Best of the Pod: Would You Shut Down Your Most Successful Product? The Arc to Dia Story

[00:00] Last year sucked. We were running a company with... [00:03] 70 people, millions of people using the product every month. Nope, we're starting over. Oh, cool. What are we doing? We're doing it. [00:10] We don't really know. We have this realization that all of software needs to be rewritten for this new world because primary interfaces we have with computers are going to change. What feels like your computer... [00:19] in five or 10 years is actually going to be this layer that sits across all of your devices. [00:24] this personal intelligence layer that is a wild invention of humanity that's gonna help you do all sorts of things, and it's gonna be miraculous, but don't forget, [00:30] The browser is just the enabling technology underneath. [00:34] *music* [00:48] Josh Hirsch, welcome to the show. [00:51] Good to be here. Thanks for having us. [00:53] Of course, so good to have you. So for people who don't know, you are the co-founders of the browser company. [00:59] um you are the makers of arc and now dia um and i'm psyched to talk to you about dia talk to you about the journey to get there um as as you can see if you're watching the show uh me and hirsch are together we're we're in upstate new york in a little in a little uh cabin we've been friends for a long time and i think one thing that people should know coming into this is like we have been uh sort of like on parallel journeys together as you've started the browser company um [01:29] I started every around the same time as you guys started. And we've been close for a long time. So it's been really fun to get to watch the journey from afar. And yeah, I'm just really excited to get to talk to you about it. So thanks for coming on. Well, I'm just going to say for the record, I'm not there because I didn't get the invite. So we're not that close, Dan, I will say. I'm kidding. No, it's awesome to be here. I didn't organize. You got to blame Paulina Hirsch's wife.

[01:59] yeah, honored to be here. Excited to do this. We've listened to so many of your podcasts and obviously you have so much space. Yeah. Yeah. So, um, another, a couple other disclaimers, um, [02:11] I'm a small investor in the browser company. So people should know that. I also spoke at your wedding. So we'll put up a little picture of me speaking there. [02:22] And another really interesting little bit of history that is fun for me to remember now is I actually, there was a point at which... [02:32] I was considering or you guys were considering me to be the CEO of the browser company, which is crazy to think about that now. Because, you know, you would come up with the idea to incubate. At the time, it was superhuman for browsers was the pitch, as I remember. And I was working on super organizers, which is a little newsletter, which would become every website. [02:53] Oh, did that turn into every? I didn't even realize that. Wow. Yeah. And so there was a moment in time where we were discussing working together on this. So it's really fun to come full circle five and a half years later and be like, wow, we're all so much older. We have so many more wrinkles. I have a big beard. I think I was pretty clean shaven when we were doing that. So yeah, that was a crazy moment. [03:23] Wow, so super organizers is like your arc. [03:27] Yeah. Yeah. Yeah. Well, I hope Dia goes as well as every, I've like, fingers crossed, like I'm into it. Thank you. I think we all have a lot to learn from each other. Um,

[03:39] So I think the first thing I want to talk about is I watched this pivot from... [03:47] Arc to Dia. [03:48] I think we were in Thailand together when we were first starting to get like really psyched about AI. And I just watched Hershey's eyes just like, [03:58] He had this like spark. It was like a sparkle. It was like, oh my God, this is crazy. And I was like, yes. I was like the little devil on his shoulder being like, yes, you get it. Oh, it's your fault, Dan. I always wondered where that came from. [04:14] And what's been really interesting to me to see is... [04:20] I've obviously been watching this journey from ARK to DIA and watched the decision happen to be like, okay, we're not going to do ARK anymore. And then watched the public reaction, which was so negative. I think people just loved ARK so much. And then watched you guys have to like [04:40] bear that while you were [04:43] building dia and i just want to know like what was that like [04:47] Thank you. [04:48] Um, [04:50] Honestly, it was a little unexpected. I don't think we had predicted the reaction that we got when we announced that, especially because we had been in this headspace for quite a while. You know, we'd been trying to all throughout 2023 or so trying to figure out how do we get from early adopters to the mass market with ARK? [05:07] And ARK had this problem, which we call the novelty tax, where there's so much new stuff in it, which really attracted early adopters, but made a lot of the mass market more hesitant to try ARK because it just took a lot to get to learn and use. And so, I mean, Josh can probably say more. We tried so many things to figure out, OK, how do we make ARK sort of approachable to the mass market? And then simultaneously, AI had been taking off. And I think Josh has this video last year of Act 2.

[05:37] realization that actually browsers are going to change pretty dramatically and how we use computers are going to change pretty dramatically from a user computer interface point of view you know we're going to talk to our computers they're going to do stuff because this new play-doh allows you to like speak in english and they can use tools and suddenly computers are something completely different and so uh we went through our own evolution over many months to be like how do we nudge arc in that direction and for a while we called it arc 2.0 um and we were like okay how do [06:07] we loved ARK so much that we were hesitant to change it. You know, like the internal discussions about like, what do we do with the sidebar was like at the wrong level, you know, when we have this new Play-Doh. We really had to start from scratch. And so maybe, you know, I'm curious for your take, Josh, actually, on like how do you feel about the external messaging? Because for us it was... [06:28] over such a long period of time. And so obvious, you know, that we had to, okay, we have to start from scratch if we're going to build this thing from the ground up to use AI and to be accessible to the mass market. But I think it was probably a shock to the external folks that, oh my God, we're building a whole new browser. Yeah. I mean, I think, I think the thing I would say is you're catching us two weeks after we really released Dia to this private beta. And I'd say in terms of Hersch and my hopes and expectations, I don't think the first two weeks could have gone better. [06:58] so proud. And so I think I felt an adrenaline, [07:02] release. [07:04] And I think it's safe to say, [07:07] Last year sucked. Like it, it really did. And I think it's one of those things, like many things that I think are most fulfilling and rewarding in life that if you knew what you knew now, would you have done it?

[07:19] And yet at the same time, in retrospect, it so obviously was the right call. But I think a thing for especially listeners to know, and you alluded to this, Hirsch and I, I've never made software really without Hirsch. [07:31] I met Hirsch when I was 20 years old. We both left university, barely knowing each other to start a company together. I think flash forward over a decade now, I think it's a big thing. [07:41] If I'm being really honest, I think it felt so obvious to me and Hersh. Intellectually, emotionally... [07:49] The thing I think we both... [07:52] underappreciated was it's not just me and him anymore. Right. I mean, I mean, we knew that. But at every moment that Hirsch and I looked across the proverbial table and said, hey, [08:03] Should we do this? Should we make this thing? [08:05] It was me and him. [08:07] Right. And we share value. Like we it's just a different relationship. We were running a company with. [08:13] 70 people, you know, millions of people using the product every month. And we definitely, I think, underappreciated. [08:22] The public reaction. I mean, when we started this company, Dan, you know, when we when we were trying to recruit you to be the CEO, one of the questions like how to get anyone to care about their desktop web browser. In fact, I've never even thought about my desktop web browser before, let alone have an opinion on it was the what we saw the most. So I think we were very surprised by the reaction for sure. And when it does some things differently. But then maybe more importantly, definitely more importantly, is our team. [08:47] Showing up to a company where at that point in the company's evolution, people came because they loved ARK themselves. They joined for ARK more than the browser company. Like, hey, that thing we poured our hearts into for years. Now we're starting over. Oh, cool. What are we doing? We don't really know. But it'll be great. We're just going to go that way. Right. Sounds like a great plan. Right. You know, like so. Oh.

[09:09] It was a winding journey, but very proud at this point. That's, yeah. Okay, there's a lot here. I really want to go into... Tell me about the... [09:19] initial moment where you guys were coming to the decision of like, [09:24] oh wow, there's something totally new here. And not only is ARC maybe not working in the way that we want it to, but there's something new that we need to go play with that we can't even really describe [09:35] what it is? Like, what was that like? How did that whole thing come together? Yeah. And one of the things I think I want to be more forthright about here, I don't think we've really said publicly before, is that, [09:46] I almost think blaming Ark is a scapegoat, [09:50] Right, if really Hirsch originally, and then myself, didn't get so inspired by this new material that are LLMs, [09:59] We would not have pivoted Ark to something else. You know, it's easy in retrospect, but I was growing literally the novelty tax. I 100% agree with her that that is true. But our first instinct was actually just to make Ark better. So I really think the origin story here was that like, [10:16] you can't stop thinking about it. You're staying up till 3 a.m. Every bone in your body is just like, this is why we got into software. But I really got to credit Hirsch. I think I have this part of me that is very... [10:29] anti-Silicon Valley hype. [10:31] which honestly, for people that follow me on Twitter or something, they're probably like, I think of you as like a hypey dude. And yeah, let's have a therapy session later about that. But for me, I just, the crypto wave was the one that came before. And just for me, it just, it felt icky to me, even though I agreed with idealism. So here comes around AI, and my knee-jerk reaction is like, this is a bunch of hot air. But really, like when I think about, I didn't, not thinking about this interview before we came in, you know, when you just said

[11:01] I think of a conversation I had with Hirsch in San Francisco that to me was like, oh, we got to do this thing. So, I mean, Hirsch, when you really were first to see this world that was coming and feel it in your bones, like, do you remember even for you where it came from and when? He's giving me a lot of credit. I think it was probably, I came at it from my own perspective. I think realizing that. [11:23] Part of it was excitement about the Play-Doh, but part of it also was the realization that if we don't go this route, I think we're going to get obviated. All of software needs to be rewritten for this new world because, again, the primary interfaces we have with computers are going to change. But I think a lot of it also came from your experimentation with ArcSearch, right? And actually playing with LMs and realizing, oh, we can build something cool here. This is Play-Doh we're really good at using. [11:53] that knew about it. And the number one request for Arc was, I need a mobile app. But the browser is very different on mobile than it is on desktop. So we said, OK, let's just build a companion app. You can have your spaces and your tabs on your phone. [12:08] And, you know, in the backdrop of launching this kind of Arc mobile app is when all these kind of AI tools started popping up. And as much as we wanted or I wanted to resist the hype, I'm an intellectually curious guy that loves new technology. And so there was some part of me, especially driven by Hirsch, that was like, [12:26] okay, but if you really just tried to make the best kind of AI browser experience you could, what would it look like? But it would have been a huge distraction for the team to do that on

[12:38] desktop. So we actually hired a contractor externally. I was living in Paris at the time and kind of hired him to the Skunk Works project. He's like, hey, just for funsies, it's not going to be a big deal. What if we build a mobile browser that only does one thing, [12:55] Again, we had a little bit of the PTSD from ARK already. It only does one thing really well, and it has something to do with AI at its core just to prototype and learn. And so the idea was, what if in this mobile browser, [13:11] Instead of pulling back links, [13:13] We made you the perfect webpage. [13:16] So instead of typing in a query and trying to find the right link from the World Wide Web that is the closest approximation to what you're asking for, but not exactly what you're asking for, let's understand the intent of your query and just make a web page on the fly for that thing. Just a very simple idea. And it was so not a big deal or meant to be a big deal. [13:35] I tweeted it on a Sunday. [13:38] before boarding a flight, which for our company is not how you launch new products. And it just exploded more than anything we've made so far. And so, you know, we had a number of takeaways from that, but one of them that really influenced DIA was what Hirsch is referencing as the novelty budget. [13:53] Keep it extremely simple. [13:55] extremely focused, [13:57] change one thing, [13:59] And have that one thing be the thing you talk about, which I realize in retrospect is like, yeah, duh, welcome to like making consumer products for real human beings. But is the antithesis of ARK, as you know, in terms of the product philosophy.

[14:11] I think Arc Search also gave us sort of a strategic realization, which is if you think about what a browser is, it has sort of a desktop browser, especially it has two components. One is it's sort of a funnel for intent. You know, the Omni box is where you type in all your search queries. And if you have a certain intent, it goes to search and then you can go find that thing. [14:30] And secondly, it's an application platform. And so you run all your web apps and it's sort of the Web 1.0 and 2.0 sort of portions of what a browser does. [14:41] The intent portion is going to be ripped [14:43] drastically changed by AI because a lot of our intent is going to go to these models that get a bunch of data and then spit out the actual answer. And because AI can use tools and the browser has access to all your apps, it can also really support the application component. And so I think ArcSearch was our first foray into realizing [15:08] browsers are going to change, you know, and so we need to rethink what our strategy is because the entire ecosystem has changed and the browser's place in that ecosystem is going to change drastically. And so that was sort of our first moment of like, oh, we got to rethink, you know, what this product is. And I think the bit about search is really key here that came to influence Diaz, you know, keep in mind, our kind of approach to building Arc was really as much about [15:33] you know, urban planning and interior design as software. And our recognition was you spend hours a day in this rectangle in the space and people don't have any feelings about it. Could we change that? And so the sort of conversations we had is like, okay, can we enumerate where those minutes and hours are spent? And arguably the core action in a browser is,

[15:53] is and was search. [15:55] But up until Arc Search, we sort of said, you can't touch that. You know, we're not Google. Even though Search, opening a new tool, I mean, the Command-T text box is the most popular text box on your desktop computer, according to Apple, and I'm sure Microsoft as well. But we kind of said, no, we're not a search company. We couldn't possibly do anything with that. You know, that's Google's domain. Look, even Microsoft threw tens of billions over many years, and Bing sucks. So what are we going to do? [16:24] Oh, wait, as Hirsch said, this is this is the choke point for the Internet. This little box routes you to places. And does that mean we can now route you to new places and new things? So that was also the big like eye opening moment from ArcSearch, which was not the intention at all at the time we started the project. [16:42] So take me from that moment from like, okay, you make some realizations around ArcSearch to, you know, [16:50] Um, [16:52] Probably, okay, we're going to do Arc 2.0. [16:54] But actually, maybe this is a totally new product. Like what happened in between there? [16:59] I've never said this before. I think indecision and a lack of, you know, excellent leadership on at least I'll say my part in that I think Hirsch and I knew deep down in that moment. [17:10] what we had to do, whether it was from the Arc perspective of like Arc Search feels so simple and clean and resonant with people. I mean, that was the first time we had people telling us like, hey, I really don't like Arc on desktop. Sorry, I just like too much, don't get it, but I love Arc Search. So I think the combination of Hirsch and then my kind of increasing conviction that this was going to change all of software, as bombastic as that sounded, combined with the

[17:40] hearts knew what we had to do. But what happened next was, [17:45] That was in probably February of last year. And it was in June at a company-wide offsite that we said, we're going to build something new. But then even then, it was like, it's going to be ARK 2.0. Like, we don't even know. Like, TBD. So I think it probably wasn't until September-ish of last year that we made the call to say, it is a completely new product called something completely different with no connectivity to ARK. [18:15] especially in this moment. But I'd say that period is the one that I regret the most in terms of just like not calling a spade a spade and just like ripping the cord, you know? And to our point earlier, especially with a team that large and a team that joined post-Problemar Confit, you know, like getting the team's buy-in and aligning with everybody and figuring out what are we doing. And it took a little while. Dan, one thing I'll add too is like, you know, I love every one of my favorite publications. Not an investor, wish I was an investor, but it's fantastic. [18:45] is remarkable. Actually, even reading your coverage is, [18:49] Keep in mind where AI was. [18:53] 14 months ago. You know, it was, you know, ChatGBT was, you know, long in the market. And I think for most people paying attention, you could kind of pull the curve forward. But there was also a lot of technology macro risk at the time as well, which is like, are the scaling laws really going to continue? Like, are we going to figure out hallucinations? You know, even this concept of memory, which is core to what we're doing. I mean, I can't remember the timestamps, but I don't think

[19:23] the most central idea behind DIA. And so, yeah, it's easy in retrospect, I think, to be a little like, you know, self-deprecating. And it was not obvious to everybody or anyone. And I do remember we put out a video that was like, we think AI browsers are the future. And everyone's like, what are you talking about? Like, you're being a hype boy. What are you talking about? And, you know, literally today there was a, you know, MG Sigler last night, right? This whole article that's like, yeah, AI browsers are definitely going to be a thing and everyone's definitely [19:53] AI is going to make one. And so I don't say that to gloat. It's actually, you know, competition that we have to wrestle with. But in the moment in which I'm like, oh, the indecision, there were so many reasons why you would not do what we did. So it was not that clear. I totally, I totally remember all of those things, like how... [20:11] if you're really using those tools every day, you're like, wow, there's something here. But if you weren't, you were like, this stuff is like not good. No serious person could ever use this for their work. And that has completely shifted, at least for a lot of people over the last 18 months. I would even say six months, Dan. One of the things that I remember in the moment was, again, it's almost like, it sounds like a love story or a relationship, like me and Hirsch in quiet moments together, [20:41] conversations we would have about this. Handle it, you know. Yeah. Like looking at each other's eyes, glistening about our future intelligence or whatever. But keep in mind, like I, Hirsch and I are both people that, you know, really care about the people in our lives outside of work and put a lot of effort into those relationships. I'm really close with my family and the people in my life in that moment.

[21:06] We're not into the idea of AI. Keep in mind the discourse. That was peak, this thing's going to ruin humanity, and it's stealing artistic works and all that stuff, which I'm not saying doesn't have merit. But you go to the dinner table with my family, showing a demo of this ChatGPT thing that I'm staying up till 3 a.m. and just finding deep inspiration in it. And they're like, [21:28] "What are you talking about?" And so that's just like another moment where you're like, "Okay, the story of Arc at that point was great. "You made a great product for tech people. "Anyone that's in the software industry loves what you did, "or at least respects it, but your problem is "you didn't build something fundamentally mass market "and essential for everyone that uses their laptop." And so to the extent we were gonna build a new product, it was only if we thought we could achieve our original mission of changing the way the average laptop person uses the internet, [21:57] In a moment where the dinner table conversations with the people we deeply love and care about, we're often saying, you are too in the tech world, man. You're telling me that the thing that comes after ARK is this AI nonsense that's spitting out gibberish? So it's a real act of introspection to say, have I lost it? Have I become disconnected? Or do we just see what other people don't see because of what we're doing? And at least for me, it was this constant back and forth in my own head and with Hirsch of like, [22:25] what is reality? I don't know. Yeah. Yeah. I mean, not only for our friends, but also employees, you know? Uh, so that was, that was tough. Yeah. What was that like? Um, [22:35] when you had that company meeting and you were like, hey,

[22:40] This is a new product. [22:44] Do you remember, Josh? Was that at the offsite? [22:47] Yeah, I will say... [22:49] Since narratives get kind of added in retrospect, it really was more of a gradual evolution. So there were some key moments. There was this offsite in Montreal where we definitively said we're at least going to try to build the second product, whether or not it's Arc 2.0 or not. There was a moment when we said it is going to be a second thing, but it was obviously more gradual. I think one of the saving graces... [23:09] and I do not think we could have done this with a different company in a different moment, was that, you know, really to Hirsch's credit, we founded this company with some core values that we really live by. And one of them is assume you don't know. [23:23] So we have this culture of no one has any idea what they're talking about, and we'll have no idea until we try it. [23:29] And so that's why we have such a prototype-driven culture, experimental culture, for good and bad reasons. And so I think in general... [23:37] at the browser company, if anyone says, I got this wacky idea and I'm really excited about it, I'm not, you know, let's try it. People's first reaction is like, cool, prototype it and let's talk about it after that. And so I think that was one of our saving graces is that we could just sort of frame it [23:54] as what it was at the time, an inclination, a hunch. And again, I think as you built more and more things, I think it became clear to more and more people over time. And I also think we just built up a lot of trust. Right at that point, [24:07] We'd been building ARK for a number of years. People loved it. People joined the company because they loved it. And I think, you know, we had added some trust points to our bank account. And Hirsch and I drew them down to maybe a negative balance, you know, up until Dia got further along. But, you know, credit to the team. You know, it's not like there was an organ rejection. There were people that had their own concerns and anxieties. As is normal, it was a big call that was existential for the company.

[24:37] Thank you. [24:37] all right, let's see, you know. Yeah, and one thing I'll add is we also built the capabilities over time. So as we were prototyping, we were getting a better sense of like both where the value is here, but also like what are the things we need to get better at? Like evals, fine tuning, you know, from a technical perspective, from a design perspective, how do chat interfaces work? Our chat interface is the right mechanism, all of that. And so again, as we gradually built that capability, our confidence grew. Yeah. [25:06] - What do you say to people [25:08] Because I think this is probably on someone's mind if they're listening to this. [25:12] And it certainly was like, it was a thing I was thinking, which is, [25:17] Okay, I totally understand that maybe it's not gonna be like this gigantic, or it's just not gonna be this gigantic mass market consumer product or whatever, but like, it had millions of users. That's pretty funny. [25:30] Good. [25:31] And isn't like VC distorting your... [25:37] like what you count as good and aren't you kind of like killing something that so many people love [25:43] just because it's not like, [25:46] billions of users and like talk to me about that because yeah I want to understand how you think about it. [25:52] Well, Dan, actually, it relates to origin story when we asked you to be the CEO. So the origin of the browser company was I was working in a venture capital firm. And I was just shocked that all of the coolest, fastest growing Silicon Valley companies that were coming across our desks were all of a sudden desktop web apps.

[26:10] And they're desktop web apps for work. I mean, you started a newsletter called Super Organizers. This was the moment of productivity software, right? And so the idea was like, let's make a enterprise browser for work. [26:22] And as you said, take the superhuman, the notion, the Figma playbook and run it. And so Kirsch was going to be the founder and we were looking for a co-founder CEO. And the first piece, I call Danny's my best friend. He does this newsletter called Super Organizers. It's perfect. And maybe it would have been perfect. Maybe we'd be having a different podcast about how wildly profitable ARK is. [26:44] But. [26:45] The reason I say that is because the reason I ended up joining the browser company CEO, in addition to really just wanting to work with Hershey again, was this aha moment that actually the browser is arguably the most consumer piece of software in the world. Like there are very few pieces of software that your mom and your second cousin and your partner all use. [27:07] Like. [27:08] 4 billion people use Chrome every month and nobody cares about it. And most people don't have like a second browser for Netflix and shopping. And so if you care about what I personally have always cared the most about, which is how we as a society broadly use technology, then the motivation for me and I think the origins of the company was to build something at that level of ambition, even if that's not the idea that we pitched to you initially. [27:33] And so, [27:35] That was never part of the calculation, honestly, for me and Hirsch when we talked about like, oh, do we want to build a product with millions of people? We would have capitalized this company completely differently if that was the goal. We would have done so many things differently. So I think the idea of like, why wouldn't you be OK with, you know, a couple million people browser that, you know, makes decent money just comes back to why I personally got into software and what motivates me and Hirsch.

[28:05] a lot of things that were hard. There are a lot of things that were unclear. The why are we here? Why did we start this company? Has been consistent and unwavering from our perspective. I also, like... [28:16] Again, we're sort of fixating on Arc, but the reason for this was not just what do we do with Arc and how do we grow it. It was also how do we meet this moment, you know? And like, we were just so inspired by, oh my God, like this platform shift is happening and what software is and can be can be so different. And so much more capable and look very different. And we have a specific skill set and team that's capable of building that interface future. [28:46] for what's possible. And again, we didn't start with the idea we would start from scratch. We started with this idea we would evolve Arc in that direction and then that ended up being much more difficult than we thought and so we started with Dia. And I also think as Hersched, it's like as the technology was evolving, as we were getting more familiar with it, it was getting more clear to other people what was valuable about it and where the points of leverage were. We also started to realize that the browser layer was going to be very central. You know, at first it was sort of like, wait, we don't want to be that company [29:16] inspired doesn't mean you should pivot your company, right? They're inspiring things I see all the time. I'm not like, hey, let's go start like an indie film company, right? So the thing [29:27] Okay, I'm going to go. [29:28] It feels really disconnected from everything I'm doing every day. [29:32] I'm spending a lot of time copy and pasting and context switching and exporting tabs as PDFs and weird stuff like that. So being able to meet people where they are. Again, the idea of memory and context, like people didn't say context was king a year ago. Maybe you did, but most people did. I did. Yeah, there you go. This is why you got to subscribe to every, you would have been a year early. But, you know, starting to realize that, okay, the more that you tell these models, the more useful and personal they get. And I'm spending hours a day on this thing.

[30:02] as Hirsch said too, hey, we don't know anything about agents. This is sort of out of our wheelhouse at the time at which was coming up, but it seems like the big promise is they're going to do some things and tools for you, but wait a minute, that's not useful unless they're your tools with your authentication and we have your cookies. So it was also, again, to Hirsch's point of inspiration, realizing that structurally, [30:22] And I honestly felt like we had like... [30:24] you know, been at the end of the rainbow before the rainbow came and all of a sudden like a pot of gold appeared. And so it was sort of like, all right, I feel like we should probably try to open this pot. Like this feels like the thing to do, you know? Well, let's let's talk about what you found when you opened the pot. Like you guys made Dia and... [30:41] um i will say and i think this is this is an interesting testimonial given my uh my relationship to both of you and my relationship as an investor is like [30:50] I... [30:51] I was always an Arc user, but it was never my default browser. Safari was always my default browser. And Dia is now my default browser. I switched from Safari. [31:01] Hey. Which, like, and that's not, like, I love both of you, but, like, a browser is too personal for me to switch my browser because I like you. [31:14] Clearly, Dan. [31:17] So this is totally not for podcast purposes or anything like that. [31:21] just switched to it because I think the, and this is maybe gets to your point. One of your points about Arc is like, I used Arc for certain like heavy research tasks and for podcasting specifically because like Riverside doesn't work in Safari. You have to use a Chromium browser and I'd rather use Arc than Chrome. But Safari is so lightweight and fast and Arc was never like that.

[31:46] And Dia is so fast. And I was like, this is really good. And then it has the AI stuff in it, which... [31:55] it opens up a whole new way of using web pages for me that I really love. So it's like it's coming. It's definitely I think a lot of what you're saying is coming through for me. And it's been really cool to see it like get launched and get the reception. It's gotten. Well, Hershey, it's honestly it's worth talking about the architectural decisions because that was a whole nother thing. We decided to just be like, why don't we do this completely differently, too? Well, yeah, we really had a. [32:21] So Arc was really built around how do we prototype as fast as possible. And we made some shortcuts in the architecture to enable that, which had some performance impact. Interesting. And so we were really optimizing for how do we prototype and build new features to do that. [32:36] And in a way that we can do them quickly. What was the, just quickly, what was the architectural? It used a kind of Redux type state driven architecture, which is a little crazy for a desktop app, but allowed us to move really quickly. Over time, it became really complex to manage, which was an issue, but also performance was a big concern. And Redux is like the meta like React framework or thinking about how to pass information from UI layer to the back to the server and all that kind of stuff. [33:06] Exactly. And so, you know, we had a real choice last summer on when we're building DIA, do we build it on the same architecture or do we build something new that's really fast? And I think part of the, I mean, certainly the engineering team was really excited about an architecture that we were excited about that fixed a lot of performance problems that DIA, that Arc had. But also it was core to the strategy.

[33:36] just be a really, really good browser. What did you figure out about how to make something performant? [33:43] Honestly, a lot of it has been Swift. Swift has this new structured concurrency mechanism that allows you to just move a lot of stuff off the main thread. So just leaning really heavily into that has been helpful. And then shout out to some of our incredible architects, Max and Adam, who built this entirely new architecture that allows us to move faster than Arc, but also deals with a lot of performance problems. [34:06] - We figured out a way to still do prototyping quickly. - Yes, yeah, exactly. [34:10] And the reason I think it's relevant is one of the ways I think we got the team really on board is we sort of went to a bunch of our best, most creative... [34:19] People and said, [34:21] Hey, you get a blank page too. [34:23] So, you know, this had been a topic that sees architectural challenges and latency. These have been themes we talked forever. So we just sort of went person to person and said, well, [34:32] Hey. [34:33] Start over. You got to do your most, be your most abundant self. [34:38] dream. Come on. What are the things that you always regret? This is the moment to do like, you know, so I will say once the ball started rolling, people were able to find their kind of corners and pockets of what we were trying to do to either fix mistakes we made with Arc or do that thing they always wanted to do. You know, you can, you can feel it in Dia. There are parts of it there. I hope there's some kernels of the browser company that are consistent across it, but there

[35:08] no, we're really going to flex on this one. And we're not going to flex on this one in a way that's like ARC coded, but in a way that is what this should be right now. So I think you can feel that intentionality personally in the product relative to ARC. [35:21] And one discussion we had when we were building this out also was, it's funny seeing Angie Siegler's post now about how the eyebrows are going to take off and we're sort of kicking off this race, but that was a discussion we had about the architecture, which is like, [35:34] We have so much conviction in this route that when we launch this, if we are the first ones, we're going to kick off a race. And so how do we build an architecture from the ground up that allows us to... [35:45] move quickly in that race by the time we launched this. And so that was a lot of the input as well. Yeah. Hirsch and I had a very seminal conversation for us, at least, on that San Francisco trip I mentioned. And Hirsch flew out the next day before me and sent me this long, long essay in a way that, you know, I don't get long, long essays from Hirsch, you know. And one of the things that Hirsch had the foresight to see [36:11] in this new world was that it's not about the browser. [36:15] The browser is almost incidental, right? The real opportunity. And I think the language Hersch used at the time was this truly personal assistant. I think all of us right now are grasping for what the right nouns are for the work that we're doing. But Hersch had the foresight to say, hey, it's not really about the browser part. That is necessary. [36:32] But that's not where the value is going to come from. That's what's going to enable the thing you're turning to it for. Almost like the iPhone isn't really...

[36:39] a cell phone. I mean, it is, and you get that utility out of it, but that's not why the iPhone is so powerful. And I think in the moment for me, I was like, seems a little far-fetched, but hope he's right. And as time has gone on, I think his instincts were, in retrospect, absolutely correct. You know, even the original... [36:59] This vision video we put out for ARK talked about this idea of an internet computer. And we very much said, we have no idea what that means. But based on the trend lines in our industry, it seems absolutely certain that what feels like your computer in five or 10 years is actually going to be this layer that sits across all of your devices of all shapes and sizes, because all of your apps and files are on the web now. [37:29] desktop browsers and mobile apps and all sorts of things. [37:33] And on top of it is going to be this personal assistant or this intelligence for you. And it's going to benefit from the fact that it can intercept your queries on the new tab page and that it sees the docs you're writing in your browser tabs. And it remembers the things you did last week. But you're not turning to it for the tabs. You're not even turning to it for the desktop part. What you're turning to it for is this personal intelligence layer that is a wild invention of humanity that's going to help you do all sorts of things. And it's going to be miraculous. But don't forget. [38:01] The browser's just the enabling technology underneath. It's not the main event. So you can imagine how getting in front of the browser company of New York's employees, makers of a popular web browser called Arc that helps you with tab management in novel ways, couldn't really nail that messaging in the moment. But in retrospect, I think that long email her sent me was extremely prescient, regardless of if we win this race, if it's even zero sum.

[38:31] kind of underpinned arc. [38:33] That's a couple things that I think are interesting here about lessons that it seems like you guys have learned about building the future, and some things that I've definitely seen at every, like trying to do the same thing is, one, when... [38:47] GPT-3 was first a thing. Everyone was like, this is going to be an incumbent enabling technology. Like Google and Microsoft, they're just going to put a chat box on the right side of Chrome and they're going to put it in Google Docs and everything is just going to be for the incumbents. And I think one of the things you were saying earlier is, [39:08] just getting to go to people at ARC and be, or people at the browser company and be like, you have a blank page. The level of creativity that you get there is so high. And the level of power that you get is so high that it's actually, I think, impossible for most incumbents, unless they're starting from scratch, which is really hard to do, to just like slap AI onto this and actually build that layer that you're talking about. And another thing that I think is really [39:38] it can be really hard to pitch the future to talk about the future in a way that anyone understands um you have to build it you have to like prototype it you have to like feel it and see it and that's the only way that that then you can talk about it in a way that people understand and it sounds like that's that's sort of what you found is you had this sense that you could express but any way you expressed it everyone was like you're

[40:02] crazy. But then once you started to actually like prototype it, that's how you found a way to talk about it in a way that people can understand. [40:09] - Yeah, and two stories actually that come to mind that are very relevant. So Hirsch and I, our first company was acquihired by Facebook when we were, I don't know, 22 or something. And I'll never forget this internal-- - RIP branch. - RIP branch, RIP branch. [40:26] for the real ones out there that remember, uh, [40:29] I will never forget this internal all hands at Facebook, where Zuckerberg stood up and had this whole presentation about how in five years, [40:39] everything was going to be video. Social media was going to be video. And I thought he was nuts. I was like, this disconnected billionaire, does he know how slow it is to watch a video on your phone? And then like, even if you could fix the buffering issues, why would you pick a tiny little screen in your pocket? Like, what is he thinking? And in retrospect, that proclamation and honestly, others that he made, even if he didn't capitalize on it, [41:07] He was right. You know, I remember five years later, it's like, yep, TikTok's definitely a thing. And so I think part of what Hirsch and I took from that experience in different ways and similar experiences, like, hey, if you're going to start a company, one of the you mentioned Branch, me and Hirsch's first company, the origin story was a hackathon prototype. [41:22] And my takeaway from that and that experience was like, if you start a company based on a prototype, you don't have a kind of fundamental... [41:31] internal guidepost when inevitably your prototype's gonna fail. And so the intention of the browser company, inspired by our experience with Branch, observations of Zuckerberg at Facebook, was to

[41:43] Don't pretend like you have all the answers, but even just directionally. You're in LA, you're going on a road trip, you're going to New York or Miami. You may not know how you're going to get there, and you may hate Miami, and Miami might not be as cool as you think, but pick a destination that is five, ten years out and run towards it. And so the internet computer idea was that for us. And... [42:05] was pretty central, I think, to our ability then with Dia as well to kind of just go for it. What I think is interesting too about that is just to push on it a little bit, because I think there's a thing that if you're a builder listening to this, [42:19] you might take away that I don't know is actually what you actually mean. [42:23] is that you can see the future in front of you and you're just going to go figure out how to build it. And yeah, you don't know the route, but you know you're going to San Francisco or you know you're going to New York or whatever. But I think what's interesting about [42:34] the internet computer idea, [42:36] is, [42:37] What the internet computer meant changed. [42:40] Yeah. And so you're sort of, [42:44] you have this broad sense that there is going to be a different way of interacting with computers. [42:51] and that you wanted to build that, but what you thought that was going to be, [42:56] is actually very different than the thing that you ended up building. Because you had no idea that AI was going to be such a thing when you started five and a half years ago. And I think that's really interesting, too. And that's definitely been consistent with my feeling about every, you know, like, [43:11] When we were talking about super organizers, I was like, I want to build this newsletter, and then I'm going to launch products to the newsletter.

[43:19] we're doing the same thing now, but it looks so different. Yeah. And and why it works is so different because of AI. And I think that's a really interesting. [43:30] is to watch that vision unfold and fill in the details of what that actually means as the vision, [43:39] the world evolves and your product evolves and you learn new things. [43:41] Yeah, I think that that that point applies to that video story that I told you where at the moment Zuckerberg was saying that. [43:51] you're going to share the same types of personal vulnerable updates that you do today in video now. And while there's some of that, [44:01] I remember the moment I made the connection back to what Zuckerberg said at all hands and even realized that video had become a thing was I had the opportunity to meet Evan Spiegel from Snapchat. I think it was like 2019. And I was like, I got to ask him one question. So my question for him was like, what's one thing that's really inspiring you right now? And he told me about this podcast. [44:21] app from China or called TikTok, which again, it was pretty popular at the time, but I hadn't really come across it. And I was like, oh, why is it inspiring? He's like, it made it so anyone can be a celebrity for a day. [44:33] Yeah. [44:33] And I remember being confused by the description, but you can kind of see that he was right and kind of... [44:38] capturing what was unique about TikTok. That was very different than what Zuckerberg said. [44:42] That is not friends sharing intimate videos with each other in a news feed. And so the point of being able to say where the world is going directionally is definitely not, you know, the details.

[44:55] But it's still, I think, you know, really important. And it's tricky. One of the things, especially in our media cycle, is, you know, we built Ark in public. [45:03] But we also built it with prototypes and, you know, sharing everything. And so when we shared the Internet computer video, I also remember people internally being like, man, this seems kind of like, why are we adding this branding? What exactly does it mean? But at least we had built a little bit of trust to say for people to say, like, OK, I like the things they're putting out. So I don't know, maybe they know it. At the time with Dia, it was even more of a kind of high conviction bet on where the future was going. But we did not have an audience that wanted any proclamations from us about AI. [45:33] sort of question like, you know, [45:35] were we right? And so one of the interesting things about Dia was there was a larger disconnect developing it than Arc between the internal feelings and beliefs that we had and like what we could say publicly. And so that was also kind of an interesting part of building Dia versus Arc. [45:54] What made it trickier also was because we bet on the scaling laws and the curves. And so a lot of our conviction wasn't even possible last year, you know, and so. [46:04] Oh yeah, GPT-4-0 made Dia work. Like when we started working on Dia, I might get my dates slightly off, but I distinctly remember messaging people, [46:15] our person at OpenAI and being like, [46:18] this model made our thing. [46:20] happen. Yeah, it's crazy. Like, we've had that same thing where [46:24] you're you're working on something you're trying to build something it's not working you're doing all this prompting you're doing all this like architecting to like string string different models together and then we had this happen with autopus 4 recently like

[46:36] It just worked. And it was like, okay, we can throw out three months of work because now it just one shots it. Yeah, I mean, a great example of that from Dia is, you know, to me, the core idea of Dia is this idea that it should get better with every tab you open. Just like every time you swipe a video on TikTok, it feels like it's getting to know you and it's more useful to you. That's how Dia should feel and get better with age. [47:06] the thing that ended up differentiating when all of the models and intelligence commoditized or looked that way, it was going to be the thing that was the most personal. So Hirsch, [47:15] And a team worked on memory for nine months, [47:18] And at some point, we killed the project. [47:21] We just toss it away. We're like... [47:24] Guess it's not going to have memory. [47:26] And then... [47:27] six weeks before launch. It was like, wait a minute. I think the fact that, yeah, this context window over here, like, well, no, no, let's try this approach. Like, let's do memory one more time. And oh my God, that was a thing to deal with internally. That project that we just tossed away after nine months, having this high, like top down conviction, like, [47:43] we're going to do it again right before this big launch we've been working for for a year. And again, I don't want to say it's work because it's too early, but like, [47:52] it's looking pretty prescient again from Hirsch. Um, but, [47:56] But again, that was enabled by, in a span of six months, the fundamental building blocks changed enough in our understanding of how to wield them that a thing that smart people banged against for nine months didn't happen, and then it happened pretty much overnight. And cost and latency and all that stuff.

[48:15] Tell me about, you know, you've gotten it out in the wild. You have this thesis that it's going to get better with every tab you open, which I love. I think that's such a clean way. That's such a Josh, like Joshism, like perfect. I love it. Means a lot coming from you. [48:33] What are people using it for that? [48:36] has surprised your or, you know, what have you what have you learned getting it out in the wild that you didn't know? [48:43] Here's your take, Josh. [48:45] I think I am surprised that people see what we see as quickly as they did. [48:51] Thank you. [48:52] I thought this was going to be a really painful moment for us. [48:55] Thank you. [48:55] And I prepped, and Hirsch and I prepped the team for that. [48:58] hey, it's going to be a somewhat tricky, brutal summer because Dia is so basic. Because again, keep in mind, the idea of Dia is not the browser bit. It's not the tab management. It's not the things you can see. It's all of these capabilities that Hirsch and the team have built under the hood. They're like, oh, if we just like then piece them together, we can build this application platform for AI. But that wasn't what we released two weeks ago. You know, we released the building blocks under the hood, but that was pretty basic. And we thought it was useful, but we thought it [49:28] - Yeah, I kept asking Hirsch for access, and he was like, "Oh, like, later, like soon, soon." - And he did in the outfit for what it's worth. We got to absolutely try. - My friends and family were so nervous about me, because my style of friendship and working on what I do is I enthusiastically want them on early in their feedback,

[49:46] Admission, my wife, who I've been with for over a decade, didn't try Dia until launch. [49:53] That makes me feel so much better because I was like, Hirsch is just like, I don't know what's going on, but like, are we growing apart? [50:00] And that shows you, I actually felt like an artist, not in terms of an artist in terms of our quality, but just like the nerves to put this like personal work on. So the biggest surprise is people got it right away, despite in its biggest in its current state in terms of what it's useful for. What do you think they get? [50:18] I think... [50:20] MG's post, [50:22] really captures it, which is you can sort of feel... [50:26] the new latent possibilities underneath the surface by fusing these things together. Like really what Hirsch and the team did is they broke apart the browser [50:36] And they rebuilt it, put the pieces back together with models in every core part. And so you can't visually see that, but as you start using the product, like this morning, actually, I saw someone shared that their, [50:49] They clearly been using Dia for a couple weeks. So people discover these little Easter eggs, like, oh, I can do this. And what they found is when they look up songs on YouTube, they can ask it for the tab chords so they can play it on their guitar. [51:01] That's sick. I want you to use that. I actually do this all the time where I like send YouTube videos to Gemini and then have it tell me the like the transcribe the piano. So I'm going to have to do that. [51:14] And so I think people finding the like, it's subtle, but it's so much more convenient and so much more powerful to do it when the one plus one equals three. The big surprise on the product side was...

[51:24] Sort of that observation taken to its extreme, which you would never put, hey, you know what you can do in Dia on our website? You can look up the tab cords for the guitar when you're watching a video on YouTube. It's so personal. It's so the TAM of that is tiny. And a couple weeks before launch, because our big vision for Dia is that actually it's this kind of application platform that people will build things on top of, was like, we should gesture at that just a little bit. Right. So let's we have this concept of skills, which are effectively AI apps. [51:54] We have a couple that are first party, but they're really basic. - Just for people who are listening that haven't used it, like what is a skill in DIA? [52:01] Yeah, a skill in DIA is the equivalent of an, or we hope will be the equivalent of an application of, [52:08] on an iPhone or on a computer. But in this kind of AI world, what it is, is it's a system prompt. [52:15] a model or number of models, tool use, and stringing those things together to do something for you. So someone might make a skill that is my job is to do sales and I look at leads on the internet every day. And when I see a lead, I have to look up X information about them and extract Y information and consider Z framework. And they kind of put that logic and framework and tool use into a skill

[52:45] version of the product that we released, the platform is still extremely limited. And so we weren't going to let anyone make their own because we were waiting until later this summer when we could actually expose those, the really cool building blocks, reasoning models, memory, and the like. But we're like, okay, we got to gesture at it because that's why we're here. We were this big internal memo about the app strategy. And so we expose the ability to make the most basic custom skills possible that are basically just little prompts and added a week [53:15] in terms of what people are doing and is now, essentially the whole company is just building out the skills platform now. And the reason that that relates to the tab cord example is that one of the things we used to say with ARK was that, [53:31] you know, no one cares about their browser, but they spend hours a day in it. It should feel like your home on the internet. Why do we all live in these like drab hotel rooms online where all of ours look the same run by Google? It should feel as cozy as this background you're in right now. And you're seeing that principle play out, whether it's the first party features like chatting with your tabs and people finding use cases that are new and very personal to them, to people be like, wait a minute, you're saying I can make my own one of these that like when I'm [54:01] exactly what I wanted to do. It's almost like the normie, and I'm saying that for myself, I'm a sociology major that doesn't know how to code, like, everyone's looking around and seeing Cursor and Replit and Lovable, and these companies just, wow, you can make new things, and there are people like me being like, what, I...

[54:16] I want to make new things. And I think we're tapping into something there of like, almost like handmade software. Like people wanting to make things for themselves. Like me that I literally had to hound Hirsch at a hackathon in New York when I was 20 and beg him to make something with me. Because it was the only way I could express my ideas in software was to hope that someone like Hirsch would like, [54:38] want to make something with me. And so I think it's starting to tap into that nerve. And while that [54:42] observation is old, did not expect that this early in the product development, since we felt like that was like to come and to be revealed. This was an amazing insight by Josh, actually, that AI enables sort of a new class of software where rather than us thinking about what a software we can create that is [55:03] useful for a large TAM. Instead, we can build software that allows people to make their niche software just in time, you know, whenever they need it. And so it's just a different way of thinking about what software can be. And that's because of AI. Yeah, I think there's a couple of other examples of I mean, chat GPT is maybe a good example, where Excel is a good example, maybe notion where it's not one big use case. It's not one job to be done. It's like millions of tiny [55:33] Which makes it a little bit less like... [55:36] a traditional, like, even like SaaS product where you're like, well, we solve this particular problem for a user and a little bit more like creating a language where you're creating building blocks and then anyone can use the words and the language to express whatever they want in whatever situation they have. And that's like,

[55:53] A, a uniquely powerful type of product, and also B, it's a uniquely challenging type of product, because getting that language right and then communicating it to people that you can speak in this language is really hard. [56:05] Thank you. [56:06] Yeah, I think we're particularly excited about it, A, because we love human-computer interaction problems and design problems. And B, the browser just knows so much about you. It's such a personal piece of software, and if we're building this memory system and it's getting more and more personalized to you, it can help you with that. Dan, it's funny you say that. I'm looking at my phone because late last night, Dustin, our head of design, he was so inspired by these custom skills being created. [56:36] Extremely humble, soft-spoken Canadian. He doesn't do things. He turned to me a week ago. He's like, Josh. [56:42] I need a week alone to prototype. I know what to do. And that sounds like something I would say. That's not something that Dustin says. So when he said that, I was like, yeah, okay, cool, go. And he texted me last night saying, [56:56] about what he's been prototyping and kind of show me. And I sent him a text back. I said, actually, I would have written it differently if I knew it was going to be a podcast. But, Dustin, you're legit inventing a new programming language. [57:08] Non-technical people like me want to feel technical. I aspire to know Swift. I don't. Give me a language to learn. And I don't mean language like complicated. Anyways, it keeps going. But so I guess that's the surprise. The idea that AI and the power of it is when it's going to be extremely personal to you with very minimal work was the idea of Dia.

[57:38] going to be more powerful than the base itself. That was consistent. [57:43] The idea is going to be skills and it's going to be really jumped to kind of the third party, not first party. And it's going to actually require this kind of new programming language. It's not programming. It's kind of like the Internet computer. We struggle with the words. [57:56] But the sensation is so obvious, which goes back to what we've said before. It's like one of those moments again where it's like, okay, entire company, we're going that way. 100%. Come on a podcast. Like, what is this skill? What are you talking about? It's like, I don't know if the listeners are like, what are they? Like, these guys are crazy. But like, in my bones right now, it's like, that is so clearly where the world's going. So like, let's try our best to run that way. [58:26] a while away and because of this launch and just with the incredibly basic version we have in there and how people are using it, I think we've been inspired. [58:35] I want to get down into, like, I think we've traced the whole, for lack of a better word, arc of the company thus far, which is like... [58:47] Do ARC. [58:48] big success, but maybe not the kind of success that you guys were looking for, announced that you're pivoting like a year of just [58:56] pretty much hell was maybe... Getting punched in the face. Maybe a good description. And now you're like, we're back, baby. We're so back. [59:06] There were a lot of problems too, but yeah.

[59:10] And, um... [59:12] I just want you guys to get [59:14] as real as possible with me as if there was no camera and no mic here, like the way that we would, if we were just hanging out about, [59:22] Yeah, what that what that has been like, like, I've watched you her like and sitting down with you and you're like, there's like dark circles around your eyes and you're just like, oh, like, what did I do? But but even in that moment, being very like, but I know that this is right. But also this sucks and I hate this. Yeah. Tell me about what has that been like? Like, I've seen I've seen a lot of the ups and downs. I've seen. [59:48] um, [59:48] I think people probably don't realize how much... [59:52] For example, when you're pouring your life into something and... [59:56] everyone on Twitter and maybe people internally are like, you're making the absolute wrong decision. You're, [1:00:02] this is awful. Like, yeah, [1:00:04] And on the flip side, now everyone's like, this is amazing. Like, what is what has that all been like? [1:00:09] Uh... [1:00:11] It's tough. I think Josh said it best. You... [1:00:13] It really feels like you're getting punched in the face a lot. And it's this weird dichotomy because, as you said, we kind of like, at every moment, we were like, no, the strategy feels correct. You know, it feels right. And everything that's going on in the industry and as we are making progress, we're like making good time on it. But yeah, when your family and your friends and the outside world and your employees are... [1:00:38] maybe less than enthused about the direction, it's really difficult to sort of manage the two. I think there's probably two aspects to it. There's one, just the head game of being like, okay, as founders, we have conviction here. And to credit, so many of our employees were on board and excited as well. But a lot of it was just trust in us. And that dichotomy of like, are we crazy? You know, are we just like completely full of it? I mean, you are, but you're right about this.

[1:01:08] And then there's just like the day-to-day tactical stuff where it just, [1:01:13] hurts a lot. I think press and then managing morale, people leaving and attrition, hiring becomes way more difficult. It just feels like you're on very hard mode during that transition. Then you had a kid a year ago-ish. Yeah. Josh has had two kids. On no sleep. You're [1:01:39] Yeah, it's a meditative exercise to stay present. I just get through every day, I think. - Yeah, I will say there was one session I had with my coach, where he was asking me something about the product strategy. No, he asked me how I was doing, and I gave him one of these kind of excited, genuine, [1:01:58] takes on this, whatever the equivalent of the Dustin prototype was, which was earnest. We said, no, Josh, [1:02:03] how are you doing? How do you feel? And I was like, I've been waking up every morning with a pit in my stomach. [1:02:10] and he was like, let's talk about that. And so this year professionally was by far the heaviest I felt in my body. And I think the things that get you through it are both Hirsch and I have, we're really close with our partners and our families and each other. I think one of the things that Hirsch and I have this dynamic is we both go through funks and highs and we both have the things that trigger us.

[1:02:40] these moments where like I'm really low [1:02:42] People may not know, but Hirsch knows the telltale signs and he knows how to pull me out of it. And I'd like to think the same is true with him. And vice versa, yeah. [1:02:52] No, it's not. [1:02:54] fun at all. [1:02:57] Obviously, there are things that keep you going. And I think for me, it has really been just trying to stay present and quiet and focused on the internal... [1:03:07] there is something in there that's making you get punched, that you're okay getting punched in the face. And so, but no, it was a year of constant doubt and ups and downs. Because I think the other thing people miss is the browser company had a Cinderella story first few years. We basically had no, I mean, it didn't feel like this in the moment, [1:03:25] But in retrospect, our biggest challenge was at one point, ARK was so popular and so taking off that we had these performance and reliability issues where it was crashing. And that was like the hardest thing we had. We were sort of like this darling in terms of people in our industry raving about it. The guy who ran Chrome for 16 years came to work for us as an IC. Like it was just a storybook. [1:03:55] not a runaway success. We had our challenges and I'm sure you remember them, Dan, but it did sort of feel almost like our divine right that it was going to go this well. And so I think all of a sudden having it so clearly feel different, [1:04:09] a never, almost a never ending to the, every morning there was some new thing. Someone was leaving or some competitor this or whatever. Um,

[1:04:17] Yeah, it's tough. I will say it makes... [1:04:21] Part of the reason I think I feel so elated right now is not because we're out of the woodwork. In many ways, we are... [1:04:27] and the most competitively challenged market in the world, right? Like there's a lot is still ahead of us. But part of what it is is this sort of vindication might be a strong word, but like, oh, we were right. So if anything, I don't think that will ever happen to me again professionally. Not that I won't have tough moments, but it's almost like taught me. My dad told me that one thing about getting older is like you get to like build your intuition because every time something happens, you can look back and say, okay, was I right or wrong? [1:04:54] What was I right about? What was I wrong about? And you gain that confidence. So right now, I'm not this confident that D is going to be the next Chrome, but I know that what me and her felt in our bones was correct. [1:05:06] And so I'm almost excited for in 20 years, if I ever started another company with Hirsch, that moment of like, okay, we're gonna get punched in the face, but like, don't let don't waver. I definitely had a lot of moments of doubt internally, privately with Hirsch throughout the year that manifested in all sorts of ways. But I'm grateful for the life experience of like what it has taught me, regardless of what happens next, honestly. [1:05:29] I feel you. I mean, I've had not quite the same kind of thing, but a couple moments where the company just like almost fell apart. And it was like, I could just... [1:05:39] stop. And it could just be like things are just not working. And I was like, no, I think I'm going to keep going. And those definitely have been the...

[1:05:51] in retrospect, now that things have seemed like they're [1:05:54] going well, like big [1:05:56] self-trust moments where you're like, there's something inside of me that I can listen to. And that I think as a leader is really, really important. And it's something you only win by [1:06:09] Unless you're delusional, you only win by like going through that like [1:06:13] oh my God, things are about to fall apart, but I'm kind of like doing it anyway. And then being like, oh, it worked. [1:06:18] Yeah. [1:06:19] The doing it anyway really feels true. Some days you just wake up and you have to play the part. Get through the day. And yeah, you just get through it. [1:06:32] I want to talk now about just like [1:06:36] Josh, you mentioned the competitive environment. How are you guys thinking about [1:06:40] okay, we've got some validation. People are excited about this. It's still really early. And now there's a big target on our back. Like everyone is starting to figure out, oh, the browser might be, [1:06:51] a really important layer to be competing at. I will say if you've been reading every I've been talking about that for a couple of years. But but if you're not reading every it feels like now people are sort of really starting to like starting to wake up to it. [1:07:05] Perplexity, I think, has a browser or they're at least working on one. [1:07:08] There's rumors that a couple of the big AI labs are really paying attention to this. So how are you thinking about how to win as a still relatively small company compared to the big incumbents and then compared to other startups? Like, yeah, how do you think about winning?

[1:07:25] On most podcasts, I'd give you the answer that I would give a VC personally. You know, it's like, okay, we think memory is going to lock you in this way and our application platform and our design... [1:07:35] I think the truth is we had the same questions with ARK. [1:07:39] And we've been in this industry long enough, like, [1:07:42] Whether it's naive or enough, I just feel like we have a clarity of thought and perspective that to me, from what I've seen from other companies and heard, feels somewhat unique. I think we have a sensibility that like we just got to run as fast as we can. [1:07:59] but not frantically and not out of scarcity of beating competitors. But if we can stay locked into just being like, [1:08:07] Yeah, that most energetic, like, this is what we believe sells, like, [1:08:11] That's the truth of how I think we're going to compete, and how we've competed so far since the beginning of the company. [1:08:17] Again, I don't want to like diminish all the thought we've put into the ways in which like, okay, if someone clones our browser pixel for pixel, what can we do? And but I think at the end of the day, I think one of the things we're grappling with as an industry is what are the moats anymore? [1:08:29] I think so many of the traditional moats for product, so much about this moment is questioning these dogmas. [1:08:34] that I had just taken as natural law of how our industry works and products and competition. So that, I don't know, maybe I'm feeling a little bit too hippy-dippy right now, but I'm sort of just like, it's all noise. It is all noise, just like it was noise before. It's just a distraction. Like, if anything, I regret spending so much time, [1:08:55] time this past year, thinking about things outside of our company, instead of just like,

[1:09:00] putting our hat down and having fun with it. And I think everything that has resonated so far were parts of the product of the strategy. We're just like... [1:09:06] Can I curse? Like, f*** it. Like, let's just have some fun with it. If we're going to get punched in the face, let's just go and do this thing. Let's just be unapologetic. And the places where we did that, [1:09:15] People can feel it. There's like a character to it. And the places that were very like... [1:09:21] like top down and, and, and not even top down, just like, [1:09:25] you know, let's get ready for the HBS case study. Just like, you just feel it. - Here's how I should do this. [1:09:33] I think, you know, or maybe that's maybe that's a weakness as me as a leader. And but I that the answer is like, damn, we're gonna do our thing. Like we've been doing our thing. My learning from the last years when we do our thing, like at least some people resonate with it. And we've got some [1:09:48] foresight. So yeah, I'm excited to see the other. I think what you're saying is vibes is the moat. [1:09:56] And you'll figure out the rest, which I actually love. I think that that's totally right. I mean, here's the thing that I've thought a lot about. I am a DAU. I love ChatGPT. I love that product. There are some things I would personally do differently, but like, I love that product. Remember this conversation with OpenAI like a year or two ago? Oh, now like every, you know, [1:10:18] And then it's like, [1:10:19] You use those products and they feel derivative. They feel like someone said, oh, we got to go like that. Yeah, that's the feature. Let's go do that thing, too, from an intellectual perspective. But you can feel whether or not you like it or not, you can feel this is a research lab that like decided they were going to make their life's mission to do this a long time ago.

[1:10:40] And now their conviction is building and those intuitions building for what matters. I'm not saying the chat GPT is going to be the end all be all or there won't be other things. But I think what would Sam Altman have said, you know, I'm not comparing myself to Sam Altman, but two years ago when it's like, all right, now what? It's a simple product. It's just chat. Google has all the TPUs. And it's like. [1:11:00] Why has Chachapiti continued to be the fastest growing company of all time? What's their moat? [1:11:06] So I don't know. [1:11:08] Yeah. Vibes is definitely one. I also think there's some systemic things about our company that I am excited for in this new race. One is, I think the assume you don't know, I think we've built so much of our culture and how we build products to build entirely new things, which is very different from copying or, you know, seeing what's in the market and sort of like taking inspiration, perhaps. [1:11:38] inspiration to like in this entirely new world figure out what works um and then uh i think the second is taste you know i think like our josh and our designers um i'm just like blown away by what they come up with on a weekly basis i would also say for the for the every audience i think there is a there are conversations we have a lot internally that i would be shocked if the perplexity team or whoever else talks about this in the context of the browser which is when i see

[1:12:08] we started too. It's like, it's very engineering tech forward. Hey, there's some new computer use model that, [1:12:17] And we're going to go have a do a bunch of tasks for you. And you're going to like, we're going to book that thing for you. And it's like, it's, it's, we started that. Like, I get it. I totally get it. What I've found is I'm having this like very interesting, I was a sociology major. And I, and I'm sure you've seen this too, Dan, where I, I find myself talking to people, not in tech. It'll be at like a party on the weekend. I'll meet someone somehow it'll come, AI will come up to you on GPT. And almost to a person, everyone's like, [1:12:42] Yeah, like it's kind of weird, I think it's weird, I kind of feel weird talking about it, but like, I'm like asking it for advice, or like, I got its help with this like, health thing, and like, I don't know if it was right, but I felt better at the end of it. And there are people, there's like an emotional intelligence to these models, [1:12:57] that because of the origins of what the research labs and the sorts of people that are at the forefront right now that are more driven, I think, by the benchmarks and the raw IQ, which again is spectacular. We're really excited about going back to this home on the internet. Like, okay, if we have, if we know what you do every day, personally, professionally, and these models are [1:13:18] can be subjective and think and give advice and joke, at least give the perception of it, brainstorm. What are things you can do at the intersection of, yes, you got it, you have a job, you have a job, [1:13:28] You don't want a browser that's your therapist. No one wants that. That's not what I'm saying. But actually, I think, okay, when you think about the tasks and the workflows that have high economic value that you get paid for that are on your to-do list, there's a component that's clicking buttons for sure. And we should click those buttons for you. But the hard part about it is not checking out. That's a minor nuisance that, yes, computers should solve and are on their way to solving. The thing is, all right, I'm going upstate with my family to where you guys are at the end of August. Which town do I go to? Which one would we like? What's kid-friendly?

[1:13:58] Like we don't really want something fancy, fancy, but also like, I'm kind of like, don't want to sleep in a tent, you know, for this vacation. And those kind of qualitative, subjective things. [1:14:09] reasoning, in our opinion, is actually where a lot of like if the value is going to keep going up the stack and like Maslow's hierarchy, I guess is another way to say it. So if we're just going to assume we have this like cheap intelligence, commoditized across models, accessible, can click buttons, can do things for you, that's important. [1:14:26] But where's the value going to be then? I think it's going to keep moving up. [1:14:29] to some of this more emotional intelligence stuff. And so I'd be surprised if the other... [1:14:35] AI browsers have that [1:14:37] so central to their North Star versus the light. You're going to have these agents that are going to do these things for you. And that's important, but it's to us is missing the force for the trees a little bit. [1:14:45] On that sort of like emotional intelligence and sociology point, I had Nash on this podcast maybe a year ago, who's your head of storytelling, who is incredible. And one of the things that I... [1:14:57] really loved about her perspective, which I assume, Josh, you share in this approach, is we were talking about how you guys came up with the laptop class and how you think about marketing and [1:15:09] It was very like, well, I was like listening to this like Janis Joplin record and I was thinking about like that era of music and and the social change that was happening. And so it was very kind of rooted in what are what are historical examples of moments that are kind of like this, but also thinking about moments and technology and music and art as being part of a conversation and being the next turn in a conversation.

[1:15:39] is true about your approach is you're always taking inspiration from [1:15:43] really diverse, weird, interesting artistic sources to help inform the product direction, how you think about things. And I'm curious what those are right now for Dia and for the features that you're thinking about building. [1:15:56] Yeah, I think you're right that we've always drawn inspiration as much from... [1:16:01] places and moments outside of technology industry and today. One of the things that's interesting at the time we're recording this podcast is I think just like every week, it feels like there's new breakthroughs in the AI space from a technology perspective. I feel like our own intuition and understanding and conception of what we have in front of us is changing as well. But I'd say the things... [1:16:23] on my mind right now, I'm going on vacation next week. So what am I going to read about and think about? I had this conversation with, [1:16:30] My closest friend, my Dan Shipper, best man in my wedding, who is just he's not in tech, but he's always been the one that was like, hey, the Snapchat thing is going to be huge. So if I ever started VC fund, like I'm hiring this guy just to be like the mystic suits there. And he was the first one, again, not a tech industry guy. [1:16:47] that was like, there's not a thing in my life I do [1:16:51] Josh, [1:16:53] that is an important project for work or my personal life that I'm not [1:16:57] working with AI, getting advice, I forget the word he used, advice, collaborating with. So to what I said before, that feels like the profound shift now [1:17:07] in the world that I see outside of the industry is happening.

[1:17:12] And so that sort of new relationship with technology and with this intelligence makes me want to read about – [1:17:20] Honestly, romanticism. I wish I could go back to college. Someone I work with, Abby, actually dropped a bunch of books on my desk for this vacation with little notes. They're all romanticism related and what chapters I should look at. So I'm really excited to understand how in the past, when your relationship changed with technology and even what it was and what was possible, how did society and art and culture react? And then the second thing I'm [1:17:50] objects or places in your life that given the centrality, that importance, if you're doing every project in your personal and professional life, how do you build comfort, [1:18:01] a sense of safety in those spaces, even when maybe there aren't as comfortable as you wish they were or not as safe as you wish they were. How do they feel like yours? How do they, you know, this is really squishy stuff to some, but, um, you know, I, I played the, uh, percussions and drums growing up and I remember, you know, I got a really cheap drum set. I was a kid. And at some point I was like considering getting a new one, but it was like, man, the way that you wear in the snare and like, again, I know this sounds kind of like, or as a baseball player, [1:18:31] References are a little overdone, honestly. So I'm excited to go like a click deeper on that. You know, I picked up Christopher Alexander as like stereotypical as it is. Like I'm going back to Christopher Alexander. So to me, that is going to be, whether it's us or someone else, I want a timestamp in 24 months.

[1:18:50] I predict that the AI interface that people feel the deepest affinity and value from will be the one that they actually have the deepest kind of emotional connection to, and not in a way that they're like having [1:19:05] with the bot or anything like that. But actually just from the perspective, this intangible feeling, [1:19:10] In the same way, I guess, you know, iPhone versus the Android at this, [1:19:14] point. There's not a good... [1:19:16] I don't know. So anyways, you can kind of hear the answers kind of all over the place, but that's honestly how we kind of do it at the browser company. I think this is great. I mean, I'm very, very also into romanticism. And for you who are listening and they're like romance novels, it's the reason I like romanticism, I think it's relevant. And I'm curious what your take is, is like. [1:19:36] Um, [1:19:37] Romanticism was a movement in the 19th century that was a response to, um, the, [1:19:44] enlightenment rationalism, basically from the 17th and 18th and 19th centuries, where physics was so successful that it sort of took over the rest of culture. And we can still sort of see that in a lot of ways, but starting with, you know, Galileo and Newton and [1:20:00] that revolution, there was a push to reduce everything in our life and our experience to things that we could explain in... [1:20:08] physics, more or less. [1:20:10] which is still the case. And romanticism was a 19th century movement to be like, actually, there's a lot about human experience that is more like vibes-based. Intangible. Intangible, and that that's actually like one of the best,

[1:20:23] parts of human experience. And so there's a lot of art and work from that period in Europe and in the United States that touches on that. And I think language models are [1:20:36] a weirdly romantic coded [1:20:39] technology because they are so squishy. It's the first example of software we've ever had that you can't reduce to like rules in the same way that you can regular software and you have to work with on on vibes. And I think that's a huge shift for software and for culture generally. [1:20:58] And does it sound familiar? You know, if you if someone was listening to what you just described, it sort of describes the world today in a lot of ways. And it's one of the things I still feel a love hate relationship with this kind of AI industry, if you want to call it. It's extremely rational, extremely mechanical, extremely the benchmarks and the capabilities. But, you know, reviewing these on every podcast. [1:21:19] You can see the SWE benchmark, and then you can feel it. [1:21:24] And there's a difference there. And then it's also, I think, very relevant in terms of the like, [1:21:30] There's what it can like intellectually do and how powerful it is. But what do you want? I think that's what we're going to be grappling with as a society in the next five years is these sort of like almost things that sound philosophical or like you should be in university talking about them. I think it's going to be really important again for us for the first time in a long time. Like. [1:21:47] What do we want from this life? And I know when I've heard podcasts and people talk about that, I'm like, oh, my God, I got to turn this thing off. So hopefully people don't stop listening right now. But I will say I would encourage people, even if you're just going to ChachiBT or Dear or whatever, like do a couple queries about romanticism and find and replace some of the nouns with where we are today. And to me, when Abby said,

[1:22:12] kind of reminded me of this. It just was, I can't unsee it. So I'm excited. I'm going to do like a Romanticism 101 class for the next week in the mountains. And, you know, sorry for everyone at all hands the Monday I get back. You're going to get like a lecture about, you know, awe and magic. [1:22:31] I love it. If you want to... Maybe that's what our every essay should be about. Maybe we run a romanticism essay. That would be great. [1:22:38] I love that idea. How do we not? Well, Josh, Hirsch, this is amazing. Thank you so much for coming on and talking to us. And good luck with Dia. We're so back. We're back. We're back. [1:22:50] Thank you for having us. We're back. Thanks for having us, Dan. [1:23:08] because this show is the epitome of awesomeness. It's like finding a treasure chest in your backyard, but instead of gold, it's filled with pure, unadulterated knowledge bombs about chat GPT. Every episode is a roller coaster of emotions, insights, and laughter that will leave you on the edge of your seat. [1:23:25] craving for more. It's not just a show. It's a journey into the future with Dan Shipper as the captain of the spaceship. [1:23:32] So do yourself a favor, hit like, smash subscribe, and strap in for the ride of your life. [1:23:38] And now, without any further ado, let me just say, Dan, I'm absolutely hopelessly in love with you.