creation. It's much more precise. Yeah, much more precise. And the code it generates feels like some senior engineer wrote it instead of Claude Code. It kind of feels like some junior engineer wrote it. I know it's a big statement, but it just feels like with tons of comments with tons of—the logic it types, it's much more like not well thought out. But if you look at Codex code—

Dan Shipper

Yeah, yeah, I think you're right. GPT-5 code does feel like if I want a precise change and I want to be confident that it's doing the right thing, I can use GPT-5 and it'll do it well. And so it's got that precision thing. I think of Claude Code as being high in industriousness. So it's just going to do a bunch of stuff, which is actually helpful in certain circumstances.

So, for example, I wanted to commit some code to Cora. And so I needed to get my repo running locally. And Claude Code is just really good for that because it's just going to keep banging its head against the wall until it figures it out. It's very industrious. And GPT-5 will be like, well, I can't do this because I'm—you know, whatever. There's like 15 precise things that it thinks are wrong, you know, and then you have to go through each one with it.

And yeah, so I've been liking Codex as well. I'm curious—I've heard that a lot of these models are not that good at writing Swift. So what has been your experience?

Naveen Naidu

Even during Underwrite development, which is the Grammarly alternative that I was working on, it's technically complex, meaning you have to show these red lines on top of any app. So that's a very highly technical challenge. You can't just use some external library to build that out. So during that time, I was using o1 a lot. o1 is really, really good. o1 and o1-mini, I remember being really, really good at writing Swift code and writing highly technical, complex apps. So I always have a soft corner for o1 series models—the thinking models—for Swift because it has that extra intelligence when compared to any other model on the market. So I use that a lot.

And so if you compare—Claude Code is really good at implementing the features. You can just give it work and it does that, and it's good at Swift. Actually, Claude 4.5 is really good at Swift. I use it like you can see that.

But Codex—GPT-5 high—I recently was experiencing hotkey work, which is very complicated work because you have to go into niche CG event, which is written by Apple in the 1980s. That's code that's still running on Mac, right? And it has very, very bare minimum documentation, but GPT-5 high is so good at ripping through all this complex code. So I think that's one of the biggest advantages for me.

So I think that's one of the biggest advantages for me. Why I didn't hire or get help from other developers. It's just so crazy thinking about like before this stuff existed, solving that problem would be—you might hire somebody for like a year to solve that. It'd be like a $200,000 problem.

(00:50:00)

Brandon Gell

Yeah. And now it's just not—it's so, it's kind of, to me, it almost makes me sad to say because there's been so much blood and sweat expelled to solve these problems. And now it's like we live in a world where we're like, nah, that's not a problem.

Naveen Naidu

Yeah. Yeah. Before, I used to be very scared of these kind of technical problems. Now I'm like, okay, I have GPT-5. I don't have to worry.

Brandon Gell

You got your partner?

Naveen Naidu

Yeah, I got my partner. I have Claude 4.5. Those are my team members. I'm going to give those tasks to them. So I think that gave me a really big confidence boost. Okay, I can do anything I want.

Dan Shipper

I love that. All right, so we're about to launch Monologue 1.0. This is out—it may actually already be launched. What are your hopes for it? Where do you hope to be in a year? What would be a success for you?

Naveen Naidu

For me, our internal goal is in one year we want to hit $1 million ARR for Monologue. Woo. It's—by the way, Every as a whole just hit $1 million or a little bit more than $1 million recently. So that's a big goal, but I honestly think we can do it.

Dan Shipper

I definitely think we can do it.

Naveen Naidu

Yeah, so that's the biggest goal—to hit $1 million ARR. But that's a big goal in one year if we do it. That would be crazy. But at least for the next quarter for me is release the iOS app. That's one thing we are getting a lot of requests around and I want to release that iOS app. It's also technically challenging because Apple doesn't make it easy for us developers, so we have to somehow find loopholes on iOS. And then end of this year, reach $10,000 ARR. So that's my goal.

Brandon Gell

Great. I love it. Oh, you're definitely going to do that. Yeah, definitely. I'm looking at the usage graph right now—I'm looking at the numbers. I think you're going to do that and more than that.

Naveen Naidu

Yeah. I think one of the really cool things is Every's audience just amplifies it a lot, so yeah, it's crazy. And I just love the way we are building because I think it's distribution first. So Dan did really hard work for the last seven years and I'm just coming in and—

Dan Shipper

Five and a half years. Yeah, it's paying off.

Naveen Naidu

Paying off now.

Dan Shipper

Well, Naveen, I'm psyched for you. I feel so honored that we get to work with you and that we get to launch Monologue. I think it's going to be amazing and I can't wait to see what comes next.

Naveen Naidu

Thank you. Thank you. Same here. I'm just—every day waking up, I'm just grateful that I have this team and I have AI with me, and I can do anything. Right now I'm in that zone where if a user or customer has some kind of problem, we can solve it. So I love it.

Dan Shipper

What a way to end. Yeah, absolutely. I can do anything. Awesome, man. Awesome.

Brandon Gell

Yeah. See you.

Thanks to Scott Nover for editorial support.

Dan Shipper is the cofounder and CEO of Every, where he writes the Chain of Thought column and hosts the podcast AI & I. You can follow him on X at @danshipper and on LinkedIn, and Every on X at @every and on LinkedIn.

We build AI tools for readers like you. Write brilliantly with Spiral. Organize files automatically with Sparkle. Deliver yourself from email with Cora. Dictate effortlessly with Monologue.

We also do AI training, adoption, and innovation for companies. Work with us to bring AI into your organization.

Get paid for sharing Every with your friends. Join our referral program.