Full Transcript for Tobi Lütke on the AI agent revolution and future of Shopify

Toby, there was something that happened recently and we had already scheduled to have this conversation and I saw it and I was like, man, I'm dying to know what Toby thinks of this. And it was when Amazon recently blocked Muse. I would love to know your reaction to that when you saw that. Yeah, I found out after posting that we are embracing Muse.

So that was kind of a fun experience. So that looked pretty cheeky because I think people sort of draw to a contrast of Shopify and Amazon taking basically polar opposite positions to the incredible, authentic capabilities that are coming out of the companies, Muse specifically, I think is such a remarkable product. It's funny because people's reaction was very surprised that Shopify and Amazon found themselves on different sides.

But to me, that's been the last 20 years.

There's a difference of philosophy, I think, in commerce. And I think neither of them is necessarily a right position. It's just like these are two different approaches that just I think Shopify and Amazon have always represented. It's like one is the digitalization of a department store, which is like, you know, bringing forward the merchandiser, the like everything store, you go to one place, you get everything.

And it's like, you get everything. And it's like, you get everything. And it's like, you get everything. And it's like, you get everything. And it's like, you get everything. And it's like, you get everything. And it's like, you get everything. And it's like, you get everything. And it's like, you get everything. And it's like, you get everything.

And it's like, you get everything. And it's like, you get everything. And it's like, you get everything. And it's like, you get everything. And it's like, you get everything. And it's like, you get everything. And it's like, you get everything. And it's like, you get everything. And it's like, you get everything. And it's like, you get everything.

And it's like, you get everything. And it's like, you get everything. And it's like, you get everything. And it's like, you get everything. And it's like, you get everything. And it's like, you get everything. And it's like, you get everything. And it's like, you get everything. And it's like, you get everything. And it's like, you get everything.

And it's like, you get everything. And it's like, you get everything. And it's like, you get everything. And it's like, you get everything. And it's like, you get everything. And it's like, you get everything. And it's like, you get everything. And it's like, you get everything. And it's like, you get everything. And it's like, you get everything.

And it's like, you get everything. And it's like, you get everything. And it's like, you get everything. And it's like, you get everything. And it's like, you get everything. And it's like, you get everything. And it's like, you get everything. And it's like,

you get everything. And it's like, you get everything. And do you think that's the reason that some of these large platforms have already kind of started to instinctually throw up their walls to these agents is ads to protect? Or is it more than that? I don't know. I mean, the ads business, I'm not that deep into it, because to your point, we don't have one.

We fairly intentionally built the business in such a way that up and down the stack, we are on the same side of the table as our customers. And this is specifically the reason why we haven't built a marketplace in the traditional sense, because the marketplace is a three -sided situation. You have, of course, the corporate entity with its own interests.

But you have buyers and sellers. And you need to make a sort of choice between them. The correct choice in that moment will always be the buyers, because those are the ones you optimize. And you just like easy observable data, too. So it's easier. We specifically, like, just have, we look out for merchants. We would like make it easy to for entrepreneurship.

We, you know, this is sort of our obsession, like in this place. And so, you know, with the emergence of a genetic world, I think it's even too early to tell if this is truly playing into our sort of worldview better than the aggregators. I think what's the, I think the aggregators have a little bit more concrete need to stay in control.

I think it's really, I think it's really, I think it's really, it's really, it's really, it's really, it's really, it's really, it's really, it's really, it's really, it's really, it's really, it's really, it's really, it's really, it's really, it's really, it's really, it's really, it's really, it's really, it's really, it's really, it's really, it's really, it's really, it's really, it's really, it's really, it's really, it's really, it's really, it's really, it's really, it's really, it's really, it's really, it's really, it's really, it's really, it's really, it's really, it's really, it's really, it's really, it's really, it's really, it's really, it's really, it's

really, it's really, it's really, it's really So, like, for the city, like, Shopify is very well known for its online stores, but really, it's being used, just for, for so, in so many other situations, right, like, it's, obviously, most people probably purchase things on Shopify through maybe an Instagram ad or something like this, that's very common, sort of, we call it the LA you of small businesses, you know, and so, Shopify is fundamentally just allows you to sell, and it's not, like, for online stores, your home base, it's the easiest thing to see, it's the thing that merchants control, it's that they design, they represent themselves, they do their storytelling,

but, like, people sell through any of the social channels, people sell through a point of sale, people sell on crafts fairs, and people sell, like, the more places people sell their products, I think the more valuable Shopify becomes, so it's very natural for us to, like, make sure that, you know, we get the products that, like, like, I think Shopify is presiding over the most interesting product catalog on planet Earth, because most of new businesses building, you know, making, you know, the best fashion designers, all the way to the newest electronics, I guess, they, they, they have a Shopify store, they start on Shopify often, and, so, we want to bring that catalog to,

like, everywhere, so, it's very natural for us. This interview is presented by Hyperagent. Hire AI coworkers that take on real roles, collaborate with your team, and help run your business. Learn more at hyperagent .com. Thanks also to Granola, the AI notepad for people in back -to -back meetings. It works everywhere you do and lets you focus on what matters.

Try it at granola .ai slash sources and use the code sources for three months off. This episode is brought to you by Jira by Atlassian, where teams get the context, visibility, and control to move work forward. Try it free at jira .com. That's J -I -R -A dot com. This episode is also brought to you by Mercury, AI -native banking that's loved by more than 300 ,000 entrepreneurs, including me.

Visit mercury .com to learn more. Mercury's a fintech, not a bank. Check the show notes for details. I'm curious how agents are already, in Jira's view, changing how people are shopping. Are there early trends you've noticed? Yeah, so it's really, I mean, it's definitely increasing. It's obviously, we enter this year with a fairly low base.

So the percentages are absurd, but like they also don't mean much. What's really interesting is the sort of changes, like the types of purchasing that's happening. Fairly often, it's the sort of project -based purchases, like help me get everything I need for setting up my home gym. It's like the kind of action that we see. The planning, we can't see this, but imagine this conversation might even start with a picture of a room or a floor plan.

Like an entire phase of figuring out what should be in there or what you would like to do. So deeply considered, fairly high order value purchases happen with increasing frequency through agents. The things that I personally think are really, really fantastic.

I mean, I guess the research is one part, but things such as I would like to do a project. Let's say I have a 3D printer or whatever. I would like a project with my kids on a weekend, buy me everything I need that will be in the end time. And like this is like it's such an actionable, really delightful experience to say, hey, we're going to build a whatever it is, something with a little bit of electronics in it or maybe needing some magnets and some filament.

And it figures this all out and just like places for order and places from the local stores make sure everything is in place. So I think this is the kind of thing we see more and more. But honestly, this is super quickly evolving. And there's entire forums now dedicated for people like just like posting, you know, I think the Muse case word is a fun one that the Muse team has started using for just finding delightful experiences that people are doing with agents.

It sounds like a long tail is forming that maybe didn't exist before based on the examples you just gave. Yeah. And again, this is it keeps getting more and more. I mean, I just enjoy how futuristic today feels like, you know, in this way, like just like I just marvel at this. Like, you know, I grew up reading lots of sci -fi and like the world was always like I just enjoy seeing the sort of imagination by the authors about, you know, like how do we describe the future? It's a really, really daunting task if you think about it, you know, extrapolating from the 60s to the 2020s or whatever, figuring out what the world is like, would be like.

And so and some did it extremely well and some, you know, sort of got the details quite wrong, but like plausible. You look at today and you just like when you actually hear what like especially the really sort of agentically pilled people do. It's like it's I don't know. I have endless like interest in hearing these like crazy stories like I like to stay on the music example.

Like it's quite easy to put news on what's called your like your tail net, which means that can like access devices in your house, which again, you have to be comfortable with this. But for that weekend project of 3D printing something, you can also just connect it directly to your printer. And like it can reorder the filament when it's empty and like it's starting to like take more active roles in this.

And this already works in some instances. I think we are currently in a phase where the agents are largely helping people do things that they are already doing better. And I think what usually follows afterwards is a phase that we discover completely new things that weren't possible before. And I think I like I think you're starting to see the glimmers of that and like that's kind of that's usually when technology gets the most fun when people are finding sort of a digitally native new possibilities.

The sci -fi point you brought up, you're right. Sci -fi is often very prescient and it also motivates a lot of the builders in tech to build what they build. I was curious and I know you read a lot. Is there anything from sci -fi that you've read that you're seeing now take shape that you're like, wow, how in the world did they predict that? Yeah, I recently reread Cryptonomicon from Neil Stevenson and which I loved reading when I was like early 20s.

And my God, like my son just read it and we were talking about it. So I picked it up again and I was like, like literally everything in the book just happened, you know, and like and it's all sounded pretty fantastic, but like, you know, just I he had an absolute read even over cryptocurrencies and sort of fractional ownership and so on. It's like, how how did you put that off? Yeah.

And so, you know, it just that's more near future. I think when it goes further out, I think it's more like the overall picture. It's like it's like no one got it right. But like if you put all the pieces together, so you sort of like what I've what I really, really like doing, frankly, is like go through the books that I enjoyed and also just check when they were like written.

Because at this point in my life, I know enough about history. So that I realize that a lot of sci -fi books actually tell you more about the time they were written in than the time they were written about. Oh, interesting. Because, you know, just like fundamentally, like you can even look at the Spider -Man comics. It's like during the nuclear age, it was an atomic spider, right? It's like it's just like it's like it's always like these are things are representations by the time they come from.

The question like the sad thing is that there was probably a 15, maybe 20 year period that I don't think people wrote a lot of sci -fi. That was like it was very much when it was sci -fi, it was Hunger Games. It was very dystopian. Outside of maybe the Andy Weir books or so, I think there was like maybe a little optimistic sci -fi. And I think that's actually contributed to sort of a tech mobiness in the industry.

Is this around the era of a Watchmen and stuff too? It's exactly. It's just like these kind of it just I think I mean, you could also take a sort of commercial lens to this. It's just turned out that writing dystopian sci -fi or like near future scenarios ended up becoming bestsellers. And that was never true for the sci -fi books. It was always it was always niche.

So it wasn't people just wrote the book they had in them, I think, instead of writing them for an audience. So whatever caused this trend, it certainly was that like the technology industry turned quite pessimistic. And what's so lovely about I think now is like there's just like like there's just more optimism around again. And I think that's that's just really, really good because technology is just progress.

It's a sort of core ingredient in progress. And like for the first time in our lives, we have software that we train rather than write. And we then basically have to interview it for its capabilities afterwards. And it just seems like even that's like sounds like straight out of sci -fi, doesn't it? And so I feel like I always want to look around to say, my God, I can't believe like I have like this infinitely patient, agentic sidekicks around that like just help me kind of do cool shit all day.

Like and then do the stuff I have to do easier and make my life a couple percent better. And so it's quite fantastical. It's a wonderful time to be in tech. Do you think that optimism is just because of the feeling of invention and invention as a thing that naturally gives humans optimism? Yeah, I believe that's exactly it. I think it's you hit the nail on the head there.

I think humans need a frontier of sorts because I think we work best when there's opportunity because obviously, again, theoretically, if there's no frontier or not enough perceived growth around us, fundamentally, the correct strategy ends up becoming competing for what's already exists. And it sounds like a small thing, but the difference in what it feels like to be in an environment of where zero -sum thinking predominates or positive -sum thinking, where the opportunity is about and like people are going for new markets, new ideas.

It's about the biggest difference you can possibly imagine. These are two separate sites on the points on a very, very, very, very long spectrum. It's the thing that I think is so intoxicating joining a startup that's winning because there's so obvious growth around you. It's so

and Alex Heath and Alex Heath and Alex Heath and Alex Heath and Alex Heath and Alex Heath

and Alex Heath and Alex Heath or finding a niche, a very small group of people that everyone agrees that it's okay to be optimistic which is usually comes these days in the form of a company or a startup. And so, but once pessimistic, like once the physics returns to this, the reversion to the mean means that it's sophisticated to be pessimistic and that usually means you don't take advantage of as many opportunities which then leads to lack of growth and lack of appreciation of opportunity which then very quickly leads to zero -sum thinking and competition.

And I think somewhere in the story is also sort of a fate of what happens often with companies that they go through a maturity cycle, with industries. I think this is precisely the thing I described earlier that happened to tech. What's again, like I just like, I'm so fortunate and we are also fortunate, I think, to have opened a completely new frontier, not just one that is like, oh, there's like, you know, new little island to explore for 15 minutes and afterwards we are like, we're good, we're going back to what we had.

It's like, it's like magical territory in which we teach sand to think and afterwards we have to reinvent everything and like it's like, it's just such a wonderful opportunity to just like make like the best version of, you know, all bets are off, path dependence is out of a window, we can reinvent everything, we can fix everything and our computers now fulfill wishes.

It's like unbelievable. The way you said that is so interesting to me, zero sum thinking and the contrast. To me, when I heard you say that, it's like, why is zero sum thinking such a negative thing? And I feel like there's a futility in it. There's a futility in having a zero sum view of the world, because there's only so much to fight over.

Yeah. And what you're describing is like abundance. Yeah. Which, I'm wondering if you agree with that. No, 100%. It's like, I mean, again, zero sum thinking means that you have to compete for every piece of resource and that means for every time you win, someone else loses. And that is a mindset that's common, right? And it leads to, I mean, if a world would be like this, it would be a pretty, pretty bad place.

Like it's like, we do not want this. We are also lucky because it's, so some thinking is a choice. If you understand economics reasonably well, or at least you understand how value is created in the world, you realize that there's, we aren't probably even at the first basis points of value, that's been converted into economic value from the potentiality around us, right? Like we, I mean, this is what AI tells us ourselves.

Like the AI industries already has added like probably trillions to, of, of, of, of new objective value into, onto planet Earth. Like, and so all that value was hiding in floating point math. So the difference, like we've had floating point math for a very, very long time. We, we knew about matrix multiplications for a very, very long time.

Uh, yet I don't think we realized that, um, at the limit, we could extract trillions and trillions of dollars of growth from just doing a floating point math and matrix multiplications much faster. And, uh, in a particular way, right? So I think this is where you have to think of everything, including resources. Like before we knew what to do with uranium, we had no idea what like the entire value of nuclear energy was, uh, locked behind this piece of knowledge.

The second we figured out how to break, um, the atom, suddenly there was an infinite amount of energy and then we choose not to use it for some reason. And so we, like this is the way the world truly works. It's like, we are, we are, we are gifted with utterly unfinished,

potential value. Every product that could exist is out there. Many, many, many of those shouldn't exist for very good reasons that someone might or might not discover. Um, but many products need to exist. And I mean, this is what we see in like in millions of businesses are formed on Shopify to, um, create the thing that someone feels like needs to exist in a world.

And that is not like, it is, it would be so impossible for me to be, um, uh, thinking in, uh, zero sum, because I, I see this every day, you know, you know, someone, you know, someone quit their accounting job to, um, create an, uh, like a completely amazing business that, that now hundreds of thousands of people appreciate the products of.

And that was a choice at a lunch break at some point. Right. And again, if that lunch break would have gone slightly different, um, uh, you know, then, um, that value would just not exist. And so I, I see this every day and it seems very, very natural to me, but I feel the need to bring that up every once in a while, because it seems, um, like most people don't have such a fortunate, uh, vantage point, um, to, to, to, to, to, to, to, to, to, to, to, to, to, to, to,

to, to, to, to, to, to, to, to, to, to, to, to, to, to, to, you know, where you are in the world, obviously your economic status, but be curious to hear based on everything you just said, the positive vision you're laying out, how you make, uh, of, the moment we're in and how people feel. I am probably not the most representative person to talk to, um, on this.

I, I don't think I can see all these new businesses. And, you know, I imagine there's more businesses being started on Shopify now than ever. A hundred percent. Like, think about it. Like it's, um, entrepreneurship is one of the probably jobs that is most benefiting from AI, uh, imaginable, right? Because, um, again, you are trying, as an entrepreneur, you have to kind of do everything.

Um, at the very least you have to understand every job because you have to hire someone and then judge if they are doing a good job for you. It's extremely hard and very, very daunting. And, um, once upon a time we had about 80, almost 80 people, 80% of people on Shopify, uh, said yes to the, to the question. Um, if they have someone in their life who would answer entrepreneur question in 24 hours over email or text or something.

Um, that didn't mean that like these people are abundant. That means that it's survivorship bias. It's for people who build the businesses of a people who can ask the questions to someone. Right. And so this is now democratized. Like, AI is very good at answering every entrepreneur question often, um, better than some of the people who might have replied to my emails.

Um, turns out there's a couple of core myths that Western civilization is built on. And one of them is a Promethean myth myth. It's like the Promethean myth means, Prometheus brought fire from Mount Olympus to the humans and, and got the whole party going. Um, and then of course was, uh, doomed to, um, torture forever, uh, I think this myth is, uh, casting a more powerful spell over us than, people generally admit.

You have a very strong, uh, sense that everything that's good. So has negative, um, consequences of equal proportion. No doubt. It's like, like, potentially the most powerful thing that ever will be invented by humans, um, or definitely in the top, few top, ones. So it's a lever and so obviously it can be used for really good, really bad.

On an individual level it will have enormous positive influence on some and negative on others. So all this sort of as a proviso, I just don't think this usually plays out this way. I think the positive effects are going to be so vastly higher than the negative effects. And I get to say nothing about the sort of PDOOM scenarios which I just like my PDOOM is zero.

It's like again at the end of the day it's floating points. There's going to be consequences but like my god. Again I'm a technologist so I can say this like I think to me dystopia is always just 30 years ago. It's just like do you know how bad the Wi -Fi was 30 years ago? No Wi -Fi. You know like just like you did teleport me 30 years back and it would seem horrible from today's perspective.

And I think this is just we're just going to keep this going. I think things can get really, really, really great. I think maybe where it's coming from is you're describing AI as leverage which I agree with. And I agree with you on the PDOOM thing and it's math at the end of the day. But I think the reason a lot of people feel very strongly about this is some leaders are talking about AI as if it's an entity that we can't control.

Yeah. That's a god essentially. Yeah. You obviously seem to disagree with that. Yes. Is there anything that can happen to... Because it feels like it may get worse before it gets better in terms of the sentiment. And in the US we have midterms coming up and all these things and it's going to be this political lightning rod. Yeah. And it's a thing to which people can assign a lot of pent up frustration they have with their lives and the economy and the government.

Yeah. And I'm not sure it's going to get better before it gets worse. Yeah. I think people are right to pay attention to AI a lot. And it would be extremely hard to sort out what opinion to have about it given the public discourse. Because again, I'll just go back to pessimism sounds smart. Optimism sounds naive. And there's just like the incentives are such that it's a lot of people chose the pessimism route for one reason or another.

And that's the predominant narrative you're going to get. I think this is not going to take too much because personal experience of AI is going to become just like much more common and people will very, very, very much like their eyes. Like, again, just, again, we've been trying to make software that makes it easier to make entrepreneurship for like 22 years now.

And what we hear from our customers is that they now feel that computers actually are as good as how we techies used to describe computers. You know, like when, like I remember this as a teenager, I tried to tell my parents, it's like, these things can do everything. We network them together. Like it's revolutionized everything. You just like, you just write this bit of thing and then it has this new capability.

And like, to me, that was very normal because that's what I could do. And then it turns out that was like a skill that ended up not being that well diffused because it kind of, you had to frankly spend your entire teenage years in a dark room cultivating your programming ability. The terminal is intimidating. Which was exactly, not exactly what, to me, this seemed like, why would I do anything else? Right? So this is like, not exactly, like I probably don't even know if I was anything else I could have been doing during this time.

So it's like, I suddenly didn't find out. And so I couldn't stay away. It was too interesting. And so not everyone's like me and that's, thank God. And so it turned out this, like using computers in this way I can describe and I always felt about them was a very, very rarely diffused skill, which then meant that people like me should go and build as generic as possible software so that other people can experience at least a semblance of this through like software like Shopify and others.

But now we have a completely, again, I keep going back to this. Now you kind of have a person by your side and your computer fulfills wishes. Like it's like the future versions that people will experience of Shopify or, and that software like it are going to be just like remarkable pieces of engineering. That's not actually very engineering.

It's like, it's a remarkable pieces of collaboration between synthetic intelligence and real intelligence trying to create something that's like, like it just can completely give, we want to give people cruise control for their ambitions, right? Like we would just want them to be successful doing their thing. And I think so many more people can do that in the future.

And I think that's like, I don't know, people will experience this. I think the stories will wash the pessimism away. There's some interesting stats in the US about how net new business creation is skyrocketing. And I'm curious to hear if that maps to what you're seeing on Shopify. Are there more Shopify storefronts being open now than ever? Can you talk, what's the curve? Absolutely.

It's, it's, it's, it's, it's, it's going really, really well. And earlier too, it's like it's younger. I think the new generation is extremely entrepreneurial. It's really, I mean, there's a combination of AI use, but also their media as much as I think every generation parents sort of maligns the media consumptions of the next generation.

You know, with no break. So I don't know how often Mr. Beast got banned in various households, but like, I'm not so sure that was the right call because it's clearly entertainment, but it's very, very entrepreneurial by itself. Like, like the kids grew up seeing money being used to build incredible, you know, games and stages and competitions.

And like, like the, the, the particular content is, you can think about whatever you want, but the meter content was like deeply entrepreneurial and very, very clear. And then. Well, you can see that in the creator too. You can see their progression. Yes. Then gathering more resources, becoming more established, building things. Yeah. Even just showing, Hey, this is the kind of things that you can do when you get resources is like actually useful because like, why are you like, why strive? Why, why, why, why, um, you, you can make good money working 35 to 40 hours for someone else.

Um, you know, why spend a hundred working for yourself, right? Like it's, um, and so I think, um, that, and then also just like the clever use and like the clear sort of progression in, in scale of this, um, felt probably too many. Um, at least my, my kids sort of confirmed this just like felt like a journey they were on with Jimmy. Um, and, um, that is an intoxicating experience and you want to be part of more such journeys of, again, journeys of growth, journeys of opportunity.

We see a lot more business formation now. Um, and I think this sort of goes to your previous point of, um, attitudes towards AI. Once again, it's leading in the U S significantly. It's, I think other, uh, parts of the, um, Western world are on an early incline on that sort of map slope of America, but like, it is just remarkable how America is a country of early adoption.

Um, and, um, so I think it's, it's, it's interesting to see even these sort of societal, almost stereotypes, um, bear out in the data. So, um, I think, uh, I spent some time in Europe recently and the AI attitude is, it's not that it's more negative. It's, it's, it's, it's a little bit, it's definitely more cynical, which, um, I think it's just Europe and European, so I can say that.

Um, and, And definitely German. But like, which is also fair, like things should, you know, it's fair for a country to decide, hey, like stuff has to prove itself first, that's all good. But it's almost like it just in effect, like it's when you talk with some people about AI, it's like, oh, yeah, yeah, I understand AI. I installed ChatGPT when it came out.

It's, it hallucinates a lot. I mean, yeah, okay, there's like five more layers to this now. You know, like, it's like you're not completely in the meter there. So it's like people reason on bad priors. And I think that's hopefully like changing now too, because, again, but Muse is not available in Europe, right? So this is also a problem.

In Canada, we are a little bit luckier, so we usually still get the release. I was going to ask, how is it in Canada compared to, you said, you know, you talked about US and then Europe. What about Canada? I think Canada is as so often, somewhere in between, maybe slightly closer to the United States and on the spectrum than Europe. I always called Canada as America for Europeans.

And it's like, it feels like pretty apt. Canada takes a lot of pride in having massively contributed to AI, like here in Toronto.

Like, I think Alex and Ed got trained here in U of T. And of course, Professor Hinton is here and go here. And then there's some labs now too. But like really sort of, especially in the during the nuclear winter of of Machine Learning.

Here was a pocket where people continued pursuing it and laid the foundations for what came afterwards. I think that helps a little bit here. So anyway, I think it's not as unbridled as I see in Silicon Valley, but it's very clear this is the future, we need to use it to make better companies, make better government services and bring productivity out.

These kind of conversations are being had here. We were saying before we started that you go to California about once a month. I'm curious, given the energy you're describing there, does this moment ever make you want to move Shopify to California?

No, I'm very happy to be outside. I love Silicon Valley really. I think a big skill of mine is being an outsider and observing things at distance and allowing myself to not take the exact package and put it into my neurons, but actually do a little bit of an edit for sensibilities.

Silicon Valley raw is not for the faint of heart. It's a lot. Yeah, yeah. The swings are massive. I think of Silicon Valley as a pendulum that just goes rapidly and I believe there's always some green zone somewhere and I try to spend more time in that. So, I don't know.

My criticism of Silicon Valley is that companies are very, very convergent. Cultures are really not very different. It's like try to find a company that truly escapes the OKR system. And, you know, just like it's just at some point gravity is too much to escape some of these ideas. And it means you can't really reinvent things very well and time gets better.

Silicon Valley has a unique thing of so much new business formation and an ability to grow new business so quickly that sometimes someone manages to build like enough of a different culture to then have almost a pull for everyone else to to adjust to this. So it is able to evolve slightly. Um, but like I think you like I really enjoy company building.

I enjoy thinking about systems. I enjoy having the ability to make decisions that are just like go like a very different from the rest of the industry. And like either from first principles figure out what's like why no one's doing it or discover something new and then tell others. I think that's it's not the most important thing in the about every individual change we make.

But like I think in the aggregate, it's like I think there needs to be diversity of companies like companies for different types of people. Um, I think, um, it's bad when companies try to be everything to everyone. And, um, so I, I find that just much, much easier to do from a distance. Um, you know. We were talking also earlier about, uh, you know, you've been on the board of Meta for a while.

And I'd be curious to hear what it's like, uh, working with Mark and, uh, Mark's been on the show recently. We were talking about Muse and I assume you got to try Muse early as did I. Uh, and what it was like seeing that kind of take form inside the company. And then obviously the world feels like it's changed now, but yeah, maybe a little bit about like how Meta fits into what you just described about Silicon Valley.

Yeah. Meta is, I, I think it's such a fascinating company. Mark is, uh, clearly one of the best entrepreneurs who's ever, um, uh, lived. Um, he's, um, just a remarkable, uh, individual in the way he, um, like really orchestrates a company. But is able to just change on like, I mean, I think, I think everyone to have got a bit of a front row seat here then.

Um, you know, I, I was incredibly thankful for the gifts, uh, at Meta made to the, really all of us by open sourcing the Lama models. It really started the trend or like, actually, I, who knows what would have existed if any open source would have existed without that. Um, it, um, it created, uh, like a norm or like, uh, it's like if Meta can do this, others can do it too.

And, um, I think it's inspired more companies to, to, to, to go for this. And I think we all, um, um, um, that is, I think, I think without the open source availability of, uh, AI models, I, I would modify my, uh, optimism slightly. I think, I think my optimism is, uh, like, it would still not be, um, it would still be sky high, but like, uh, maybe not as, uh, via own cruise control as, uh, as it is now.

Um, because I think, um, you need competition, um, in, in a space. And I think, um, having, so, so this was a massive contribution. I was incredibly thankful for, uh, for that. Um, and it's a, one of those decisions, you know, like if Mark wouldn't, like if founder wouldn't have been present, there would be no way Meta would have done this.

So that was great. But then like, there was a big setback with the Lama 4, which is like, this happens, this happens at every lab. It just happened very publicly with Meta that there was a mistrain. It didn't work. There was some, uh, ideas that didn't quite plan out. That's, you know, just another result essentially in science. And so seeing what Mark did and how quickly he jumped into action to reassemble like an amazing team that's now training, uh, like just absolutely insane models.

And like really just like the quickness by which they came back is like, and anyway, so I'm fascinated with this. Um, I'm also really thankful from personal perspective here just to be having this view. Because every time I come back, I really feel like, um, thankful that Shopify is not 70 ,000 people. So like, I don't think I, I'm a leader who could deal with that.

I think it's, it's like, you have to be Mark quality entrepreneur to do this, not Toby quality entrepreneur. So like it's, it's, uh, it's, uh, there's always another level and it's nice to live at, um, that kind of scale vicariously rather than having to, um, figure out from first principles that I probably can't handle that. So employ AI staff to run your business with hyper agent agents can find new customers, run campaigns and handle your business operations.

They work in the cloud, collaborate with everyone on your team and pass work to each other. Like real coworkers, hire your AI staff at hyper agent .com. I spend a lot of time context switching between meetings often with no time to process one before the next starts. Thankfully, granola runs in the background the whole time. It's an easy to use AI notepad for meetings that works everywhere.

Even on phone calls, I use granola to recall what was said in meetings and create helpful summaries. I use it every day to stay on top of what I need to get done with my team. It connects to my email and suggest follow -ups for me to quickly review and send saving me valuable time. Granola isn't just a core part of my workflow. It's basically my second brain.

Try granola at granola .ai slash sources and use the promo code sources for three months off. As you move faster and bring in more AI agents, keeping tabs on what's actually happening gets harder. Who's working on that? What did that agent just change? Are we still on track for our goals? That's why teams rely on Jira for visibility across their work, whether it's done by people or agents.

Jira is powered by Atlassian's teamwork graph, which connects work across Jira, your most used tools and apps, and your agents. That means your agents share the same context your team has, like the goal behind the work, the decision made yesterday, and the project plan, so they work faster and more accurately. You get oversight, track agent progress right alongside human work, spot blockers sooner, and connect your work to goals.

You don't have to choose between shipping fast and One of the biggest pains I've had running Sources has been managing my books. I'm not an accountant, and all of the software out there is unintuitive, clunky, and detached from where my transactions actually live, until now. Mercury recently launched Mercury Books. It's AI -powered accounting that works with your Mercury banking and credit card transactions, plus external cards and payroll systems as well.

You can choose between cash and accrual accounting and quickly generate reports for your cash flow, P &L, and balance sheet. Like everything Mercury does, the design of Mercury Books is super approachable and clean. I finally don't feel overwhelmed when I'm trying to understand the big picture of my business. I love that AI does most of the heavy lifting, including for things like auto -categorizing transactions, and that Mercury's command AI agent can handle tasks end -to -end.

You don't need to be paying for separate accounting software anymore, and you can even invite your accountant or CPA to work alongside your Mercury Books with you. This doesn't cost extra, and Mercury will even help you find a human accountant if you need one. Visit mercury .com slash books to learn more and apply online in minutes. Mercury is a fintech company, not an FDIC -insured bank.

Banking services provided through Choice Financial Group and Column NA, members FDIC. Framer is the AI website builder that powers the Sources Podcast website at podcast .sources .news. Framer brings AI agents into the same canvas where your website is designed, managed, and published, so you can move faster without giving up your taste or control.

I use Framer to make the Sources Podcast website be the destination for everywhere you can find the show, plus you can also see recent issues of my newsletter. Start building with agents for free today at framer .com slash sources for 30% off a Framer Pro annual plan. Framer .com slash sources. Rules and restrictions may apply. You were saying, which I agree with, like, if Mark hadn't been there that moment with what's happening now and him rebuilding things probably wouldn't have happened, or even the original llama, which I agree with.

That made me think, are there things that, and you've talked a lot about how, like, the pandemic was very hard for Shopify, obviously. Are there things that you've reflected on over the last several years where you're like, man, if I wasn't here and I didn't have the setup that I have as the founder and CEO, like, Shopify would be in a really rough spot right now? Is there anything you look back on and, like, this is the call that I had to make and it was hard and I had to convince everyone, but here we are? You know, like, I think companies that have their founder slot filled are just completely different beasts because they have an asset that's, like, impossible to

replicate, right? Like, it's not actually as much about, like, I mean, if you have Mark Zuckerberg in a founder slot, then that's really, really, like, it's not just, like, you have a founder slot filled, you also have Mark. So that's excellent. But just having a founder is actually incredible because there's just, like, an ability to cut bureaucracy, cut through red tape, make decisions fast.

Like, one example of this was, like, then as a technologist, I want to have a really good read on the future, right? That's important. We make products that take a while to create and so we have to have, we have to build for a future as it will exist rather than one that does exist right now. If we do this and we get it exactly right, we might actually change the sort of particulars of our future.

Maybe not, it would have happened anyway, but we have, might have an impact on it, which is really, really fun. But once I build conviction of, like, here's something that needs to happen, I have tools that, like, a professional manager who would be hired just doesn't have. I can go and say, send a memo to the company about, hey, we are now requiring reflexive AI use for everything.

And I'm going to ask you every time you need something from me about how has your AI experiment failed? I assume a committee did not review that memo before you sent that. So that's very true. And, yeah, committees. If they had, maybe that wouldn't have been sent. Yeah, I'm pretty down on committees. I was just like, I call, I, it's, it's, I, I don't.

And it was controversial when you sent that. There's a lot of headlines about that. But not in Shopify, which was fun. So, like, in Shopify, actually started, I think more people ended up paying attention to the memo after the press made a big deal out of it. And so, oh, that's like, my mom asked me about this. It's like, oh, yes, I could have this wonderful conversation about all this AI usage we are doing.

I actually just thought this sounded exactly like what you say at Town Hall. So, like, that was actually gratifying because it was seen as controversial but really not inside of a firm. That just means we managed to transition already pretty early. Now we are, like, a year and a half later. So it's just nuts what's happening. So it's, like, really, it's, I wish people could have a good look at what it's like in the specifically highly AI -pilled companies right now.

It is absolutely fantastical. It's just, like, in a way that, like, I could not have predicted. And I just, like, I go from review to review. And I'm, like, just, like, as just, like, I'm a true technologist, right? Like, I love computers. And I hire people who love computers. And I don't think this is any problem. Like, it's just this is perfect.

This is what we should be doing. I just, as someone who loves computers and loves the power that can be extracted from them and is just delighted when someone invents something, I just go from meeting to meeting. I'm, like, I see the craziest thing I've seen in my life basically once a week. It's just absolutely wonderful. So, you know, I don't know.

Is it possible to point a company so directly into a direction by, if your mechanism of management and leadership is speeches or the creation of an annual plan that then you sort of build equity in that then everyone follows? I've never had an annual plan. Like, it's just, like, it's, like, we, like, we run a function over all the inputs to figure out what's the next step every day, right? And it's, like, it's not just me doing it.

I teach the function. And I want everyone to run this and just do what they see as, well, here's the next best step. And we try to get everyone to move into the same direction but as loosely coupled as possible. And so I think this is how you get to high creativity, to innovation and for– into a place where people build hopefully world -class products, right? And so all this to say your question is whether situations where I was needed.

My hope is not. If you read my letter that I wrote when shopper in public, it's 11, 12 years ago now. I included a little letter from Toby in there, which was a nice thing that the SEC allowed us to do. It sort of just came around to a view that it's okay to include a section that is not every single claim substantiated. It's, like, it's okay that it comes from something.

It's very new. So I used the opportunity to write something. And in it, I write a couple of things which are similar to what I'm just saying. And specifically, it says I would like Shopify to be a company that sees the next century. And that's my ambition. I want a company that's, like, doesn't need me. But, like, while I'm here, I'm trying to also try to be the biggest asset to the firm.

And I think I– I hope I'm doing this reasonably well. What was Shopify's market cap when you were in public? It was sub 10 billion, right? It was one and a half, I think. Okay, sub two. So I think you're doing okay. Okay. That memo, though, I want to come back to that. There were two things that people focused on. It was the bring, you know, AI into the performance review aspect.

This was in 2025, which I think is also why it was controversial. It was people doing it. Exactly. And then there was, you know, basically you have to ask the AI question before you add headcount. You know, can I do this with AI instead? Are those two things still happening inside Shopify? Are those still the rules? Yeah, I think they are happening entirely organically now.

I think this is like, I don't have to enforce these kind of things. I mean, the performance review, I think we went through a few sites. I actually don't know if it's in the current one. There was like a question about, like, this is a peer rating kind of thing. It's like how reflexively does so -and -so show up, like a prototype from AI or something.

It was, you know, sometimes you play a little bit of tricks there. Like, it's like this was in a performance review questionnaire, but of course it didn't matter, like what the answer was. But like the funny thing is, just the fact that it's in there means like, you know, sometimes you can just sort of give a company a reputation to live up to.

And then it, you know, it does. And so there's very big AI skeptics or like even people like, there's many, many open source projects that have now banned AI contributions, right? Like, which is, to me, it's like the source of endless entertainment that people somehow have morality conversations about whether or not the code got hand chiseled by humans or somehow computed by humans.

It's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like,

it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like,

it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it

like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like, it's like a performer in a company where everyone around them is just using

like, a data center of agents to accomplish their tasks. So I'd like, it's just not going to happen. And so just like, make it explicit, because I think everyone deserves to work in a place where they have, where they either are, or have a really good shot at becoming one of the top performers. And so once we know this is no longer a possibility space, we also want to make, help you get back to industry to look for such a place for yourself, right? There's this consensus forming right now with these personal agents that the business model that will support all these is some kind of take rate business.

Mark has talked about it with me for Muse. Noah's talked about it, I think for Instinct. And I hear it with founders all the time I'm meeting with of these, it's a, it's a absolute Cambrian explosion of Asian startups right now. Yes. And they all seem to think they're going to get a slight cut of transactions that flow through their platform.

And that's going to be how these businesses get built. As someone who is running one of these take rate businesses at scale for many years now, do you think that's actually the model? Because I'm trying to figure out, okay, where does this, where does this take rate come from? Yeah. Like, will Shopify, I think you, you take roughly 3%. Will Shopify give? Well, most of that goes to the credit card companies.

Right, right. You have, you have to share that margin. So it's not a lot to fight over or to give. And so is there a world where Shopify shares some of its margin with these agents because it creates more GMV that flows through Shopify? Or do you increase your margin because Shopify is becoming more valuable in an agent world? Like, how do you think this is going to net out? Because I'm like, where does the money...

I'm actually really fatalistic about these kind of things. Like, I honestly just find that, and I think this is the right position to take for the agent startups, you know, assuming they can get the financing. It's just like, these agents are of tremendous value. Like, it's like, once you have this set up, the money you would have to pay someone for them to agree that they can never use their Muse again is like increasing by the day, right? Like, it's like, so on the base of a reverse money test.

Then maybe the model is just charge the user. I think it just depends on how much value you can provide, how many unique things can you do, or like, how much were you in the right place at the right time? How much did you have with onboarding? I think there's... And frankly, what is it, 2% of American households pay for AI right now? Yeah, it's very low.

It's like, it's just so early here. And, you know, I think there's plenty documented. Mark is, since we talked about him, has in many cases built enormously impressive businesses by just saying, hey, we are going to start thinking about making money when we have like 100 million users or billion users or something. I mean, of course, you need to be able to stay liquid throughout this period, which is, you know, another thing.

But like, I'm just going to... Well, this previous era of startups didn't have the compute costs that these ones do. That's very true. Although, you know, and frankly, the analogy is to .com where everyone had to go to Larry Ellison and give him a million dollars to just be able to receive a database. So, like, that didn't go so well. So, it's...

Look, I... The last time I was a startup founder was 22 years ago, right? Maybe 20 years ago. Come on, you're still a startup founder. So, like, I mean, I feel like one, but like, I'm sort of almost in the video game version, the sort of fantastical version of a startup founder. Like, I have a company that feels like a startup that is, I think, as nimble as a startup.

I think that's as fun as a startup. It's, in fact, remembering startup days, they're often not so fun because, like, I, you know, back in those days, I had to ask my father -in -law or who we were living with for a check to meet payroll. That wasn't so fun. So, I, you know, Shopify is doing very, very well. So, I don't have to think about these things anymore.

And so, that's wonderful. I don't know what it's like to actually have to go fundraising. I'm certainly just as surprised at some of these fundraising rounds that I'm hearing about as others. And I added a job as a venture capitalist about two months ago, and it's unbelievable, the valuations. And there's a sense of just throw the money in the pot and go, go, go.

It kind of reminds me of, like, maybe Uber, like, just capital as a weapon. But the capital this time, it's all going to GPUs. It's going to be Jensen. Yeah, it's all going to Jensen. And that's why I ask about the take rate thing because, like, what is going to actually sustain all this? And I don't, everyone seems to think their agents are going to be just taking 2% of transactions.

But I'm like, who gives that? Yeah, you're not going to get 2%. Like, I mean, welcome to the world of basis points. Like, people come in with, like, I am in conversations sometimes where people don't have all of their priors fully arrayed yet for some of these things. But, like, it's, again, the agents are providing a ton of value. And I think people will also pay for it.

I mean, I'm certainly extremely, like, professionally at Shopify and privately. I am very large consumers of AI tokens. And I'm very happy to pay money for them. My situation, again, is a little bit different here. And these things are very expensive and I have resources to do this and this will all get cheaper and address different spending classes at the same or better level of intelligence very quickly.

And so I think, I don't know, I honestly, this is one of those areas where I'm just honestly like unconcerned. Like there's going to be so much exchange of value here that when there's a significant exchange of value there's opportunities to monetize. And maybe it's going to go the way of like electricity metering in the background or maybe it's going to be in a way of transaction flows.

Like maybe it's going to do all of those things. Frankly, just optimizing people's lives. There's like people are spending money inefficiently, you know, like there's a lot of buildings of banks over there and they all can afford a lot of these buildings because people are probably paying more in fees than they probably ought to. So like, and those are easily things that can be discovered by anyone who has access to your email, synthetic or not.

Right. Like, and so, you know, who knows where things are going? I think there's, the world is not as optimized as people think and agents can help people save money. And I think that's the kind of position that I think will always be of value to people. With the tools you have with Sidekick and Canvas and the ability that Shopify now has to abstract away the complexity for the merchant.

Do you think Shopify eventually can command a higher percentage of all this because you're taking more and more off your user's plate? Do you think about that? Yeah, I would, I would, I would prefer a lower percentage and more, right? Like I just like, I, I, I care, I don't care about percentages. I don't care about, um, uh, gross profit.

I care about gross profit dollar growth. Um, and that's a very, very different. Um, like I can get to gross profit dollar growth by making Shopify cheaper and, uh, just having like, like, like having a larger wave of entrepreneurship be, uh, um, Explain why, explain why that's the thing you, you focus on. Because that's the mission. It's like, it's, it's like, I, I, I mean, that, that might sound not terribly, uh, uh, uh, like I'm performing my fiduciary duty here, but like, I, I just find that, um, again, I think, and again, maybe this is also one of those founder things.

It's usually companies that's like the companies that become quite big, uh, started with a mission and often transition to more a custodian ship over the point that has been reached, uh, when they are transitioning to, um, a professional manager afterwards, um, with some notable exceptions, um, and companies that reinvented themselves. Um, Shopify wants, um, Shopify wants just to, wants to see their entrepreneurship to be a more common and easier attainable thing for, for, for more people.

Like, like, the diffusion of entrepreneurship, um, deeper into the economy, um, like, deep, deeper as a, as a, as a plan, more as a plan A rather than a plan B for your career. Number one time stamp for business creation on Shopify is, uh, local time zones lunch break. It really is. It's crazy. It's like, I love that start because it means, it tells you how businesses are created.

Maybe after a bad meeting, maybe after hearing that someone's losing a job, or maybe after a friend of yours leaving, or maybe, you know, just, uh, maybe some, some form of dissatisfaction, use your lunch break to sign up for Shopify because you're going to try the plan B that you've always been thinking about, about for, for your career.

And then you build something, you probably learn something, um, uh, in the process. Maybe you catch lightning in a bottle and the thing that you wanted to build is something that other people deem, uh, of the place orders. And, um, um, you're gaining, you, you, you reach for independence, uh, finds purchase and you end up, um, building a business, hopefully a large one in the future.

But like, it doesn't have to be large. It can be just exactly the thing that you wanted to, to, to, to, to, to have. And so that's what we love. Um, and, um, so. We want to see that a lot. And, um, more and more people are learning about Jevin's paradox throughout the wonders of AI. It's like, um, when things become cheaper, they, um, uh, people will consume more and they'll use them more.

Um, I do, I often wonder if, um, this is an impediment for Shopify. I mean, Shopify is like really, really affordable. Of course, like basically most, most people use it in their first month, probably spend more time, more money on, uh, Starbucks than on Shopify, you know? Like, so, um, but it's, it's something I think a lot about a lot. I just want, I don't, I want there to be basically no impediments, um, there.

So, um, if you're working with a lot of, uh, um, in the AI world, make it so that AIs can sign you up for Shopify, you know, much deferred point. Like, like when you, again, when you use Shopify to gain value, then we want, there should be like the beginnings of, uh, um, uh, exchange of value, um, which then as you grow becomes fair, I suppose.

But like, yeah, like Shopify has always been built in such a way that, um, uh, we have millions of businesses on, on them and, um, they are, maybe subsidizing is too big of a word, but like there's certainly a reason why we can build such amazing software for new businesses. It's because we also have a lot of businesses that are grown to be billion dollar businesses.

And, um, um, um, so we can amortize the, all of our, um, uh, engineers and, uh, tokens over, you know, a large pool of people. And we spend a lot of time thinking about how to make the beginning easier, even though that's like clearly not the source of most of the revenue. So again, my company is very much tied to its mission. Um, um, and it's pretty through and through.

And I think that's important. I was going to ask what you think the risks are to Shopify in this agent world. And it sounds like it's the impediments you could possibly have to people signing up. Well, I mean, like, I mean, no, the risk is much more acute. It's like, um, you know, someone might build a better Shopify, right? Like I always think a lot about this.

In fact, I spent the beginning of every year trying to figure out how I would build a better Shopify. And then spend the rest of the year trying to make it more like that. And so like, this is my, my entire loop here, um, myself in enforcing the loop, so to, so to speak. Um, I don't know. I, I, like, obviously I have the best vantage point, probably a very privileged position here.

But like, I, I, I've, I've always found opportunities to go after Shopify. Um, uh, uh, it's, it's not, it's not perfect. We fall short of our own standards all the time because we are, it's a huge piece of software. It's, uh, it's, uh, we have a massive diversity of industries and customers. And again, this thing needs to work perfectly for someone in the net signing up in a lunch break while also supporting billion dollars businesses, um, um, that, that everyone's heard of.

And, um, uh, you know, that's just like, and product managers starting this company, they're like, holy shit.

It's like, this is a, this is, this is pretty metal. This is what you, what, what, what you're doing. That wasn't quite clear. I sort of thought of Shopify as insert 10% of Shopify, you know? So, um, uh, I'm proud of that in a way. So again, but like, that's also means, uh, you know, we have, uh, disadvantages because we do have a lot of existing customer base and sometimes that makes it harder for us to, to, to evolve.

So we have to operate at a level that we are as quick, as nimble, ideally first to every new possible way of, um, uh, creating significant value for kinds of people we want, like to serve before two people in a garage plus a data center of agents can get to. And that's fun. That's, that's fun for, you know, like, I don't know. I, when I started 2004, we had a similar situation.

Lots of incumbent companies, many of them, you know, back in the day it was tens of billions. Um, uh, not hundreds potentially, but like, um, tens of billions of dollars existed and maybe one or two hundred, I don't know. We knew we had their number. It's like the web was new. It's like all client server software, it's just like, I don't know, very expensive.

We had to talk to a salesperson and probably go play fucking golf with someone to buy a software. I'm like, we know how to do this better. So I think that's the attitude that almost everyone in AI is taking now. But the thing I want to change is like, I don't love calling myself incumbent, but clearly I am, so I probably should. But like this time around, we're going to have everyone around for money.

Yeah. If you think maybe I am an incumbent and you have to challenge that constantly, does that mean you need to self -inflict pain to make sure that's not the case? You have to be willing to, I mean, I think Apple's probably one of the most famous examples of this, the way Steve would talk about it. You gotta knife your own baby sometimes.

Like, yes, it's terrible to have to do this, but yes, sunk cost fallacy kills everyone. And again, what you have to do is you have to figure out what does the perfect version of Shopify look like if I start right now and had a magical wand that would build all software in a nanosecond. And then you have to change, like you have to look at, probably gotta get stiff drink and sit down and then run the diff between the thing that you just like magically created in your head and the thing that you have.

And if it gaps, wherever the gaps are, that's your roadmap. I'm not trying to glaze you, but I honestly don't know if there would be anything different with Shopify right now? You have probably one of the most... Just wait at the things you're launching. It'll become pretty obvious what was wrong. The trick is to be at the solution before other people realize there's opportunity.

But fundamentally, what you're doing and the way you ride with entrepreneurship feels structurally enduring. I would venture to say there's maybe 10 companies in the world that have a mission that can endure like that. Well, and we are part of a massive inheritance here. I mean, people have traded forever. You find like cadavers in the Alps, which happened when I was a teenager.

I think Misadventure died somewhere in the Alps and clearly was a trader with products from like 4 ,000... No, I can't be right. 2 ,000 kilometers radius on him. That was 6 ,000 years ago. We've always been traders. We've been traders. We've been merchants. We've been entrepreneurs. Some of the earliest clay tablets are like, these three people are taking these products to that market and so -and -so is expecting this profit.

These are startups. We are a species of entrepreneurs. It's extremely core to our existence. People have traded and created products for each other 10 ,000 years ago. People will do it in 10 ,000 years. It's a great bet. We don't address all of that, but we're addressing a particular storied part of merchant entrepreneurship here, which we find is deeply important to us.

So yeah, I think it's a nice trick to pick as your market the thing that we've been doing for 10 ,000 years. It's just like... So it worked out great for me.

Coming back to the startups in this space, it sounds like your advice to the Asian startups is to not be too tied up in knots about margin and take rate points and to think more about capturing value and then the opportunities will present themselves. Is that a fair summary? Yeah, and then be so important to people that you're deserving of a value.

If you want to build a company that's successful, you need to build a company that's too good to ignore. And if you want to have a fantastical career, you need to gain the skills and aptitudes that make you too good to ignore. That's the best advice you can possibly have. Even in the sense that it'd be too good to ignore when competing for the limited resources that people have and that they will, I don't know, stop having the extra cup at Starbucks once a month to then be able to afford you.

Like they should be able to... Most people, luckily, at least when you're building a company in the West, you're competing with people spending money on coffee, right? Like that's what you're really competing with, at least in consumers. In enterprise, everything's different. It's so funny, when I was pricing my newsletter a year ago and I left my traditional media job, I was thinking about cups of coffee on a monthly basis.

You're totally right. You were already fully on it. You're exactly right. So, you know, just like people like their cup of coffee. It's hard to get. I mean, we didn't manage here. Yeah. It's like these, you know, these places always are, you know, very curated and everyone's setting it up nicely. Yet I have my just mug here. Yeah, like switch, you know, I'm sure everyone behind the camera right now is upset with, so I'll update our state presence.

But like, you know, it just tells you, like, cups of coffee are really important to people. If you build something that's better than a cup of coffee, you are going to probably endure. And what is your advice or how should the companies that are wrestling with their aggregators, they have large advertising businesses to come full circle here, and they're throwing their walls up to agents or they're figuring out how, what do I open up? What do I not? How should they think about it? The best way to do it is to, you know, like your advertising system and your monthly fee.

Those are all solutions to that you have created at some point in your company's history. But the company exists because you're solving a problem. And if you need to go back to problem -solving mode, if you do that very well, then you will get to a point where you can implement whatever solutions, again, like present themselves for making money.

You seem to be somewhat pessimistic about advertising. I'm not, actually. I think it's very strong. I'm not. I'm just not, I'm not sure how it's going to net out with agents. Yeah. I think it'll probably shift its shape slightly. But like, yes, it will definitely evolve. And maybe it's not even going to be quite advertising. But like the fundamental thing is like there needs to be a discovery and growth engine that's probably run as a market that allows companies like the ones on Shopify to connect with their future thousand true fans, right? Agent impressions? I think there's so many ways about doing it.

The problem, like the problem people are facing with agents, and this is a real thing to overcome, is that it seems to be that the biggest growth for agents is, I mean, it's also implied in the world, that people like these things are agents on behalf of me and you. Yes. Right? Like, and so advertising is challenging here because if it would turn out that your assistant is being paid to make, to choose products, you know, like if, I don't know, this real world stuff happens, they're pocketing the, I don't know, the air miles or whatever.

You know, that's a breach of trust, right? And so this is a very active conversation we are also in, and we love to help, and we have these conversations about like how can we make more socially acceptable ways of doing this, right? Because we are clearly extremely interested in this connection. Like the one thing we always wonder and hope will not happen is that sort of the changes and the emergence of AI and so on is going to cause a lot of head concentration.

Like all the business goes to the largest businesses, which, you know, like the world of department stores and old school merchandising was very much so. It's like, you know, you were the biggest consumer package good company in your space. You had best shelf space. Therefore, you had all the sales and it's like all. So the beautiful thing about the internet and why it's, and why I also really, really, I'm so into Shopify is because it helped with breaking this.

And like it meant much more interesting products could be created. Not the one that works for everyone, but rather the one that garnered a thousand true fans. Once you have a thousand true fans, you created a product that people love and have, or that some people have a capacity to really love, which means you have overcome the largest challenge in product creation, which is indifference.

You can also create something that a lot of people hate. Even that's better than indifference or harder to monetize. Usually you get both, hopefully in proportions that you can deal with on social media. More products people love is really the goal. And that necessitates that there's a fairly liquid market of like being able to bring a new product, a new company that might not be like the most well -recognized company, but bring it to an audience that then they can consider to do I want this? And, you know, so this is like the thing that has worked so beautifully by especially Matters advertising engine, which has been the greatest boon to small businesses of all time, I

think. And also Google's, although slightly less so, but still good to have both of them. And I think this is why we've seen so much information. Hopefully we will see much more of this. Like in a funny way, advertising, other things are quite maligned in a lot of ways, but it's actually really, really beneficial. And it also acts as a little bit of a proof of work in a way.

It's like a business that can't put money behind their product hasn't made money. And in a way, it's actually a fairly good filter, but like buying from especially small businesses that can advertise because you're likely buying into something that's growing and that's needed. And I think this is– so creating continuity there is really, really important.

Figuring out how that works for agents. Yeah, and an incredible business opportunity to figure out. Again, launching new forms of advertising or matchmaking markets is all these– especially with the great primitives of blockchains we now have that are much more mature. And so it's like this is so much opportunity. And so I think this is the fun thing.

Like we get to reinvent everything now. You posted on X recently that you swapped your wine cellar out for a server rack. My rack cellar. Your rack cellar. I promise I'm not going to make a distillation joke. But I do want to know, what are you doing with that rack? Oh, well, okay. So look, there are layers to this one thing. It's like there's a long– again, Harley is my– Did you throw wine out too? I didn't have enough to fill out the room that we already had.

So like this is also– it wasn't technically– it wasn't very well used. And it was sort of central. It's like glass. It's beautiful. It's nice to have a staircase. I'm like– I always look at an empty room. And I felt the need to– like I found at some point, like maybe I should just like do this thing Harley does where he goes like to an estate sale or whatever and just buys a lot of wine at Montgore or whatever.

Like this is so not you. And then I'm like– I sometimes consult my inner 16 -year -old when I'm sort of like stuck. You know, just I find that often more interesting than sort of playing the regret minimization game because that just like usually makes you be fairly conservative. So what would be awesome to do here? And like, you know, I have a lot of tech.

And so anyway, you just got it like, you know, found someone who was excited about it and was excited to do a couple of days of cable management and a lot of procurement and just build something really, really cool. I think service– again, I love computers. It's like this is– you will always hear this from me. I think these things are beautiful themselves and you can just put them in scene.

So now I have a rack seller as Lulu, Christian did. And it has like one of those GB300s from Dell in it. So it is actually like fully training models. And being a public company CEO, it's a little bit harder for me to experience some of the spoils of technology industry with agents and so on if I would have to go out to a lab to send– again, there's lots of things I can't send there.

So now I can– it can all run local. So I have my own version of– The Toby Claw there? Exactly. Something– Toby Claw that has– it's completely local and it's remarkable and runs all in my house. And I think that's awesome. So is the historical corpus of Shopify in that rack seller? Yes. You know, I just– I'm lucky in that I've been pretty good– like I'm sort of an atoms minimalist, but I'm a digital archiving maximalist, oddly, which is clearly a contradiction.

But like now this works really, really well. I just– I basically have everything I wrote and lots and lots of logs. I know what I've done on my computer basically for the last like 15 years. Like it's just like– it's actually– again, I just– given how I spent my last 20 years or even more longer, a lot about me is in the pre -trains.

So like I think my experience using agents is actually pretty close to being ideal because they already– like they already know about me. Like I've played this game actually when I set up this new machine that Claw was running on because Claw like asks you like, like, okay, well, let's introduce ourselves. I'm so -and -so right now. You can rename me.

Who are you? And it's like, guess. And what's the day in the life now with your Claw? Are you getting– I've heard you say before you get audio digests in the morning that you work out and listen to. And then like are you getting like hourly updates from it? Like what's the– It'll run through like a script. I might just finish that story.

After I said– I said– give it a guess. Of course, the user was Toby on Unix and like it geolocated to Toronto and it figured it out. And like– I have a screenshot of it. It's actually– Wow. I was like, fuck, that's crazy. So it set itself up perfectly. Like what are we going to work on together? I love commerce or something it said.

So it was ready to go for a job. So, you know, my use case– I really, really, really am excited about the Claw. I think it's probably the most compelling product I've ever encountered in my life. It does so much. It does a lot of research. It does– it writes briefing notes that when I ever read or not.

But it's– yeah, it runs a script that sort of makes a collection of all sorts of things, including my Fitbit data. Like– and, you know, obviously it has my calendar. It has a little app. It knows my geolocation. Like it just like– again, it's logging. Every once in a while I come up with a completely– like, you know, all the best ideas in the world are stupid ideas, really.

Or at least the best products usually start as a stupid idea. And like recently I realized– you know, I did a lot of– you know, for my hobby, I do a lot of health -style analysis. But then given that I have always data– like I have a heart rate data for, you know, for this. And I figured out who my most stressful meetings are with. Oh, God.

It's like– it's just a query. No. Oh, man. Have you told those people? Oh, yeah, yeah. Totally. It's like, guess– here's the funniest thing I just find out. Guess what? And then they didn't think it's nearly as funny. That stresses them out. So anyway. So it's just like– I don't know. You can just like do funny things. That is a terrifying thing to hear, Toby.

So I don't know. I just think it's all funny. And– You can ask it how your heart rate was after this. Yeah, yeah, yeah. Exactly. It's like, you know, who's– maybe I'll send a report card about my heart rate to my podcast hosts. Oh, my gosh. Do you feel like you can't live without it now? I can live without it. But it was just like– I would be like– remember the days when we had this amazing technology? Like it would feel like one of the most significant regressions.

And this is why I say like I think dystopia that people fear is always just like 30 years ago. Like if you actually would go back there and realize all the stuff that's missing that you really, really appreciate, I think this would become more clear. So which I like to point out when people are saying, oh, we're going towards a dystopian future.

It's like, you know what? Like let's actually just like not– let's just like keep this train going because it's been pretty good for us. It would really, really suck to lose it. And I think you can adjust to anything. But there would be a lot of things I would sacrifice on the altar of thought experiments long before my claw. Well, Toby, those are my questions.

Thank you. Been a pleasure. Thank you very much.

Sources for 30% off. Rules and restrictions may apply. Granola is the best AI notepad I've tried. It works everywhere on a video or phone call in person or an Apple Watch. Try it now at granola .ai slash sources and use the promo code sources at checkout for three months off. Jira by Atlassian gives teams the context, visibility and control to build with clarity and ship with confidence.

Try it free at jira .com. That's J -I -R -A dot com. Banking should feel like modern software. Get everything you need in one place. Visit mercury .com to learn more and apply online in minutes. Mercury is a fintech, not a bank. Check the show notes for details.