Transcript: ‘He Built AI Agents to Launch a Million Businesses’

The transcript of AI & I with Prehype founding partner Henrik Werdelin is below. Watch on X or YouTube, or listen on Spotify or Apple Podcasts.

Timestamps

1. Introduction: 00:01:33
2. Dan and Henrik on the new breed of entrepreneurship that AI makes possible: 00:02:50
3. Why Henrik believes the future belongs to 1 million million-dollar companies: 00:11:08
4. How to build “relationship capital” with your customers: 00:16:14
5. Why “customer-founder fit” shapes lasting companies: 00:21:35
6. Everything Henrik learned about himself from a decade of building companies: 00:23:01
7. How Henrik finds focus and meaning in the daily chaos: 00:31:44
8. How Henrik is parenting two kids in the age of AI: 00:34:17
9. The way AI can fix what social media broke: 00:50:33
10. What happens when AI agents become part of how we tell stories: 00:56:59

Transcript

(00:00:00)

Henrik Werdelin

One thing I was thinking about the other day, I’m doing a keynote for a summit, and so I’m trying to find themes. So, you know what’s an interesting theme that I don’t think that a lot of people talk about is part of portfolio entrepreneurship, which is kind of what you’re doing now, which I guess we did at Prehype, I think a lot of people intrigued about, but also it’s been, because so much of the conversation about entrepreneurship has been defined by VCs telling people what real entrepreneurship is, portfolio entrepreneurship always been kind of like a bad thing.

Dan Shipper

I totally agree. VCs have kind of given it a bad name and there’s this thing about oh, you have to focus and so having a portfolio sucks. And it’s well, you’re the VC, you have a portfolio.

Henrik Werdelin

Exactly. You should work 24/7, except you take eight weeks of holiday in the summer.

Dan Shipper

And one of the things I admired a lot, or admire a lot about you, and I think people should know there’s a lot of you created this incubator Prehype and there’s a lot of spirit of Prehype in Every. Every came out of Prehype, and I’ve kind of shamelessly ripped off certain things that I saw you do.

And one of those things is, I looked at what you did every day sitting at Prehype. You were running Bark, which is now a public company, and you also had all this other stuff going on, which you’re absolutely not allowed to do. You’re not supposed to do that. And I was like, in my heart of hearts, if I could be Henrik and have this sick venture-backed company and also have this incubator that I run that’s separate, and like writing books and doing whatever it is that you were doing, that sounds like a pretty good gig. And it only took me like six years and I’m definitely not close to your level.

Henrik Werdelin

I’m sure you’re eclipsing soon.

Dan Shipper

I have definitely moved more in that direction. It’s quite a good life. So portfolio entrepreneurship—great idea.

Henrik Werdelin

You know what? There’s a lot of things that kind of, I guess now that I’m looking back over the years 2010 to 2020–2021, which were kind of a bit of a golden age for, I think, for New York entrepreneurship. And I think some of us were kind of lucky to be in the midst of it all. There’s a lot of components of that that I’m now trying to understand what actually happened. And there are these kind of things that we don’t have terms for, right?

Some of the things that I’m exploring now—everyday entrepreneurship—so all these people are trying to build businesses that will never become unicorns and never probably will get venture funding. There’s these kind of things like the in-between time, which I think is this interesting thing that entrepreneurs have. Because most people, if they’re a lawyer or a developer, their next job is as a lawyer or a developer. But as an entrepreneur, your next job is definitely not going to be what you used to do, right? Because you’re probably tired and burned out of it. But so much of your personal narrative is linked up in what you do. So you’re kind of in this weird waistline in the middle where you used to do something and now everybody’s like, what’s next? And you don’t really know. And so there’s this beautiful, very vulnerable time, which is this in-between time.

I guess the last thing that I’ve been thinking a lot about is this notion of how do you create environments where amazing things appear? Specifically, how do you do it now when people are increasingly remote? And I visited your studio the other day and kind of got to think about like sometimes you actually have to create an environment, like a physical place where people go and they jam and they feel inspired and stuff like that. So yeah, I’ve been very reflective of that time too.

Dan Shipper

It was a great time and obviously the Golden Age ended as soon as you moved back to Denmark.

Henrik Werdelin

I’ll take full credit for that.

Dan Shipper

New York just not the same.

Henrik Werdelin

Yes, yes.

Dan Shipper

But maybe we’ll be back. Maybe it’ll come back at some point.

Henrik Werdelin

I’m sure it will. That’s actually the cool thing about, I think every time I visit New York now, it is a little bit like a college campus, right? You can definitely see there’s a new class of people, like there’s a new crew and they’re figuring their thing out, right? And building their kind of places. And that’s what’s so amazing with that city.

Dan Shipper

Definitely. I love it here. There’s definitely a lot happening and I hope to be one of the people building that crew. But I want to go back to the portfolio thing for a second. Why did you feel like you were allowed to do that? And how did you even pull that off?

Henrik Werdelin

I mean, I think as many entrepreneurs, it wasn’t there. So I kind of just made it. I don’t know how I pulled it off because I think even more so back then it was considered kind of a bad thing to do. And I’m also trying to figure out like what it is that I did. And I think one of the insights for me is that I don’t really think like people who put out a goal and then work towards it. I’m more kind of exploring things. And so I kind of find stuff that I find interesting and then I look at the projects that I’m involved in. I go, which, where would this fit? And so it’s a little bit kind of a roundabout way of working. I think honestly the way you get away with it is that you add enough value to the projects that you’re involved in so that the people, the other people involved, don’t throw you out.

Dan Shipper

Or you convince them that you do.

Henrik Werdelin

Exactly. Like very good jazz hands.

I think most of my partners, like on the good days, will kind of say that they’re very happy I’m involved. On their bad days will go like, how does this work again?

Dan Shipper

It means you struck a good deal. Yeah. I actually think, you know, when I think about the overlaps or why I ended up wanting to copy some of how you’ve built your career, the desire to be curious and do a bunch of different things. Like some people get exhausted by doing a bunch of different things and I am just, I just want to be doing lots of stuff. And then some of those things turn into larger things, and I just apportion more time to them as they become larger. But it’s a very bottoms-up exploratory process. And I think the more standard startup thing is a little bit more like what a Peter Thiel or a Keith Rabois would do.

You know, you just like set a big, you call your shot for this big goal and that’s the only thing you’re focusing on. That’s certainly another way to do it. But I think the bottoms-up, exploratory startups as a curious way of life is a legit path to entrepreneurship. And it is actually particularly a thing for media entrepreneurs. I notice a lot of media people have this because if you want to, if you’re writing stuff all the time, it’s because you’re a curious person. And so I think you may have, well, you have a media background too, so you have a little bit of that media thing about you too.

Henrik Werdelin

Back to the point about having kind of only one way of describing things. I do feel that there’s many different flavors, many different pedagogy of entrepreneurship. And it’s almost like we have told ourselves there’s only kind of one way. Or maybe there’s a few, right? There’s the way that they teach you at a business school about finding a big market and go after that. And maybe there’s the lean startup, Gary Vee had the hustle. You know, there’s like different. But I do think that there’s many, many different flavors of entrepreneurship and we probably don’t have a very refined vocabulary of how to describe some of these things. And I think what you’re articulating here is maybe the media kind of background approach to entrepreneurship, which works for some people and doesn’t work for others. And that’s fine too.

Dan Shipper

100 percent. And I think now that’s a really logical segue into you are thinking about new shapes of entrepreneurs, specifically in AI, which I actually think the media stuff is a particular shape that is now newly possible. And we’re sort of figuring that out. But you have another set of entrepreneurs that in the company that you started, Autos, and in the book that you wrote, you’re thinking about how they can start companies too, and what that looks like. What have you learned?

Henrik Werdelin

It’s a moving target, right? So I’ll give you, I know you had Nicholas on a bit ago, my co-founder at Autos, and we’re about to release our version two of the platform, which I can go in a little bit of detail of what we learned. And I think some of it is akin to, aligned with some of the thinking you have. I guess like on a few kind of parameters. First, I think entrepreneurship is increasingly being democratized because the entrepreneurial technology toolbox is increasingly being democratized. And I think it really started back from when the cloud kind of came about and then SaaS kind of made it even easier. And now with AI you can no-code and—

Dan Shipper

It’s so funny. I thought you said Claude, not cloud.

(00:10:00)

Henrik Werdelin

Yeah, yeah. I did say that and I realized, listen, my brain is completely going to shoot. I know I did say cloud. It’s like, ugh. Anyway, so now, you know, most people can do stuff and 60 percent of Americans say they’d like to start something and only 8 percent do. And historically I think there’s been a good reason because it required resources and required specific capabilities. And now I think there’s increasingly less of an excuse. And so the area that I’m very interested in is basically how do you make not one person make a unicorn, but more how do you make a million people make a million dollar turnover business, what we call Donkey Cos.

And so what I’ve learned is that first, that we’re pretty early in the phase of inventing what agentic businesses are. I’m a little bit reminded of when I did a fairly large startup back in the two thousands in the video space. This is before YouTube. And we raised a bunch of money and everybody kind of knew that video would go online and then we raised all this money and then we were like, hey, wait a minute. Nobody kind of knows how it’ll go online. Things that are now very basic were just not invented. What will an electronic programming guide look like on the internet? Will the video auto-start or will you have to play it? Will the next video then start? Or should you stop? You know? And all those different things that now we take for granted are things that we had to invent. And I think as we’re building Autos right now, we’re realizing that a lot of the agentic frameworks are just not invented.

Take Claude, for example. What should be the name of a bot? Should that be a human name like Claude? Should it be a utilitarian name? Should it be a brand like Perplexity? And with each of these small decisions we take, there’s a lot of kind of thinking that has to go into it. And I think we’re just kind of chucking along and trying to fix this up.

Where we are right now is we still have the ambition of launching hundreds of thousands of businesses a year. We’ve had our first cohort of a few thousand people come through and tried to help them make a business. And I guess what is new and interesting for you, I think one thing that we’ve totally gotten convinced about, which I think you have too, is that there won’t be that many one-agent businesses. I think increasingly we believe that a founder will serve a customer and then they will have multiple agents that will have to be part of this portfolio of tools that they offer to their customers. And so the portfolio thinking that we talked about earlier, and I think you’ve been writing about also, I think we are kind of now subscribing to. So a founder will come up with a customer they have to serve, they’ll come up with an initial problem that they can solve for them, and then they’ll create multiple, basically agent businesses or solutions for that specific space.

Dan Shipper

That’s interesting. And I do think that that’s one of my learnings about, let’s say portfolio entrepreneurship that I had to figure out by myself. I couldn’t just rip this off from you. Well, you obviously have come to this conclusion based on what you just said, but serving the same customer feels actually like a very important part of portfolio entrepreneurship.

Like with Every, the thing that makes it work is that the people who come and start companies with us or build products with us are from the same pool of people that we’re serving, that are our readers and our subscribers. That the tools that we’re building serve those same people. And so the ecosystem builds and becomes this cohesive thing where when we launch a new tool, hopefully everybody in the community likes it because we built it for ourselves. And a lot of the people in the community are kind of like us and we’re just sort of building this ecosystem of people who are all on the same page vs. I think the mistake that I see a lot of people do when they come to me and they’re like, I’m going to do an EIR thing, is they’re running one business and they’re like, but this is going to be in a totally different area with a totally different customer that I don’t even understand.

Henrik Werdelin

You need to compound.

Dan Shipper

Yeah, you need to compound. And if you do, it’s like every end thing you do just gets better and better. You know your customers better, they know you better. And so you’re building one thing instead of many things. Even though it looks like many things, it’s actually just one thing. And I think that’s one of the things that people miss about this kind of entrepreneurship and what works.

Henrik Werdelin

I mean, like the whole book is the premise of what we call relationship capital. And what we argue in the book is that it’s going to be one of the few, if not the only moat that’s going to be left in a time of AI where everybody can do everything. And so, I mean, like if you look at everything that came out of Prehype over 15 years, it’s basically these relationship capital companies. It’s companies that are not defined by what they do, but who they serve. And you take BarkBox, a good example. We never defined ourselves as a company that puts stuff in boxes, because then our next business would’ve been the cat box, which many people asked all the time. What we were about is we were about making products for dogs and the people who love them. And so the problems that we saw were problems for dog lovers and their dogs. And so our next big business was an airline for dogs, not another stuff in a box business.

And you take Ro, which obviously also came out of the Prehype world. They basically solve medical problems for men who have something they’re a little bit embarrassed about, being hair loss or ED or weight loss. And so I am a full subscriber of that. The pedagogy that I very much understand is you can’t start with an idea. You can’t start with a technology. You can’t start with a market size. You have to start with who do you want to serve for the next 10 years, and what is a fundamental problem that they have that you feel that you can solve.

And then from that, we basically think that relationship capital has these three areas and we can go into them in detail if it’s interesting. One is basically depth, which is how much does a customer feel seen by you? And I think you articulate very well. People feel very seen when they know that you’re kind of a little bit like them. But it is also like, do you respond to their posts? When they comment on something online, stuff like that. How quickly do you answer their emails?

The second thing is density, which is really how much do they feel that you belong in the community that you serve? And I think same point again, people know that people in Patagonia or Peloton, they’re part of the community, right? Sometimes they are even the community.

And the third part, which I think is going to be very important for portfolio entrepreneurship and for companies in the age of AI is basically durability. And that is, are you allowed to offer more than the initial thing that you offered? And some brands do this very well. You can imagine what a Nike hotel would look like, but it’s very difficult to compute what a Hilton shoe would be, right? So some companies and brands just have like very clear permission, authenticity, authority to serve their customers in different ways. Some don’t know.

Dan Shipper

I think that’s really good. And my add to that, my build is the way that that has worked for me, you know, because I think where you started here is you have to identify a customer you want to serve for 10 years, is instead of knowing who your customer is, know who you are. And that’s not just a, I’m going to journal about who I am, although that can be part of it, or I’m going to go to therapy, which can be part of it. But it is a dynamic process of building things, putting things out in the world, and that will tell you about who you are and what appeals to you and what you like. And you have to just iterate through a bunch of things to make that picture clear. Because a lot of it sometimes it’s not very flattering. It’s not actually what you want, or not actually what you think you want to be.

So for example, for me, I think when I was at Prehype, I really wanted to be a writer. But it was very hard for me to admit that because I had a whole founder identity, and I felt like it would interfere with that. And it took me like three years getting into Every, like for three years the whole thing almost fell apart, where I was like, what am I actually in this for? It’s like, I really want to write. How can I organize my day so I’m like mostly writing and then, or at least half writing and then operating? And how do I make a business with that? Because I also do really love business and I love technology, all that kind of stuff. And so as I’ve gotten clearer about who I am, the products I can build and the company I can build becomes a much clearer, solid idea that has staying power. Because I can be like the force or energy behind it in a way that’s different from I’m just putting on a show so that I can like raise the next round for the next year or so. And I don’t really love banking for, you know, I don’t know, male, former male people or whatever. I’m just making it up. I don’t really love whatever I’m doing, but I’m just going to put on a show. And that’s actually, it’s hard to, at least for me, it was hard to learn how not to put on a show and just do the thing that was most reflective of me. But that’s the, I think that’s the game.

(00:20:00)

Henrik Werdelin

I think the other reason why that is so important is because if it’s not a natural customer for you to serve, if there’s not what we think of as customer-founder fit, then you don’t have authenticity and authority to do it. And then I don’t think that people will really buy into the solution that you’re solving. And I think that’s going to be even more important again, when people can do everything. What do you, again, I think I sometimes hear people come and go like, oh, should, wouldn’t it be cool if you built whatever? And I’m like, yeah, that might be cool, but not for you. I mean like that’s, it doesn’t seem like a business that you should do at all.

In the book, we have this framework we call the Five Ps. It’s basically powers, passions, possessions, positions, or potentials. And these are five ways that you can basically look at yourself and figure out where do I have something that is innate in me that I can extract from, and then build something around that. And so I think different people come to that conclusion different ways. I think you had like your journey and I had mine, but I will echo that every time that I try to build something to make it successful, I end up not building something great. And every time I basically go for, I want to build cool shit with people I like, then I end up doing something that seems to resonate with people.

Dan Shipper

What have you learned about who you are in the process of building companies over the last decade?

Henrik Werdelin

I have learned that I have a weird role where I find people who are very talented and then I see something in them and then I help them find that. And in many ways, that role is kind of like a weird supporting actor role. I kind of had this mental model because Prehype was called the studio. Then I went deep and kind of studied like music studios. And if you really study the Motown of the world or the one in London, I’m just blanking on where Beatles had their stuff. And there’s one in Sweden where Max Martin kind of did a lot of like the 2010s and 2020s pop music. What you see is there’s like a few people there that kind of like get these uncut gems in and then they help them find the baseline, they help them kind of cut the music and then they create this magical stuff. And so the amount of like successful musicians that have come out of like Motown or Evergreen Studios or the one in Sweden is just incredible. And so I see myself, I think I’m learning that that is kind of both what I enjoy, but also what I seem to be good at—finding people and then helping them figure out what should be their thing.

Dan Shipper

What else? Give me something deeper.

Henrik Werdelin

Oh, something deeper. I guess the thing that comes to mind, which I think was most surprising, was that after we went public, I took a little bit of time where I didn’t do much and I kind of thought that I would, I guess borderline retire. And suddenly I felt that I was, I went to this very high-end conference and my feeling was that I had like a year where I still would be invited back to that. And so I got really worried of losing relevancy and I was trying to understand why I was so worried about it. And I actually don’t think it was about vanity, although there’s probably some of that. But what I really, really, really enjoy is to get to talk to people like yourself and other people who have new original thought. And if you’re not relevant, then you basically don’t get invited. You don’t have interesting observations to trade with, and so you just kind of wither. And I think that the urge and need to be part of that club of people who think about new stuff, it was something that I guess I learned.

Dan Shipper

I think that makes a lot of sense. It actually dovetails with something that I’ve been thinking about because you’re right, entrepreneurship is a weird profession where you expect to, okay, I’m going to work really hard. I’m going to work five or 10 years, maybe more, and then I’m going to get an exit and then at the very least, I’m going to take a lot of time off, but maybe I won’t ever work again. And that’s maybe part of the point. And a lot of people who have done that, you included, I’ve had this at a much smaller scale, we didn’t IPO, but you know, where you just get to the finish line, and then you get to the number that you wanted or the outcome that you wanted, and then you’re like, well, I have to live my life the next day, but my whole life is now gone. And that’s like really hard. And I was thinking about that. And I think there’s two ways of running companies. It’s startups is a means to an end, and startups is a way of life. And if you run a company as a way of life, ideally you want to run it or be doing the things that running companies entails for a long time without like a specific end date in mind. And ideally, you’re even if you sell one thing, your life is set up so that it’s a way of life. It’s not like I’m going to get to this end date and then do something totally different. It’s like I have set up my life such that I don’t, maybe I want to take a vacation or maybe I want to slowly pivot into something else, but I don’t need to get away. I don’t need to have that, like one final, like it’s over. I’m out, and then I could do something else. Because that’s, those are the periods that are, I think, hurt the most or suck the most for entrepreneurs or feel very unnatural.

Henrik Werdelin

That’s very much resonates with me and I think you’ve been better doing self-work. You know, the doing self-work for me is a newer thing. You know, I didn’t get a coach or a therapist until a few years ago where I had a lot of time. I was like, everybody else seemed to be really enjoying understanding why they are who they are.

Dan Shipper

Look, my problems were pretty fucking deep. It was not about enjoyment, let me tell you.

Henrik Werdelin

Mine was mine. You know, I don’t think I had the same thing, but I definitely, as I then worked with my coach, you obviously realized that these anxieties and these behaviors you have like something that comes from childhood and from different things and you try to understand that better. And I found that to be obviously sometimes very vulnerable, but most of the time it’s very, very interesting. And but I think in terms of what I’ve learned about being an entrepreneur, I’ve done it for so long that I just, I don’t really know any other way, I guess. I mean, I’m not sure anybody would ever offer me a job. So I guess I’m entrepreneur for life. No, sometimes people ask like, would you ever take a job? And you’re like, I don’t know. It depends what it was, maybe. But people like assume that you just want to be an entrepreneur. And sometimes when I hear like people, VCs have these amazing, I don’t know, grass is always greener and I’ve never been a VC, but sometimes you hear about them getting this ridiculous kind of management fees and goof around. And then I was like, that seems to be a pretty good job.

Dan Shipper

I was going to say like, when was the last, what’s the last job where maybe it wasn’t offered to you, but like you had to be, you were really thinking maybe I should do that instead of this?

Henrik Werdelin

I mean, this is when I did Prehype. I had just finished a company that didn’t work out and I was basically trying to figure out what to do next. And I was keeping all my options open so I got offered a job to run BBC Digital on the international side. And—

Dan Shipper

That’s cool.

Henrik Werdelin

I got offered. And then what I did do is I got offered a job at a kind of design agency called Wolff Olins who had an incredible CEO called Karl. And he offered me a job and I kind of like pivoted then into being the first project that Prehype had. And so you know, I kind of chose my path there.

Dan Shipper

Well, I am, I’m glad that you chose Prehype instead of head of BBC Digital, because I think my post would’ve been a lot harder.

Henrik Werdelin

The reason why I figured it out was I did this thing where I wrote down all the, what I call micro-moments. Basically micro-moments for me are things when I’m in flow and things where I’m happy and it can’t be kind of like big things like, you know, having a family. It has to be like very concrete things. Like I like walking over the Brooklyn Bridge in the morning. You know, it’s just, I get profoundly happy about that, right? Or I like being in brainstorm meetings with people like you and stuff like that. And so I ended up having these micro-moments, which is these 30 kind of concrete periods where I’ve done stuff. And then I started to use that as a way to measure options that would be thrown at me. So the BBC, it sounded like it was a lot of money and it was like a cool job and it would give me kind of like I guess self-esteem for a second. But then when I looked at what it would entail, none of the micro-moments would be included. And so I was like, ah, probably not for me.

Dan Shipper

What are your top micro-moments and how have they changed as you’ve gotten older?

Henrik Werdelin

I think it’s probably after I got kids, because now I didn’t like kids much before I had them. I was like, ah, they’re annoying. And I think even when my wife was like, we should really look at this kid stuff, I was like, ah, why? Maybe. But now I have a five-year-old and an 11-year-old and it is just really, really incredible. So a lot of the moments now are just re-kind of re-exploring the world with them, through them. And they ask all these interesting questions that you don’t have the answers to— What’s a good one recently? Oh, why do some people get very popular at school and some does not?

(00:30:00)

Dan Shipper

That is, I’d like to know the answer to that.

Henrik Werdelin

Oh, what’s going to happen if I won’t be able to afford a house the size that we live in? Then I won’t be anxious. Just then I want to love you. And then that was it. Okay, obviously not like your brother better. No. But so I think some of the, a lot of the micro-moments around that, I also, through my twenties and thirties, maybe part of my early forties, I was so anxious to succeed that I don’t remember a lot of the stuff. And so I have so many micro-moments now because I’ve gotten a little bit more kind of chilled. And so I can actually really enjoy all the stuff that I do every day. And so I think a lot of the early years of even Prehype is just a little bit of a blur because I just wanted to make it and didn’t really have the energy to kind of stop and enjoy every day.

Dan Shipper

How are you dealing with your kids and AI?

Henrik Werdelin

It’s a good question. I mean, I have an 11-year-old, which are kind of just when iPads were kind of, and so we took a sabbatical with him and people were like, how is that with a six-year-old at the time? And I was like, this is great. And he had an iPad and there’s YouTube, so problem solved. And now reading what’s it called? Jonathan Haidt’s book, The Anxious Generation, I’m like, okay, that was probably not a good idea. So the younger one doesn’t have YouTube and doesn’t have a lot of the same thing.

Dan Shipper

Have you noticed a difference in their temperament and how they are?

Henrik Werdelin

Yeah, but I think it might be genetic. Like the young one doesn’t want to play on his iPad because there’s nothing interesting on it. And the older one is just completely addicted to it. So maybe that’s because of Roblox and YouTube that the older one had access to and the younger one does not.

Dan Shipper

What would be the personality explanation? What are they both like otherwise?

Henrik Werdelin

You know, I’m increasingly kind of buying into the thesis that a lot of this stuff is just super genetic. Sorry, it is all the genes. It’s all DNA. I think we, I think we as parents would like to think that we have so much impact on these kids. But I have two boys. They’re six years apart. They’ve gone exactly through the same stuff, right? Pretty much. And they are night and day. One is blonde, the one is brown-haired. And one is like a little clown that talks all the time. The other one is more of an introvert. They’re just like completely different. And I don’t think that we necessarily did that much difference.

So I think on AI specifically, I think AI is like electricity. I think it’s going to be a transformative technology. And I’ve drank the Kool-Aid and so I think they have to learn about it. And so they have full access to all the different foundational models and they use it a lot. They will have long conversations with, I think my youngest likes Pi the most and the oldest likes ChatGPT the most. But they have full access to it and I teach them how to vibe code and they know how to do stuff like that. And I’m kind of all bought into that.

Dan Shipper

And are you like monitoring their chats and if so, how?

Henrik Werdelin

Yes and no. We talk a lot about first principles and so we have first principles of how we interface with machines. And there are some of the elements. There’s just rules, like some of them are just basic rules. When you talk to somebody on the internet, you can’t give your real name, you can’t da, da, da, those different things. I sometimes do go through the chat history and just kind of get a little bit of understanding on it. Mostly with the youngest. The oldest is getting 11. He is kind of pre-puberty, teenager getting there. And I feel he deserves his privacy. And so there are things that I don’t want him to talk about. And I put rules on the network level, right? So I have something called Firewalla, which is a kind of firewall that you can kind of put stuff on.

Dan Shipper

I’m so glad you were not my parent. My parents have no idea what a firewall is. Thank god.

Henrik Werdelin

And my only, it’s going to be difficult to watch porn at this household. I’m telling you that. I mean there is a, they all know about VPNs and stuff like that. But yeah, I mean we do have a little bit of like a battle. Like for example, the oldest, he doesn’t have a smartphone. He has an Android, and it’s like completely locked down. It has black and white interface. Only you can, he can only use useful apps. And so sometimes, and I’ll get a notification every time there’s a new app that gets installed, even though I’ve locked it, so you can’t install new apps, but he’ll find ways to do it. And so we will have this like, how the hell did you get past that security measurement? He is like, oh, I don’t know, man. But he still hasn’t found a way to disable me knowing about it. And so he still gets caught when he does it. But I kind of like that little, that kind of like having like the competition of whether he can circumvent my security systems.

Dan Shipper

Little cat and mouse game. Yeah. I like it. What do you find they’re using it for?

Henrik Werdelin

I find them, they increasingly like me use it for most things. The six-year-old, five-year-old, he talks a lot. And so sometimes the family are just a little bit, just can’t be bothered talking to him anymore. So I think he just like talks to it as, it’s like somebody who’s talking and somebody asks you questions. Like he’s a very proactive guy, so he’s like, I’d like to come up with things to do this weekend with my family. And what could you suggest?

The older one is more philosophical and so he has like areas of interest. He had the Second World War, he was very interested in, he was very interested in Titanic. And so last time I checked he had spent like hours basically roleplaying that he was in the Second World War and that he just landed on the beach of Normandy and basically was kind of roleplaying himself into how that’d be. And he had what seemed to be a very, very lengthy kind of just using his imagination to be kind of, yeah, positioned that as he was a soldier in the Second World War.

Dan Shipper

That’s really interesting.

Henrik Werdelin

It’s super fascinating. I mean, he also used it for, he doesn’t use it for schoolwork that much, but they also, our school doesn’t seem to mind it. They kind of embrace it rather than kind of shun it. So yeah, he used it for that too. He is one of the things, he’s a little bit negative on it. There’s definitely a sense of, dad, can you just stop talking about AI stuff for a second?

Dan Shipper

Oh, interesting.

Henrik Werdelin

There’s a bit of that.

Dan Shipper

Is that because his classmates in the environment is negative on it or is it like a personal thing? One of the things I’ve been feeling about the AI wave is it’s the first tech wave that I can remember that 25 to 35 year olds seem psyched about it because it was created by 25 to 35 or 40 year olds. But usually it’s like the Mark Zuckerberg 18 in his dorm room and this is the first big one that’s not that. And I think that that has created a different sentiment among teenagers for it than is normal or is sort of the usual way it works. So where do you think the negativity comes from?

Henrik Werdelin

I think for him, I think it’s a little bit different from what I sense it is for the rest of the world. I definitely sense also like this two-tier world where some people really, some people are really into it and some people just think it’s really bad. Write a comment every time they see something that’s generated by AI online. I think for him, I think honestly it’s a little bit more just rebelling against me. I think it’s a little bit like, I’m just tired of you talking about it. That’s your thing. It’s not my thing. So I think it’s more him just being the anti-parent thing rather than necessarily AI.

Dan Shipper

Got it. Okay. That makes sense. One of the things that occurs to me is you talked about them asking you all these questions and that being delightful. How has that affected, how they can get an answer to anything immediately? So how has that affected what they ask you? And also just being able to get answers and go deep and just go why, why, why until you get down to like quantum gravity. How has that affected them and what they know about the world and how they think about the world?

(00:40:00)

Henrik Werdelin

My observation is that they just know so much. I mean, like my son will often tell me something that I just didn’t know about. We’ll sit at a restaurant and he’ll go like, oh, you know what’s happening here is blah, blah, blah, or this specific fruit. Or you know, and you’ll go like, what? How do you know about that?

I also think, honestly, maybe that’s going back to my childhood. I think they just talk much more with us than I talked to my parents. We are, we’re together a lot and we talk a lot about a lot of things. And so there’s that. I don’t know. I’m just very impressed by, I do tons of talks now, both at universities but also at high schools. And I’m honestly just very impressed by young people. They’re just very thoughtful and very, seem to be very knowledgeable about the world and seem to be very sophisticated about most, a lot of things. And so, I don’t know, like it just seems like evolution is doing its thing. We were like doing dumb shit because like we had nothing else to do and they seemed to be doing smarter things because I mean, like I grew up without a computer and the internet or any of those things, right? And so when you had like a weekend, and it wasn’t that my parents were programming my weekend, where these kids of course have access to everything and so they will want to have conversations about stuff and they’ll have good points. And so I’ll engage with them in kind of an adult way. But yeah, I think it’s just, it’s a little bit like it’s just two different species almost.

Dan Shipper

I think that’s what it is too. You’re watching humans become a different species. And that’s always, and that happens in every generation and that’s always uncomfortable for adults.

Henrik Werdelin

One thing that I’ve been thinking quite a bit about lately, specifically about entrepreneurship, but it kind of goes broader, is how much you should lean into your negative view of the world and how much you should lean into a positive. And I do think that people like yourself who have a lot of people who listen to you have like this choice where it’s not difficult to see some of the problems in the world. And it’s not difficult to see how, for example, AI can make things very complicated by job losses and all these different things. But it’s also not too difficult to think of all these positive things. If we talk about entrepreneurship, for example, like we have a country in the US where we’re very proud of it being basically born by entrepreneurs and entrepreneurship and the entrepreneurship that it was born by was not the MO, was not the unicorns. It was the mom and pop stores, right? Which we don’t talk about at all now.

We have this quite a golden opportunity where anybody who can identify a customer and come up with an original interesting way to solve a problem for them really has the opportunity to do that. That is pretty unique. That is incredible. And so if we believe that the future is, you know what, we make it, then, and we know that we made something wrong when we made the social web, right, which a lot of us were part of, then we can try to imagine something that is really good for a lot of people and then we can just try to go and make that and we can go and articulate that. We’d like to make it so that everybody gets inspired and do a little bit more of that. And in general, I think the world is currently pretty cynical and I think when it comes to my kids and it comes to how I would like to kind of throw energy into the world. I think it’s more about trying to find a way of showing where there could be light rather than where there could be darkness.

Dan Shipper

I totally agree. That’s what I try to do. And I think you make a good point about the social web. I actually think the reason, one of the main reasons why the default is negative is because the discourse has been so negative about the social web for the last 10 years and we’re applying a lot of the same feelings onto AI in ways that sometimes are fair and I think in a lot of cases are unfair because they’re very different. And it’s, it was this interesting, there’s this interesting swing, like when I was really first starting to get aware of the startup ecosystem and whatever in like 2010, it was so positive. The social network had just come out, the movie, and like everyone was like, this is amazing. And then in 2016 it totally flipped. And I think we’re still seeing some of that, some of that play out. And—

Henrik Werdelin

You have this interesting point about why AI is very different to the social web.

Dan Shipper

Yeah. Thank you for teeing me up.

Henrik Werdelin

Well that’s just because I’ve been like telling that story like it was mine for a bit. So this is just a little bit just guilt. I appreciate guilt, like just the guilt was coming over me. It’s probably because I almost was about to tell this story, like it was me that came up with it. I realized that I’ve done this once. I was sitting with Stacy, as you know, who used to be at Prehype. And we’re sitting at a, we’re sitting in Cannes at some festival and we are telling stories about airlines and I tell this story about how I had this kind of battle of the armrest. And then suddenly at one point like this person said, hey, do you mind? I have a prosthetic arm. And then I was very embarrassed and I tell this story like it was me. It wasn’t my story at all. It was Stacy’s story. And as I’m telling this story, I realize that Stacy’s sitting at the dinner and I’m like, oh no, I had to fess up. But so this is what was about to happen.

Dan Shipper

I’m glad that you caught it, because I actually have had that happen to me where people do my own bits back to me and I’m like, I literally, we talked about this like three weeks ago and you’re doing this back to me.

Henrik Werdelin

Oh, in all seriousness, I’ve given you full credit of that because I thought, but do, I don’t know, can you articulate it well? So I, you probably have like a more sharp version of it now, so I can use that for the next few weeks.

Dan Shipper

The very, very basic, and feel free anyone to rip this off, is the social web operated on revealed preference. So what you click on is what you want. And we acted as though revealed preference is all that people are, because that’s what you really want. And the reality is that you’re always going to click on a car crash. And so that’s where things go. They get more extreme. But I think humans are quite a bit more complex than that. And one of the beautiful things about AI is it operates based not just on what you click on, but what you say. And what you say about who you are actually, I think has, is real and relevant. And so, and I think you can see the difference in the ways that if I asked ChatGPT to make me a for you page versus my Instagram for you page, ChatGPT is super wholesome and Instagram is like the trashiest slop. And we may see that change over time, and I assume that AI will get a little bit more racy over time or a little bit more sloppy over time. But I do think that is a fundamental, it’s a fundamental difference to the way that these algorithms see us as human beings. Language models are much, much more multi-dimensional and complex. And I think that will make them healthier.

Henrik Werdelin

I very much buy into that. I want to buy into that because I want to see something that’s positive. But it is incredible. I mean, I was just like asking ChatGPT, what is my Myers-Briggs thing? And I realized that my Myers-Briggs has been debunked. It’s like not something that you should use, but it is incredible that it just immediately goes like you are yes, whatever. And it’s right, right. And so it does have a pretty good sense. Yeah. You know what I literally just posted on LinkedIn. I’m ESPN, I think.

Dan Shipper

ESPN is a TV channel.

Henrik Werdelin

So not that you could hear, I really use it in a serious way every day. Yeah. Let me tell you as you’re asking me.

Dan Shipper

I think that’s one of the funniest things I ever heard you say.

Henrik Werdelin

I know so little about these things. I was at the Goldman conference once and there’s all these famous people and I know not a lot about American sport because I grew up in Denmark. And so I’m standing there and I talked to this guy and he’s very, very, very kind and nice. And he’s just pivoting. He’s in between time and it’s amazing. He used to work in sports and now he’s doing like whatever. And so I talked to him for a while and then you know, turns out he was in baseball. And then my co-founder Matt Meeker is somewhere and I see him coming. He goes and he’s totally into sports. I was like, come on over, Matt. And Matt comes over and he gets like all weird. And I introduced him. It turns out to be Derek Jeter, and I just didn’t know who he was. And so maybe a little bit that I will tell you. I am ENTP, E-N-T-P.

Dan Shipper

Okay. I’m INTJ.

Henrik Werdelin

Extroversion, intuition, thinking, perceiving.

Dan Shipper

So we’re, well actually I used to be INTJ and now I think I’ve changed, but I’m like somewhere, I’m somewhere close to you. I’m on the border between I and E, I know that.

Henrik Werdelin

But then I wrote the blog post and then this therapist that I know basically sent me this long study from a few years ago that basically said that this whole thing had been debunked and you shouldn’t use it at all.

Dan Shipper

I think it’s useful if you, it’s useful if you think about it as a story that you can try on and it can highlight different things in your life that you find to be real.

(00:50:00)

Henrik Werdelin

I’ve got a thesis on it now. You mentioned the word story and it’s about this idea of serving a customer instead of defining your company as what you do. And the thesis is that we need to become much better, maybe it’s just because I’m a media entrepreneur as we just defined it. I think you need to be much better at defining your narrative and the narrative has to be pretty unique and very specific to you, because otherwise the AI models won’t be able to help you. If I write into ChatGPT, hey, I, the founder of Barkbox is two treats, two toys and a chew, and I’d like to come up with some new products, it’ll mention the stupid cat box idea. But if I say I’m into making dogs and their people happy and blah, blah, blah, it might come up with like other problems that dog lovers have. And so I do think that there’s this interesting point where the narratives that we create are just going to become very important because that is what we can use as we are putting on this AI Iron Man suit. And thus it’s going to be something that you need to have done.

Dan Shipper

There’s that Steve Jobs quote, which is the most powerful person is the person who tells stories, is the storyteller. And I think that’s totally right. And I’ll say I just asked both Claude and ChatGPT what my Myers-Briggs is, and they both said INTP. So we’re pretty close.

Henrik Werdelin

There you go. Then we’ll have our agents talk to each other’s agents. Wait, how far do you think that’s away from happening? The pure agent to agent?

Dan Shipper

Agents talking to agents?

Henrik Werdelin

Yeah.

Dan Shipper

I mean, it’s happening. I don’t think it’s happening now to a certain degree. But I don’t think we’re too far off. Like here’s an example. I will submit a PR on our code base and then Kieran, who runs Cora, which is one of our products, like first there’s an automated Claude that just runs in GitHub and does a code review, and then he will pull it down and then just like ask his Claude to like fix a bug or change the implementation or whatever. So it’s already like there’s like three, there’s an agent sandwich, there’s like three layers of agents between the two people. So I think it’s really, it’s happening.

Henrik Werdelin

What’s going to happen to storytelling then when my agent is going to read my Every newsletter and then rewrite it?

Dan Shipper

I think there’s a lot of much more, there’s a lot of commoditized information that will be subsumed under AI. You know, like textbooks are a good example. Most textbooks are just not that useful, or not that individual, I’ll say that. But I do think people want information from a particular type of, from a particular person or a particular source because it carries a particular perspective. And we have always consumed summaries of works of art or pieces of writing, whether that’s a book flap or a title or any of those things. And so this is just an extension of that. And I think it is better for artists or writers to have different forms of their work available for people who are at different levels of commitment and have different levels of attention. And it’s always a funnel into the thing that you hope people experience, but you can’t necessarily expect that to be the first thing that they engage with. And so I think good writers or good artists will find ways to make it compelling, make their whatever the AI summarizes of their work compelling and make it compelling enough that people want to consume the deeper real thing.

Henrik Werdelin

You know what, as you’re saying this, one thing that we haven’t announced yet through Autos, but it might be out when this gets out, otherwise breaking news, is the business model that we are looking to do with our entrepreneurs. So Autos is obviously this platform where you can build your own AI startup and we help you with everything from coming up with the idea to financing it, to making the ads, to talk to your customers and all those different things. And we’ve been thinking a lot about what the business model is on many levels, right? What is the business model for founders serving customers? It’s probably not just SaaS anymore. Like SaaS might be a part of it, but there’s probably going to be all these other models emerging and we can see that on the platform already.

But for us to the founder, what we are increasingly inspired by is basically old labels like music labels. And so the first cohort of founders that we’re going to have coming out are basically people that we have signed. We are a label and they’re basically an artist, an entrepreneur in residence, on our label, and then we give them money and resources and marketing and all this stuff that a label will do. And then we take a royalty. So we don’t take equity in their business, we just take a royalty on the top line. But in many ways, and maybe inspired by what you say about media entrepreneurship, like maybe a lot of this vibe-coded solo entrepreneurship increasingly looks like artists, because what they’ll need to do is to have an interestingness, they’ll have to have a story, something that resonates with a group of customers, and they’ll need to be able to articulate that in a way where they get authenticity and authority to do that. And then in return, they’re going to offer not just writing, which is traditional media entrepreneurship, but also products, which you guys of course are doing, but that we are now hoping to scale so that if the media landscape got disrupted by YouTube where everybody could make a channel, however niche it was, I think increasingly we’re trying to see it as like the same thing where people get on our platform, they’ll get signed and then they’ll serve customers and we’ll take care of all the boring stuff so they can take care of their relationship capital.

Dan Shipper

That’s really interesting. And are they all serving a similar type of customer?

Henrik Werdelin

Well, they’re all serving very different customers, right. I think in many ways we don’t have a good way of articulating, but we basically have this thesis that the Dunbar rule is going to work in an AI world too, but it’s going to be exponential. So we call it Dunbar Squared, so that the TAM of a good customer group is 150 lifted to the second. So 22,500 and that’s the TAM. And that is a group where you can do something new. So the people that are on the platform are people who, the one we just had, we just had a piece come out in the Economist about a woman called Sarah. She unfortunately lost her dad, and then she realized that the whole world about loss is very non-innovative. Anything from grief counseling and support to like body logistics. How do you literally get your dad’s body from A to B? And so she’s building all these different agents to help you with that. So very niche. We have like a woman from Texas, she helps women who have just given birth to get in shape after a pregnancy. And so that of course starts with making kind of a fitness protocol, but it’s also about calorie counting and sleep and all those different things. And so what you see on the platform is a little bit like YouTube, that you just have like people who have this very specific knowledge about tree cutting in Florida or survival trips in Washington.

Dan Shipper

That’s really interesting. And when’s that launching?

Henrik Werdelin

I mean like what’s fascinating, this week, maybe next week. Like you know what’s fascinating is, and I think a conversation that my business partner Nicholas and I have a lot is we are increasingly writing all the features, right? So we are vibe-coding the features and then we have a few engineers working with us and they are then basically enterprising it. So, but we have rewritten the platform. We launched in June with V1, which was very text-heavy and was very kind of identity. And now we’re kind of, as vibe-coding and vibe app creation is becoming easier, we wanted to implement that. And so we have in this next version of the feature, you basically come in and you riff with our agents on what business you want to make. And then we build the landing page. We build the first initial agents. We build this thing, which I think is very important, which is basically an agent communication layer between you and your customers. So you can have high relationship capital, like a strong omnidirectional relationship with your customers, but done with an agent layer. And then we do the ads and these kind of things and all this stuff. We have the core components and so we have the backend system, but we rebuilt the whole front end in a few months. And so, and of course like when you’re sitting and writing, as you know, when you can write code as fast as you can write blog posts, then you constantly come up with these new things that you want to add to the system. And so give us a few weeks and the V2 will be ready.

Dan Shipper

That’s great. And for people who are listening to this and want to find you or find Autos, where can they find you?

Henrik Werdelin

Autos.com is where you can go and make a company with us, and so we hope people do that. I’m active on LinkedIn, so connect on LinkedIn and hear my random inklings and my bad copies of whatever Shipper said. And so yeah, that’s probably, I wrote this book called “Me, My Customer, and AI” and I have a podcast called “Beyond the Prompt” with Stanford Professor Jeremy Bailenson. So plenty of places where you can hear me babbling.

Dan Shipper

Awesome, Henrik. Always a pleasure.

Henrik Werdelin

Thank you so much.

Thanks to Scott Nover for editorial support.

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

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

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

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

For sponsorship opportunities, reach out to sponsorships@every.to.