#2283 How Dan Shipper uses AI to build artisanal AI apps

0:00Hey there, Freedom Fighters. My name is Andrew Warner and this is a new series for me. It's called The Next New Thing. Here's what's up in this interview, then an intro, then we'll get right to it. Every new product we've built has always been like, hey, I just kind of want this, whether that's me or someone else on the team. And the really cool thing about AI is you can build that in a couple hours usually. Can you tell me how many people are using the different tools that you guys have created? I can tell you we have about almost 7 ,000 paid subscribers, 1.3 million in ARR. We grew ARR 45 % this quarter.

0:29How did you know that monologue is something that's worth spending time on? We started using it internally and then started to add data customers to it over a weekend and then demoed it. and everyone was like, oh, this is really cool. I'm going to try it. I would like to see some of the prompts. Will we be able to give that to listeners? Yeah, absolutely. How do you feel about me being a little critical of the software? Are you sure? Dan Shipper is the founder of Every. They make AI-based software and write about the AI industry in their newsletter.

1:03Can you tell me how many people are using the different tools that you guys have created? You know, it's a little actually hard to track. And that's one of the things I've been working on is how do we make the whole ecosystem that we have? We have four AI apps that we run in-house. How do you make that all more cohesive, both so that it's a better experience for users. And so internally, we can understand how well is each thing doing. And I don't have all the numbers in front of me. I can tell you we have about almost 7 ,000 paid subscribers across all the apps in the media. and altogether we do about like 1.3 million in ARR growing really fast.

1:39We grew ARR 45 % this quarter. And yeah, that's the basic state of the business so far. And if I'm looking at it, you've got Spiral, which allows me to turn my podcast into lots of different social media content. Monologue, which allows me to talk into my computer and it does dictation for me. sparkle it organized all the files on my computer and cora which saved my ass it's the ai assistant for email because it goes through big email inboxes and then tells you what you need to pay attention to and you can go find the rest in an orderly way right that's the business plus a newsletter people pay for yep okay um i've heard a couple of things about you number one you say we're mostly bootstrapped what does it mean that you're mostly bootstrapped um basically we've taken a little bit of money so we raised 600k in 2020 which is when we started the business um and then we just raised another round from reed hoffman and um starting line vc and that was a two million dollar round but uh we raised it as what i call a sip seed round so So they have committed up to$2 million, but we can pull down the money whenever we want in whatever amount we want.

2:59And we've pulled down like$500K or something like that. So a little bit over a million dollars raised over the last five and a half years and done in a way that when we raised money originally, I told all of our investors, you should expect that this is the last money that we ever raise. and I think the investors in this round also are aware that we're probably not going to go raise$10 million Series A from Sequoia or whatever. I don't even know if$10 million, maybe$10 million is too small now for a Series A, but$30 million Series A from Sequoia in 18 months. We really try to run the business at break-even and we're growing really fast.

3:47I feel very ambitious about what we are doing. And also, it's very important to me to preserve this spirit that we have, that we're a little bit of a creative playground. And we get to take risks and experiment and build new things and kind of live in the future together. And I think there's a tradeoff there when you're focusing on maximizing a particular opportunity versus playing and exploring new things. And we try to do both. Okay. So one of the things I want to understand from this conversation is how you figure out what to create. So I listened to the interview that you did with the entrepreneur in residence who built Monologue, the dictation tool.

4:29And I wanted to understand how it was built and the way it was built using AI and all these tools, fun to watch. But what I'm trying to understand is how do you know what to create? Because there are tons of dictation apps now that are all using Whisper. How did you know that monologue is something that's worth spending time on and creating? We have a very non-process oriented process for doing this that is very much an outgrowth of the way that we write. And the way that we write is, you know, we write about what we're interested in. Like whatever, whatever I'm currently into is the thing that I'm writing about, because I think that's the best, that's the best kind of writing.

5:10And obviously there's always, there's always like a nexus between what I'm into and what I think other people will want to read. And I think we do the same thing with our, the products that we build. You know, we have this place, we have this place inside of everywhere. We've created this loop where everybody inside of everywhere is AI-filled. we're all using ai to do everything from coding to writing to editing to design uh to marketing all that stuff and if you're not doing it with ai first you're kind of weird which i think is the reverse of how things are in pretty much every other company and it creates this like special thing where i feel like we're all figuring out how to make an ai first company together like a truly ai first company together um and then we write about it and the really cool thing about writing is um it forces you to put words to your experience and when you put words to your experience you actually create a shared reality that allows you to um uh talk about what you're experiencing living in the future with the rest of the team and with the audience so we're kind of like bringing this reality into existence by writing about it um and then um when you're doing that it actually becomes really clear what you want it becomes really clear what's missing and so everything that we've built uh like every new product we've built has always been like hey i just kind of want this uh whether that's me or someone else on the team and the really cool thing about ai is you can like build that in a couple hours usually like a first version um and um you just kind of want it you build it for yourself if you use it uh and talk to the team about it there's a whole group of of early adopter type people inside of every who if we see something good we're going to want to try it so um if it's and then if it spreads to the rest of the team there's a good bet that the audience is going to like it and so then we you know release it to the audience so the incubation strategy or process is very organic it's very i just want this then we see if the rest of the team likes it and then we launch it to the audience and so are you saying that you internally said we need a better dictation tool than all the others that are you did no we did not no no no Naveen just like Naveen who's the GM of monologue just was like I don't know I like I want I want a digitization app the ones that I the ones that I've been using like aren't aren't quite right for me and he just like built it over a weekend and then demoed it and everyone was like oh this is really cool we want to try it and then do you sit and say how do we differentiate this how do we make this more useful nothing like that how do we take it to our audience our customers and see if any of them has anything that they need added no not really I mean we do think a lot about differentiation and, and, or maybe not differentiation, but we do think a lot about what is the idea behind the product and what is the concept and, and, and why do we like it and how does it connect?

8:04But that's also, again, like a more, a more organic process when you're, when you're building software this way, when you're writing this way, um, and you're making something that you want, you are a very special and unique person. Everybody is. And if you're making something that you actually want it is going to be different because it's it's yours and you may not know at the beginning exactly why it's different but like by by putting yourself and your taste into it it becomes yours and so what we do is for example Naveen built it and um when we started using it internally and then started to you know add beta customers to it because we have a list and we said hey we have this new thing we don't know if it's good but like you want to try it what we started to realize is certain people loved it and um some people on the team loved it and some some people on the um some people on the uh uh in the in the beta list on the beta list really loved it and um at that point we're that then it becomes like an interesting um psychological sociological like investigation exercise where you're like why and we realized that a lot of the people that really loved it either spoke multiple languages so like they spoke english and spanish or they were um switching between different contexts where they were speaking different versions of for example english um they're they're speaking product manager speak and then they're speaking gamer speak at home when they're on discord and so i looked at that and i was like okay it's the thing that makes it different is that it really speaks your language and everybody has their own language they speak and and the people that love this um and Naveen is like this too like he's he lives in India um but he's bilingual and he's constantly switching back and forth between his native language and English and it just so happened that because of that he made something really great and different um and uh but I think that's a thing that you do after you after the moment of creation to understand in retrospect what is the what makes this different versus I'm going to go out and do a bunch of research and like figure out all the problems with, you know, dictation software.

10:18And then I'm going to fix them, which is another way to do it. But I think is just not our way. No, I've been reading Rick Rubin's book. Have you read it? He makes that case very strongly. He says first inspiration, then you, then your audience. And that's been a hard thing for me to accept because I always felt like as business people, our reason for being is to service an audience, to almost put our customers ahead of us, right? To not be the grocery store that's selling celery because you like celery, but to say, all right, I'm even a vegetarian, but everyone eats meat. I'm going to have to go beyond, but that's not necessarily true anymore.

11:00I think it also depends on when and how you're doing it. Like, I think that we serve our audience by thinking about what we want and by giving ourselves what we want, because we're in this interesting way, living a little bit ahead of the people who are reading our stuff and using our products. and so by really being in touch with who we are and what we care about uh we're doing the work for our users who don't have to think about that so much they just get the end result of that process um there is a point at which you know you have a product in market and you're like you know you just get a ton of complaints people being like i need i need quora which does email i need to like use outlook and you're like okay like obviously we don't use outlook but like obviously we should do that you know that's a service to people but i think the um in the earlier part of the product building phase when it's a little bit more of this like kind of creative alchemy it's it has to be uh it doesn't have to be but i i think it's best i like it best um when it becomes um the creative alchemy is about what who you are and what you want i have heard you say several times that you live in the future you see what you need and you build it for when other people come there.

12:17I'm wondering, like, how do you live in the future beyond the programming tools, which I've seen you talk to other people about? What are the things that you do personally with AI that people in the future will all take naturally as part of life? Every part of my life has been, you know, I don't know if AI thought is the right word, but it is involved in every part of my life. So really simple, quick example. I was in the in a coffee shop today and I was journaling and just sort of like thinking about you pass some cool milestones like yesterday I was at dev day opening a dev day and I got this like cool plaque that was like you've processed 10 billion tokens through through open AI which is it's so cool um and it just I like to just spend some time like soaking in moments like that because it's if you don't enjoy it what's the point um and I was thinking a bit about how that felt and what it meant to me and um there's some positives to that and there's some negatives because like i immediately am like you know i hope xyz person saw this because like i have a chip on my shoulder about you know whatever uh people who may be underestimated estimated me or whatever like and so i i i mean i don't want to i don't want to get into all the personal was it an investor or what type give me the there's plenty there's plenty of people there's like there's uh there's sort of investor people who are like oh you know you're like running a newsletter when you're to do something real with your life.

13:45Um, or there's like, you know, people I've worked with in the past. Um, just, there's just a lot of people who like sort of underestimated me. Um, or maybe not me exactly, but I think we're judgmental of you're running a newsletter, why you're not ambitious enough and whatever. And so I have a little bit of a chip on my shoulder about that. And I was thinking about that. Like why, why is that my, one of my first reactions is to like be like fuck you you know uh and there's some good things about that you know um there's like that that expression that's like chips on shoulders like put dollars in pockets or whatever i don't know there's something like that um uh but also it says something about me that that that i go there and one of the things i love about writing is it becomes an exploration of self where you understand yourself in a new way um so i was writing about that took a picture of it threw it into gbd5 gbd5 is an incredible handwriting um recognizer ocr you know character recognition expert um because i have like this serial killer chicken stretch handwriting i write in like red flare pens that's like a that's like a magic note and um it yeah it just pulled out some patterns and i was thinking about okay how does this relate to my family life and how does it relate to just like who i am as a person and it was a very good sparring partner for me to um pull out those pull out those reflections and and then and connect it to other things that we've talked about for example because it has a memory um So that's a stupid, very simple example.

15:34Another simple example is I read a lot. I love reading. I'm reading this. One of the things I'm reading is Kierkegaard's The Sickness Unto Death. I mean, Kierkegaard is such a funny guy because he's obviously so depressed. But he's also super, super smart. And The Sickness Under Death is an examination of the state of despair and actually the good things that can come out of the possibility of despair as human beings. And it's also just impossible to fucking read. It's so bad. and I got kind of interested in it because one of my favorite writers this guy David Milch was like really into it so I was like I'm gonna read some Kierkegaard and then I just you know I take a picture of the first page of Touch of P and it kind of helps you helps me to understand okay what is he actually saying here and why is he saying it why is this important and once I have that like handhold on it I can just sort of read it but um you know it's like If you read sci-fi, the first three pages, it's like a totally new language of all these characters and all these places.

16:52You're like, what the fuck is going on? And I think ChatGPT does a really good job of translating, again, the different forms of English, translating from Kierkegaard's, he calls it dialectical algebra, Kierkegaard's dialectical algebra to normal people speak. And that helps me get much more out of the stuff I'm reading. No, I saw you do a video on that where you have ChatGPT's voice system set up while you're reading a book so that you can just ask it questions. Doesn't it piss you off that you can't do that with Siri? Like that's what it's built for, that you can't just say, hey, or even that I should be able to press a button on my Apple Watch and talk even to ChatGPT.

17:29And I have not found anyone who can set that up except for one guy who's punking people on YouTube. It's so irritating. It's very silly. Yeah, I hope Alvin gets it together. and still you're not switching over to android it seems no it would take a lot it would take a lot to get to get me to switch it's not enough that they have gemini built in everywhere no because i can use chat gpt on my phone you know um i think i think there's a there's a big status tax with the green bubbles you know uh for sure do you know how many people i get to interview because i take their email address i put it in the iMessage box and i send them a message and like, all right, I guess I assume that Andrew and I are text friends from a long time ago.

18:12And they start the conversation with me and boom, we're off the, without it. You know, when I, when I asked you how you were doing and I was looking at your face and you looked at me like so weird for why I was asking that way and looking at you and scrutinizing so much, it's because you read such depressing books. You read the kinds of books that either lead someone to kill themselves or discover a whole new way of thinking. And maybe, maybe we're on the right side of that. Yeah, I think definitely, A, for sure, I've gone through periods in my life where, like, shit was pretty dark. Right now, things are great.

18:45I love my life. And looking deeply at the hard things in life or looking deeply at the range of human experience is a really important thing for me to figure out how to live and figure out how to understand my own life. I've tried using AI as a therapist. It's like it's a 2 out of 10. have you gotten to the place where you can go to ai as a therapist yeah not in a way that replaces my therapist like i go to a therapist every week but i use it all the time for stuff just like this like kind of trying to figure out yeah i'm journaling about my you know emotional reaction to this thing which is something that you know i would definitely talk to my therapist about but i've seen him once a week on friday so i've got like four days until until my therapist is going to be there so at least I get a little bit of stuff from John GPT and I think it um it depends on how you do it and how you prompt it so you in a lot of ways you get out what you put in um I don't think it's it's it's not a clear it's not a clean or clear replacement but as a either for certain people like a gateway into therapy or for people who are in therapy um a therapist that is always around 24 7 who can read all your texts or all your emails or your meeting transcripts It's really good.

20:03Do you feed it all your emails and all your texts? How? I do sometimes feed it texts. You can export your iMessages. There are some GitHub libraries where you just use the GitHub library to export your texts. And then you can just throw it into ChetBt or whatever. Discord transcripts, too. We use Discord as our internal messaging service. And then Granola is actually quite good for this. granola just introduced this feature that um you can ask it questions about all of your notes uh all your meeting transcripts uh and i they have this like it's like they have this recipes thing and um it's reusable prompts from different people and i have a recipe on there that i that i made that i i really liked it that was really cool which is go through all of my internal meetings and write our company handbook from what you see.

20:58And I want you to capture all of the unspoken ways that every works as a company, both good and bad, and read between the lines to tell me, what are the different personalities? How do we end up working together? All that kind of stuff. So that it's the kind of stuff that you want a new employee to know, but it would never end up in an onboarding document because no one's going to take the time to write it down and like maybe some of it's embarrassing or whatever. It was pretty good. It was cool. I highly recommend doing that. There's also lots of stuff you can do with like, where was I avoiding conflict?

21:32Or what questions could I have asked in this meeting to get a better result or to learn more or whatever? It's really good. See, and this is where you're a much better marketer than granola. So they posted about that. What did they even post? It was like, we have these new recipes and then they had a screenshot of you and a cultural handbook. And I'm going, where? I click it. It just opens up the image in a bigger thing. Where do I even go to get this thing? And you, yeah. And the fact that you would think ahead also tells me another thing about you, Dan, that you really are a very introspective person.

22:06Like I've heard you say that you will go into Granola or other tools and say, show me where I'm avoiding conflict. You created a culture handbook that lets it read between the lines and identify the ways that you guys communicate that most people would not even notice. That's such an interesting aspect of your personality. And I wonder, does that come through in your software? I don't see that. That's a good question. I think that a lot of, like all the software that we built is geared around, it's like essentially productivity software. It's software that is designed to help you do better work and live a better life with AI.

22:48and i think that comes out of to some degree the introspection a certain kind of introspection which is um what are the problems i have with you know my email and like how do i wish it worked um and a desire to make the the friction in my own experience or our own experiences um feel better maybe that's part i see you definitely have like the opposite of schlep blindness for other people you see the schlep and you see it and you and you feel it and you find a way to do it and that that tim farris version 1.0 where he was looking to find a way to like get you to not spend so much time doing busy work by figuring you got that all your software has it i feel like the heart maybe is what's missing monologue has it like even when you were talking with what's his name again uh uh naveen naveen you're talking to naveen there was this sense of i know that sometimes people say things then they backtrack and then they change midway through the sentence and they complete it right i'm gonna feel them and i'm gonna get it but take a look at spiral that's the one that i was i had the most hope for because you're so good at actually lex is like still part of the company it is not no it's a separate company okay so nathan took that and he's spun it off.

24:08Okay. So Spiral is the one that is the most your company in the sense that you are such great creators. Spiral takes a transcript and spits out a very generic feel of a post. You're nodding along and I appreciate that. The part where I think it could go a little bit deeper is I remember talking to Scott Olford where he was going from like internet marketer to this guy who's crying on the internet heartfelt and then telling you what he was doing great was when he saw a therapist for the like maybe not the first time a good therapist who said tell the world what you're telling me like you want them to see you tell right like i would love it if it did some kind of introspection and it made me realize that i have a chip on my shoulder and that it's good or bad and ask me those probing questions and then create the thing that's uh that is great to hear because that's exactly what we built um it is coming out of beta timeout this is andrew breaking into the interview i'm looking at the edit and i've got to tell you that there's a new version of spiral out and i'll let you decide what you think of it but i will point out one thing that they said immediately in their launch post it's an ai writing partner with taste and so the point that i made about it not having his taste and sensibility his style well the new version should let's get back to the interview it's essentially that um basically what we realized is you know spiral comes from that same kind of insight um where i realized that i was writing a lot of posts and then or doing a lot of podcasts and i needed to tweet and like there's a specific format that I do that like seems to work pretty well.

25:54And I want to spend more time, like I want to spend less time just doing the rote translation and more time either making the thing better or doing something else and something new. And AI is really good at taking something from one channel and turning it into a format and whatever for another channel. That's where, that's where it came from. Same sort of like introspection about schlep. um and uh what we realized is there's a lot there's a lot more part there's there's many different more parts of the creative processes and they're just that like narrow translation layer or translation step and to make spiral something that we all we wanted to use all the time every day and therefore other people we needed something else and over the summer i I started using cloud code and I was like, wow, this is the future of programming.

26:43And they did something really special here, which is they just eliminated the text editor. You don't need the code editor. You just need to talk to cloud and the CLI. And so I started talking to Danny about this and for the new version of spiral, it's like it's an agentic ghostwriter for your short form content. And so the question is, so it helps you, it still helps you do the same thing but um it's more of a ghostwriter and less of a uh sort of like translation machine unless like a zapier type thing and one of the core insights there is okay how do you write well for someone else how do you how do you how do you be a good ghostwriter and good ghostwriters uh when you say hey i want you to tweet about this or blog post about this ghost good ghostwriters don't go ahead and say great i did it here it is um what they do is they say, okay, well, what was interesting to you about this article?

27:40What was, what did you like about this podcast? Was it this or this? Um, and then you can be like, ah, it was kind of like, I liked this, but like, there's also this other thing that you missed and you can kind of go back and forth and put your, put your finger on, put your, yeah, put your finger on like, what is it personally to you? Why do you, why do you like this? Not just generically, right? For example, in the new spiral, we have this, uh, interviewer agent that when you, uh, ask it to write something for you, if you haven't given enough context, the interviewer agent kicks in and is like, okay, Andrew, I want to understand you and why you care about this.

28:18And, um, and that helps it when it, when we get to the right step, it helps it write something that's more for you than someone else, because like someone else would have said something totally different and that's the um that's like basically i agree with you and the new version is imbuing a lot of that sense into the product so dan part of what gets me interested is in the process to go from zero to version one and then from version one to version two and i'm wondering how you had that realization that you need to do the interview after you after you get the uh the original content how do you get to that?

28:56I was just thinking about how I do a lot. I've done a lot of ghostwriting in my life. How would I do that? And how does that differ from the way AI currently works? And not just that, but like what makes a great ghostwriter? Like how do I do something that ideally puts my finger on who someone is and how they talk and what they actually think in words? And the first step to that is to make someone feel seen because you're actually seeing them. And once you're grounded in who that person actually is, then you can do something that feels more organically like them. And I think no AI tool really is doing that because they're very focused on just give me an answer right now.

29:39I do see that. And so people used to dismiss this type of app as just a wrapper because essentially I could go into chat GPT, create my own GPT that's like ghostwriter GPT, and I could do all this. I always thought that was ridiculous because you still need the interface you still need someone to think through the process you don't have to do it yourself is that is is my analysis of like the way people would have dismissed it fair and where we where we're going to end up also something you share yeah i'm very pro rappers um i think of ai as a technology that's a little bit like excel where um excel is this very general purpose tool that is very simple to start with, but has like endless complexity and was the first killer app for computers.

30:32Like most people learn how to use computers because of Excel. And what happened is it starts with like, you know, people doing financial modeling and stuff like that or accounting. And then it started to like grow out into, you know, pretty much any function of a business or any part of your life, you could like turn into a spreadsheet and people created these very um complex workflows inside of excel that became how they were running their business and that became the entire opportunity for b2b sas like of course you can do it in excel but it's not purpose built for that um and i think there's something very similar happening and going to happen with ai is people have to use chat tbt and like realize that this new sort of workflow is possible and then there will be a big opportunity to unbundle those workflows into purpose-built apps.

31:21I'm just sitting there for a second because I think you're 100 % right. I'm flashing back to an interview I did with Sujan Patel years ago. He was working on AppLoy. It's basically time tracking for teams. I think that's the one, or it was something that was essentially, he said, look, companies I see are using spreadsheets to keep track of their people's time. I'm going to create software that does it. And then what he did too was he recognized recognize that people were looking for Excel templates and Google Sheets templates for it. He goes, that's the best way to get customers to give them the thing they want and say, by the way, do you want a better version of this?

31:54And that's a man. And then once he turned me on to that, I realized a lot of the people I was interviewing were essentially creating alternatives to Excel that did all that. Yeah. All right. Here's the thing I admire. In addition to your writing, you are so good with your team. Like when you just get on a zoom with them, I see the rapport. I see the mind meld. And I know that I feel like when I asked you before we got started, if you would share with me some of the prompts that you all share that I feel like maybe I was picking something that seemed too small for you. It's like, Andrew, you just want to see how we keep notes.

32:26That's not a big enough conversation idea. But I still feel like there's a shared empathy and shared awareness that comes from the way you all share ideas like that. I would like to see some of the way some of the prompts. Will we be able to give that to listeners? Yeah, absolutely. So, you know, one of the things that we do is we have a lot of things that we do around prompt sharing, specifically for editorial. So like copy edits, because, you know, we have a daily newsletter. We also have like four different apps where we're publishing stuff all the time. We need, we have a pretty high bar quality and we need to make sure that everything that we publish like has gone through that filter.

33:09So we have really cool copy editing. Um, but basically we have a cloud project with all of this stuff in it. Um, uh, that can do copy editing, can do developmental editing. So like helping you figure out like which ideas are interesting and why, or where's the hook for the article, that kind of stuff. Um, and I'm happy to share that after this, uh, for, for your audience. And so that's how you do it. You share it. I thought maybe you had a notion doc or something, but no, you do it in a cloud project. and then you know what i haven't used cloud projects i use chat gpt for things like this and my my frustration with it is that i still have to have somebody go in and upload the doc with the new instructions remove the old one there's no like i wish i could just point it at a notion doc and say hey everyone on the team you can go add to this and hey ai tool that we're using just read it and you can't do that but with projects can anyone on the team edit them and change them and add to understanding no and that's a problem i do think that we're getting to a place where claude claude is very very good right now at you know longer horizon agentic tasks and so i would and it now also has file use so i would bet that we're getting close to if not currently at a place where if you if you pointed it at a notion page and said read the notion page for all of our copyediting rules and then I think it might be able to do it but it's like it just got there I was you know because I spent so much time in Dan Shipper world this morning and I was thinking I bet that's possible I have to do it and then I said no I have to stay focused on this I will try that afterwards and see if that works I love it follow your impulse speaking of uh people I've the reason that Naveen was so connected to you is he'd read your stuff he was if I don't know that a fan is a way to do is is a way to really explain it but in the same way that the that the reddit guys were were reading paul graham and they said okay i like the way he thinks we are aligned and i want more of this thinking he was the same way with you and you had time with him you told me something that surprised me that you are now hiring people that you don't really know that well or working on the company and in addition to hearing we are the almost bootstrap company i've also kept hearing over and over again we are the fewer than 10 people company and so are you more than 10 people how is this working now and how are you keeping the culture going we are more than 10 people um we are about 15 between 15 and 20 ish it's a little hard to tell full time um we have sticks business lines now so um yes a lot uh how am i keeping the culture well one of the really cool things is well you know one is if you have a really solid first 10 people the culture starts to self-perpetuate but i think we also have another really cool advantage um which you you can tell from that new bean story which is everybody who works at every essentially starts out as a reader and if you're reading uh or watching your videos or anything like that you probably have we call it every smell you probably have like an every smell um you're probably pretty curious.

36:29You probably love AI. You're probably in some way multidimensional. So everyone at every like, you know, Brandon, who's our COO, he was like an architect, but then he ran an insure tech company and now he's our COO. Or like Kieran, who runs Quora, he was a baker before this, but he also was like, you know, CTO of a startup. So like he just, he was also a composer. So, you know, that sort of multidimensional quality. Um, and so, yeah, we're starting, like I was telling you, I just had a meeting with a designer on our team who's super talented, who just joined that, uh, he's been working as a contractor for a while, but he's started full time.

37:09And I just, um, I just met him for the first time and that's like a new experience for me. And so, um, I am actually currently working through, okay, like how do I, how do I run an organization of this size? How do I keep it feeling the way that it feels and keep it feeling like a playground and also really understand what's working and what's not and who's involved and how are they working, even if I'm not working with them every day? So like TBD, I'm in process on that, but we definitely have reached a new stage of company building that I'm learning how to do on the fly. I see. It seems like the way the people are going to progress in your world is they read you on Twitter some of your free stuff.

37:59They then say, OK, I want to read another article. I want to try this software. And you could read the article and just subscribe to the one thing or just the software. But maybe you subscribe to the whole thing and you're very likely then to end up on Discord, which your dad even loves. How did you get your dad to love Discord? Impressive. and then from there you're really in the world of other people who are thinking the same way do you do you have like a you're nodding for people who are listening do you have a way of articulating what it is that you stand for what the smell is like i mean definitely i use a lot of the words that i just i just said sort of um curious earnest multi-dimensional um ai but you haven't yet written it down wonder actually that would be interesting to have to a certain extent though I think it is maybe not an explicit thing, but it is implicitly in everything that we make.

38:51And that's almost more important. I once saw a biography of Hugh Hefner back when everyone admired Hugh Hefner. And the interesting thing is early on, he decided that he would write down the playboy culture so that everything would come out from that. And so when he would have someone write an article about speakers and there was a sense of elevation about it it didn't come because the person happened to love speakers and and just wrote it it's because they were encouraged to feel this way to care about the finer things in life like that um anyway that's why i was asking about that yeah let's talk about smuggled intelligence um you're you're still thinking about this post that you put up what is smuggled intelligence i think that there's there's a lot of conversation right now about uh is ai gonna take our jobs and the way that we one of the reasons for that is we're starting to do benchmarks which is important but um we're just starting to do benchmarks like there's a there's a um benchmark from open ai called um gdp gdp eval or gd gdp val which is um basically they took tasks across a bunch of different industries and then um had experts uh uh judge humans versus ai on how well they did those tasks like a wholesale sales analyst yeah wholesale sales analyst or yeah there's a couple there's a there's a bunch of them across a bunch or like a you know a investment banker doing a presentation on an acquisition like that kind of stuff they found that gpt5 is better than experts like blindly judged better than experts at completing these tasks like 40 percent of the time and I think Claude Opus is at 49%.

40:43And that triggers a bunch of headlines about, okay, it's like AI is going to take all of our jobs. It's already better than experts 40 % or 50 % of the time. And I actually think that these benchmarks show in a lot of ways the opposite, that it will certainly change our jobs, but it may actually create more jobs than it. then it takes them away. And the reason for that is jobs don't, when you think about the way these benchmarks work, a human being had to be like, okay, we're going to measure, we're going to, first, we want to measure this. We want to see how well, how good AI is at doing real world tasks, which in itself is like a hard thing to just come up with on any given day.