Tobi Lütke – Building Shopify and the Future of AI | Ep. 50

I think in a way people are somewhat overestimating the founders of companies and then they are really massively underestimating what you can do when the founder is still present and in charge. It's my best way to help the mission. Everyone gets to complain about the crazy founder, which is great. Do whatever venting you want and then you do very often some of the best work of your careers. You know, that's why people actually flock to these type of companies because you're surrounded by the other people that you can go on such crazy journeys with.

That's such a wonderful question. Thanks so much for having me on the podcast. I'm super excited to get to do this. Part of the reason why I wanted to start a company is because I have found before in my teens, but I have an extraordinarily hard time learning things that I don't have basically experienced the problem that they solve. Like it's just like the way in school how math was taught basically. Here's just sort of like the steps for to like I actually had to really go and like figure out okay well what exactly is trigonometry useful for? And I just had to like I started with tricking myself into finding a useful problem that involved trigonometry after which I felt I could learn it significantly faster than everyone else because I just was so motivated. Right? Like Kapoa said, one of the joys of life and one of the best things in life is like to find a beautiful problem that might occupy you all of your life trying to solve it. And if you're so unlucky to at some point do solve it that it will have plenty of delightful problem children that you can dig into. Right. First of all, I think that's a beautiful way of thinking about life's work. But also it just works specifically well for me because I love learning things. I love challenging myself. I find myself very, very interesting. I just like because almost my favorite thing is when I want to do something and I can't because it just like you kind of find the limits and then you can have this conversation with yourself about what you know because there's a very simple recipe to success. It's just that everyone I think sort of into it or knows but maybe not spend too enough time sort of following through on which is success is really, really simple. It's just you have to figure out what it costs and then you have to be willing to pay it, right? Like and very rarely is this comes in the form of money. It usually comes in the form of time commitment and discomfort and these kind of things. So like I seek problems. I love computers. Like to stay with the example trigonometry. Turns out the very early video games like Wolfenstein 3D are basically just trigonometry. Once you realize this and like play with these things they become delightful and you learn the next thing and the next thing. So I wanted to start a company because I was just like it's a right time of my life. I just moved to Canada from Germany. I didn't have a work permit so I couldn't work for anyone. I needed something to do. My wife needed lots of time for her degree and I'm like this is a good time to try it and it's probably not going to work but I will learn a lot and so this has been a powerful way for me to motivate.

Yeah, look, I certainly collected barnacles around certain times in my career too where just like I sort of fell into the trap of like trying to live to like people had particular expectations not really like more like in the department of aesthetics like I can only call it an aesthetic because people say things like this is what CEOs are like this is like the behavior that a CEO should display. You're supposed to be a statesman. You should travel and kiss babies and not. I'm like, well, that sounds pretty inefficient from my perspective, but like I guess at some point, you know, you just sort of get riddled down and like you try all these kind of things. My life was miserable, too. Again, this is a sort of junk science a little bit. But like what I'm in pursuing here is trying to make a beautiful product. I just think like we need to create products that are just like joyful and like one of my favorite quotes is a quote by Kathy Sierra of like don't make better cameras make better photographers right like and it's just like very, very deep to my psyche I feel through beautiful tools and through beautiful you can just inspire people to be the best version of themselves actually become like induce more ambition induce more skill or at least induce more ambition for yourself to develop skill beyond what you otherwise would do. Okay. And so I think that I want this to exist and this is just like sort of my guiding post here and it throws off all these problem children which are interesting which then challenge me. You're talking a little bit about sort of calm like I'm not a terribly calm person either but I actually don't want to like dial down my calmness. I want to channel it into building something. And so I've just found that almost all the mediocre products in the world, they remind me of room temperature, right? It is like it's just like it's sort of middle temperature of like the sort of what you get then no one really cares. Yeah, exactly. It's a default setting of a thermostat. Almost every great product is forged in some kind of furnace and some kind of temperature and like you got to put up into every energy source to produce that kind of heat.

I don't know. I don't know because you seem to love your customer. I think the greatest gift people make great products like I think without that crutch. I think if you have it, it should be an incredible boon.

It depends on what great CEO like I don't know. I'm an engineer. I read lots and lots of books about how to make great software. And you know what? Like many of these books, it's really funny are like when you actually check what was a project that they all worked on when they figured out the design patterns of modern software. It's like the Pennsylvania payroll system, right? Like it's just like not the most inspiring project. And but like it turns out I think just work is what you make out of it. And I think you can like if you just default to like hey let's actually build something that's like really, really meaningful and we all learn a lot about I think it can be done in any event. I have a massive benefit and really one of the greatest things and I tell people this when you know in interviews when people are thinking about coming to Shopify I point this out it's like one of the greatest gifts that this company has is all of our customers are inspiring they are just like remarkable people on doing an incredibly courageous act of starting a company themselves. The people who flock to then come to work for Shopify, the people who actually have experience starting a company or have maybe a family member or at least a deep appreciation or in fact. Yeah, in many cases there's just something wonderful also about, you know, Shopify feels hopefully very, very fresh as a company and there's no like tree rings that you can sort of read about its age, but like it has been around and so therefore when we do our annual summit or so where we all get together in Toronto for a week, there's going to be people who work at Shopify who weren't born when I started the company like and it's just like oh yeah, this is like been like little like people are being kids are so capable at young ages of being productive.

I mean no in fact what's even better is like be different because like again axiomatically if you are building the same thing other people build it can only be similarly good. It can't be actually much better. It can maybe make slightly look nicer but like you're bounded couple percentage points either direction. If you want to build something great or much better it has to be different. Now if it's so this has to be your starting position. If it then either converges on the same thing you have learned something potentially from first principles about why the solution is the one that everyone has converged on. If it gets worse, you actually learned something more important because now you know, hey, your theory here was wrong. You had an assumption which now has been validated. That learning is the thing from which you're going to pull so much value in every other realm now because like hey you now have a clearer idea about how things actually work. I think it's like the null results in science are massively underrated and so ideally then you don't ship it because the world needs better not worse right like so and I think that's actually we have tried to eliminate the term failure in Shopify and just call it the successful discovery of something that didn't work.

I mean like look you can't help but be affected by what you see around you and if like it's the best and the worst about the valley is that everyone's work on interesting things but of course that causes prior to be pre-installed when you start on a project. What's amazing is like and you see children interact with things like AI and they will use it so different from how you imagine like it's worth aspiring to having this like free mindset of like just like try to take orthodoxy or the obvious path off the table when you start because the forces of especially teamwork will always cause a convergence on the safest path right I think it is an advantage to be outside of the valley you just have fewer pre-arranged prior. In fact, a really funny effect was when I went to the valley as a visitor and met with people and for coffee and we were talking about I was trying to figure out how company should work and I ask questions and then I went home and like I had the entire flight to make my notes and like I again have sort of a bias towards do it differently. So I take what I hear and try to figure out okay well what would a Shopify version look like and what would be better of this system and sort of make it different because I felt that what you have to do but then I realized only very late in my career that like I never actually got the real story from everyone about how they work internally. I got everyone's ambition or highlight reel right like because that's what they really share. That plus the ability to then make edits to try to improve it further like meant that very often we actually found ourselves doing the things that we thought someone else invented actually they might not never actually have implemented it. Maybe that was just the thing that they just had the most previous meeting over and so I think that helps a little bit distance. I mean additionally I should say I think the word fundamentally Silicon Valley specifically has now for a decade and a bit declared war on any kind of distinction. All the talk about diversity was very much about eradicating kind of eccentricities and distinction and you know just like people are not allowed to be just quirky or funny with offcolor humor so I think the inverse that is a little bit more intact and you just kind of encounter characters and there's like often the more appreciation of like you know so and so is just like a little bit crazy and you know what that's really good I think that's coming back again a bit and so that's going to I think help a lot.

I'm actually really proud of where Shopify is with this I think this booked extremely well I made a choice is there's a type of situation that you get in running companies and large like I think any large group where something becomes clearly true and then you need to make a decision on like are you first of all are you do you act on it and sometimes people are just like failing bad right like in fact very often people fail the figure out what's true part you know like I'm sure Blackberry thought they were doing really, really I think they had the best year of sales ever than the year iPhone was released and thought they were pretty safe. So, I mean, and sometimes it's not that people are just not smart enough to see it. Sometimes they don't want to believe it or want to see it. Exactly. They know the implications are they're not predisposed for it and they kind of maybe for no fault of their own created an organization that like sort of every layer inside of a conversation like just sort of prioritizes kind lies over hard truths often because people just sometimes because there's a culture of everyone being nice to each other. There's also when you have a big enough organization, you talked about this with sort of like teams, you know, make originality difficult. I also think big groups of people, it's hard to get everybody to agree to go through an uncomfortable change. I think in a way like people are somewhat overestimating the founders of companies and then they are really massively underestimating what you can do when the founder is still present and in charge. It's not so much about the individual as in about the almost the piece of infrastructure the slot of having the founder slot filled and it's a reward distinction but it's actually as the founder you get so much social credit as having start the company that like you can just invest it's a bank and it's just like every time someone onboards they hear how the company was created and that just deposits a little bit of credibility tokens I suppose into a virtual bank account that is hard to reason about but sort of virtually exist and then I get to cash that in for you can spend it on big important changes and change management right like it's like that's one of the best ways I can speed up something that would take years of small culture change or like internal training with sometimes a memo and so in this case like I find I take that really seriously it's like not the easiest thing for it just like leads to you know just like I mean more work or sometimes unpleasantness I just find that is what that I owe the company, right? Like it's my best way to help the mission. And so when something like the AI thing becomes true and we saying, 'Hey, we have like two people. They're both equally good programmers 15 minutes ago, but one of them like is just like completely onto the AI train and just like has sort of back when I wrote this is really hard to actually get real value out of AI. It was more like the premise of AI was like what do we do in performance reviews?' And we just like the moment we said like it's like we can't we it's impacting it's called like net impact reviews they caught with Shopify. So like what's your net impact on the company and the mission and so like it's just like very demonstrably true that one of the people was more of more impact and the moment that is said it is feels like an incredibly unkind to not tell people. So just like hey let's write it out and send it to everyone. It's like unfair if you're like this person has skeleton on and I'm not going to point out hey that exoskeleton thing can be yeah if we act on this thing we should tell people and so we did and like I included a bunch of other things that are true like I've invested lots in making sure that everyone has everything they need they have an unlimited token policy that I'm sure your brother is thrilled with. Yeah, exactly. And so and so you know like we want people to tinker, we want people to play with this, we want people to use it much and it's super good. It skyrocketed the second this message came along but what Shopify did with this. The speed of the fusion of this tooth was like remarkable and I think Shopify is like predisposed for this. We have thriving on change is one of our core values and we really mean it.

I think it's fine. I mean at some point we like had a leaderboard and so on like of token. Yeah. Yeah. And then of course that leads to immediately bad like effects. And then when and we don't have that like on void is our internal system like of internet. It's probably not a term that anyone's ever heard but like it's so it's our internal wiki and everything. And on your profile it shows you how what your token usage and which percent you are in your department and group and so on just like because that's interesting. I think again also because we are tracking it, we have to be because we have to actually I mean at some point we have to allocate finances to opex and whatnot and so just therefore we show it to people so that I know

It's a huge deal compared to revenue. It's like it's unbelievable. It's like many percentage points kind of thing. It's extremely high. It's many tens of percentage points for some of these private companies. And so I am a little curious. You know, right now we're in that part of the journey with AI coding where it's just so valuable that people are just like, I have no choice but to spend because it's too productive. I do think that that won't we can't spend 70% of revenue on AI tokens forever and have valuable companies. So, how do you think this might go? It's complex in a way that like I'm very, very grateful for the stage Shopify at because we are a profitable public trusted company and right now it's just like we really, really, really like the tokens we're buying. It's they're incredibly valuable and we are doing incredible things with them. It accelerates us in road map and therefore ambition and all. I think it just causes so many leverage on your spend. It's just like it's a no-brainer. I have very, very high opinion of markets. Markets are extremely good. They will figure out what the correct clearing price for these tokens is and right now there's few providers and maybe there'll be more in the future and there's all sorts of interesting moves around distillation and so on. Like I think companies will know how to wield these tools within their budgets and we actually doing a good deal of this kind of thing as well but we are still just like charging ahead because frankly we like the tokens we're buying. It's like it's that simple.

Yeah. Honestly, I don't no one knows. It really no one knows. We are certainly I think we are 10xing the amount of tokens we want every year right now and I'm sure that's going to go up and we are 3xing the amount of GPUs that we are putting into the world. So that like is those lines are not converging anywhere good at price savings. So given that AI coding is dominating so much, what are the biggest changes to like team design like you know there used to be this EPD triad and it worked a certain way and there were you know certain ways that you know road maps were built and reviews happened and you know people would go out and talk to the customers and bring that back and what is it now like what is sort of like the most fundamental molecule. Yeah, I think the small team is my bet. Like it's like the three, four, five people team. Which actually is like funnily enough has always been Shopify's bias. This is quite gratifying to like the reason why we had to very often go past it now is because you just need a lot of specialized skills at least for certain moments on the teams. Your example is a great one. Talk to the customers. It's just like having someone who's like actually like does the customer research and just like talks with people or like what we always did is like collocated people from the support or we're looking at we routed them all the tickets and put them on the teams which was an amazing way to do it. Now just like the sort of agentic harness around our teams, it's actually routing like really, really good summarizations of what our customers say automatically basically back and so that's now available to everyone and then everyone can do more skill like again we everyone is a seven out of 10 on every scale now. So that's like really, really helpful because it allows to make teams smaller. The thing that we are working the most and thinking the most around at Shopify is that I'm big on pace like pace has to be induced otherwise it's received. Parkinson's law is one of my most recommended books and I have a 1970s 60s copy of it that I give to everyone my executives I own many of them of the original run because it's like so meaningful. And the law is basically time work expands to the time allocated to it. The book is 60 pages and full of these kind of wisdoms. This is the most important part. This is basically one of the most important functions of a leader is to just compress time windows. Yeah. To the plasma. It's not that you just give any deadline and anything happens. But like if this is why it's really, really helpful to be very technical and understand all the like sort of tasks and skills, you can ask for something that has like a 50% percentile chance of being the right ship date. Then everyone gets to complain about you know the crazy founder which is like great. It's like do whatever venting you want and then you do very often some of the best work of your careers. And you know that's why people I think flock to these type of companies because you're surrounded by the other people that you can go on such crazy journeys with. The main point here is like I run the company by the six week review cycle where we go for all the projects and spend with time with engineers and champions and PMs that existed to set a pace ceiling I suppose floor of a six week cycle which was faster than we instituted it because if you don't do it you are run by a quarter sometimes you see like the moment like the moment you see in a powerpoint first of all first flag Second flag is someone uses the word H2 or H1 which means you know first second half you're actually you really have to do something drastic. Yeah. Yeah. Exactly. So you do this. And now I think actually six week view is actually way too limiting. It's just like we can do so much more and we're trying to figure out how to like what is replacing this.

No, I and again our customers are wonderful and they are entrepreneurs. They're courageous people who are putting this out there. They build businesses. They create employment and it's like in one way it's like a it's clearly like a particular slice of people who would do this but like it's actually much more bigger than people think. First of all, it's incredibly diverse. It's just like Shopify's customers go exactly with a population map. It's like great businesses in the most in the smallest communities. It's like well it's I mean it's a point of sales doors in the town center of tiny tiny tiny little townships as well as Alo Yoga and so it's incredibly different and it's cosmopolitan it's big business billion dollars business it's the people who are trying to build a thing in the lunch breaks to make ends meet or actually because they have an ambition that they want to become entrepreneurs so what's so funny with the way the AI the conversation is projected or how it's reported and sort of what you see on social media and so on is like it just doesn't like we can't access it anywhere like in anywhere we look you're saying Shopify's data and experience absolutely doesn't what we hear from everyone is like hey you guys fix computers it's like they like you techies talked about computers being these incredible things that can do anything and just like and it was so complicated and so and then we try this and they like I don't know what you guys just sound unhinged and now we have we can talk to it and it just does the thing and it's incredible. It's like just works with me and it's like I've expanded my business and I've hired all these people now like and so I mean it fits into Shopify's vision because we want lots and lots and lots of small companies and by the way 60 70 80% depending on country of people in the economy work for small businesses they are incredibly precious and important.

I think you should sign up for more, like you should follow your ambition further. I think our customers, like if you would, Paul, I think they believe in a permanent upper class. I think we're going to get to a point where just many, many more people can self-actualize. There's two pieces of data that I just find incredibly meaningful. One is like every 36 seconds someone gets their first sale, which I just think is incredible. While we're talking here, think about what that means for how many people just became entrepreneurs. The other one is more of a higher-level observation that every single time we ship something where we know it meaningfully changes something about the early journey—the sign-up, the kind of complexities, the questions, the friction in the business—each of them can be best thought of as a hurdle that someone has to jump over. And every single time we manage to make the hurdle slightly less high because we made something just vastly better—we now let you register domains or whatever, or easily transfer them, or in fact these days have an AI that you can share your browser tab and it helps you set up—every single time you do this, more actual business comes out of it, which then provides employment and so on. People turn out early in the process if something happens that ends up being a governor for them, and then they null out, they give up, and they stop, and then the entire business doesn't exist. AI is just—never has there been such a thing that can take people so far.

I mean, it's my favorite replacement idea as a replacement for a Turing test, which of course we sailed past without, oddly, very little notice. I think it is crazy. I think acting in the real world, starting a business that people find meaningful enough to vote for to the tune of a million dollars, marketing it correctly, getting the right sources, knowing what things to prioritize, the shipping matter—I think we are actually getting there. Ironically, through a provider role, like you can obviously use Shopify without having products. We help you find manufacturers if that's what you want. We help you—there's an entire thing called Collective where manufacturers offer their products that you can then use in your Shopify store. So if your particular skill set is the marketing, you can come to Shopify and try your hand at entrepreneurship. And again, very often, I think about half of Shopify stores end up being created by people who have done an online store or at least a business before. So people just try and build things. People should have a product that the world wants. Ideally, come up with some unique take on something, and there's so much white space out there.

I think AI should then do absolutely everything else. In fact, it's actually literally what our product's ambition is to be—like maybe the vessel for AI being the brain or the exoskeleton around the model—to basically conspire to just do absolutely everything so that if you show up with a product, you can start a business.

I think the—I mean, anything in digital products. Make me some ebooks. Exactly. You know, actually books, there's print on demand for books, for t-shirts. Like honestly, additive manufacturing is getting extraordinarily good now. And there's lots and lots of great contract manufacturers for this kind of thing, like CNC and 3D printing. And then we're looking at humanoid robotics, which kind of—there's a lot of tailwinds here to the fact that making the products is going to become lots more tractable. This will probably be 10 years till I'm asking this, but the prompt like 'build me a house' and then the robots just go dig up what they need and then put it all together. I think I don't see how it couldn't get there, right? And I just think this is what's so—but what I dislike about the doomer conversation or the permanent underclass or whatever people want—how are we not headed to a world of abundance? Exactly. And frankly, that then goes into this sort of 'we eliminated all the jobs.' It's like, you know how good we are at making up jobs. Yeah. You know, like there's so many examples of things that are just delightful. Like I have friends who—OpenClaw came around, which we didn't think we needed, and then it ended up being one of the most compelling things. And one thing led to another, and now they have a warehouse full of 3D printers and things because it just suddenly—you have a Jarvis that you can say, 'I would like to do this project and I bought this crazy 3D scanner from Facebook Marketplace and I just put it—I don't even have software for it. Just go hack it and figure out how to help me.' And it's just like—I know. Holy crap.

Absolutely. It completely is, and I think this is actually the missing ingredient, right? Because people talk about the incredible infrastructure we used to build in the 60s, or just even before that—there's the Hoover Dam, this kind of stuff. Lots of individual stories about these infrastructure projects, how quickly we built subways, and now it takes forever. There is an observed degradation of a lot of these kind of projects, and it's multifactorial—there's not one reason behind it. But one of them that really is true is because we are building the modern wonders that are no less impressive entirely in software. Like the web browser, Linux—all these kind of things are projects at a scale beyond what the permits are, and so easily and without compulsion and just by volunteers, sometimes people who have never met each other in the open source world or at companies building these. I mean, something like Google or the social networks—they're incredibly, incredibly impressive pieces. If they would have physical manifestations, they would be the most impressive. You would—again, you go to something like a refinery and you see the pipes and so on, it's so impressive. It's not even in basis points territory in terms of complexity compared to a browser. Exactly. And we don't appreciate that, and neither—that's not our problem.

I would put it under—it's one of the wonders of the world. I mean, for so many reasons, like it could never ever be introduced today.

In so many ways. Like, okay, you go to a website, it's like we don't trust that website. We just run code. But also, it's like actually your computer in front of you magically reconfiguring itself into someone else's vision for what should be there, and like without limits. It's actually like you download software that then just exists for this one moment to do literally everything, and it is self-responsible. Like Shopify is a hundred—over hundred billion dollars market cap company with hosting millions and millions of businesses that might otherwise not exist, and all of this just happened because anyone can just put a server online, gets an IP address automatically, and then it's like—and it all hangs together. Why doesn't it all crash? It's like it's the most reliable thing. It's just—and no app store on planet Earth would allow the web browser if it would be introduced today if it didn't exist, right? It's just like no one would allow this because it just doesn't even sound like—it sounds like an insane pitch, yet it exists and we just don't think about it. And so it's just also like the font rendering alone is like one of the most complex—the fonting alone is a Turing-complete system that just because we want to be able to read text slightly better than our displays can allow them, and it just keeps going. And I can go full nerdcore on this, but the larger point is actually the following: we have not stopped as humanity building incredibly impressive infrastructure. What had changed is the infrastructure that needed building over the last 30 years, and the reason why other infrastructure didn't happen is because the people who quit—it's just so much went into this—into what planet Earth needed from us, which was a digital infrastructure. And it's like if all those people worked on robotics—it's like this is happening now, right? And it's like we are—we are—this is happening with AI. We are happening with AI. Like all the software we built was a bootstrapper for AI. Software becomes, due to this achievement, something that can become personal again and become like—now you can have basically a web browser but the websites don't even need to exist for that that you would like to see.

I think the iPhone is a lucky exception in that you can appreciate it as a physical object. Because it's not the fault of—everyone is right when they say, 'Why is everything just kind of standing still or not getting better around me?' Because from the perspective and from the ability for us to observe it, it is because all the infrastructure is digital. But we are almost done with this. We are at the end of the opening chapters of all this. And so now I think we will see vastly more impact—all of our energy against this thing, because it was the foundation. But now all these people will in some sense be free to not write software.

I mean, if my friends would listen to this question they would laugh just because like it's my predominant obsession is to try to have as many data points on as many people which I then can try to match to the right superlinear or linear or sublinear curves and just figure out how they all connect and what therefore will happen in the future. Right? Like it's basically—I mean to me it's the most fun game in the world is to have a pretty clear-eyed view of what the future is like.

I mean, the AI memo is a good example. Because again, it was probably slightly too early to write it, but now if you read it now, it says nothing that would be surprising, right? So being able to give my company the gift of a head start—I mean, that's what we're trying to do, right? Because the moment then I can tell everyone something that is not quite clearly true now about the importance of these systems, and we can rebuild our systems to really reinforce this and support everyone in their own tinkering and exploration. That means that our view of the future is going to be more accurate in the things we are building, because you're always building for a future point. Software is getting faster to build, but you still have to aim at a future point of value. It's not our customer's job to tell us what they need. It's our customer's job to tell us what the problems are that they're experiencing, and we fall in love with those problems, sort of adopt those problems as our own and just solve them in an ideal way. But that's our job to figure out what the ideal way is that comes into contact with the customers, and they give us further feedback for refinement. But it's, in my mind, a complete abdication to just build what the customers ask me for. It's an abdication of product responsibility. And so it's hugely important, and it's kind of like almost everything—every decision I make, I try to make on an understanding of—I try to live in everyone else's relative future. So this is—I do a lot of it. While it's incredibly fun and helpful to talk with the labs, and sometimes there's some really important information that you can get from this. What I found, and we've sort of touched on it earlier, is one of my favorite things, especially for product teams and engineering, is to hire people who actually know Shopify really well from outside. People who were merchants are some of my best product managers because they actually understand what the software feels like when it's being used. And same with people who built apps on the Shopify platform to then come to Shopify and help make the app platform better. It's like I actually don't even think being in a lab is actually the best position. It's actually being like using everything and paying a lot of attention to how everyone else uses what the labs release, and then being in the sort of conversation that usually happens on X on what everyone's figuring out. Maybe actually building something and seeing if everyone finds it useful. I think this sort of learning by doing is just—that's actually how to get the most clear view of how everything works.

If you do it over a while, you get trajectories. And the trajectories are the point. If you're exactly on the state-of-the-art meter of the best thought on stuff and you have other data points, the future gets very simple to predict. Now, it's the hardest time ever to do it because right now the time horizons are so short. I did this very same thing throughout my entire life. And it used to be that the future was absolutely trivial to predict. Like you just looked at the first couple numbers on mobile browser usage and you just knew what's happening on a cell phone and that we needed to make mobile websites, which sounds insane now, but that was at some point people didn't believe that.

So obviously the labs really care about programming because that's the problem they need to solve for themselves, and again, it's always easier to build for various reasons. So like Opus is unbelievably good at programming, and right now it's easy to go from that and then just assume it's equally competent in everything else. And quite often it isn't. If you want to discuss like how to do a public talk or something, it just doesn't have great theoretical rooting of this. I mean, it knows about all the different ways to construct tension and so on, but it's not able to then look at something and make it meaningfully better, where even an incredibly well-optimized piece of code it often finds abilities to go and do better. What is happening a lot is that we're bringing more types of problems into the domain of programming, which is really what OpenClaw is if you really think about it. It's like make it a file system, give it tools like a programmer uses, and make a couple of files—just tell it the soul and the memory and so on, and then it uses the normal programming tools to interrogate this. And because it stays a little bit in the domain of programming, it now actually is remarkably better at this other out-of-domain thing, which is really, really funny. I think we should assume though that we are going through the entire radio chart of things that are valuable and bring them to the same point with programming. So I think there's a bit of overhype in having to work so hard on bringing things into the programming domain, because all of this is just going to get much, much more natural and easier as the models improve. Where it's underhyped is just in deployment in companies, I think, and just what role it should play. There's a huge—no one's using it enough. No one's using it enough, partly because everyone just starts with like, 'Hey, help me do the things that I've been doing all along slightly better,' or actually vastly better, which is valuable, obviously. But if you do the more first-principle things, like, 'Hey, if AI would have been around ever since Alan Turing first wrote about it, it would just have worked and we'd be just in the presence of these superintelligences all the time, and we would just invent the job we're thinking about right now—like how would we do it?' Just invert the whole thing. Yeah. And then you're like—it's just so much more fun to have those conversations. And I think it'll be—I mean, I'm exceptionally excited about this because I think it will create the environments that the most creative and best people will really, really appreciate, because everything just around them happens at the speed that they want to work, and that has never been true for any company that requires some bureaucracy for doing stuff.

It's fascinating. So this is actually really changing all the time, and actually I've had to change my mind a couple of times. But I think the best distinction point is—Shopify is 20 years old. I think the average age of Shopify is somewhere like late 30s, which is I think still pretty good. But like, this is one thing you have to look out for as a company—if there's some companies that just age themselves 12 months in a year for various reasons, and that's probably a problem at some point when a lot of new things happen. So just massively restarting the internship program has been really helpful. We take a thousand interns a year from close to Waterloo and work with Waterloo. So it's great. And just making sure that the interns are not just the students now but also the teachers, because again, they are just so AI native that it was really, really helpful. It's also interesting because initially I had the thought that there's like fluid intelligence and crystallized intelligence—basically knowledge and curiosity-driven willingness to learn. And often early career you have more fluid—you have no knowledge, therefore you're all fluid. And then I was sort of—super new, but the first people who just flocked to it and immediately got value out of it was super, super exciting. I had the thought that maybe that really tilts into their direction. But then as we started—the more the system sort of got firmed with the coding harnesses, Claude Code, and all these kind of things, you really—programming is so much not the type of typing task. It's really understanding the problem deeply. And I think every programmer in the world massively underestimates how much they're doing steering in AI right now and how much that is really them being creative and actually just like—it's uniquely theirs what they're coming up with. Like just having seen the movie before, you can sort of spot the AI just going the wrong path and you just like one, two words can completely change everything. That's a very, very long answer to say I kind of don't know, but I actually just think good people are good. Like there's a bit of variance how quickly people adopted the tools, but once they did, everyone just falls back onto their—

Yeah, I mean, I think the best way to deal with recruiting is to build a company worth looking for for the best, right? So I think people are thinking about it too much as selling and not enough as marketing, or at least even just information. Yeah, it's like how do I look super healthy? Well, you could use this, you could turn this angle, you could do—it's like, well, you could also just try to be healthier. Exactly. So sometimes things are just simpler sometimes than they seem.

Yeah, just don't admire people in bureaucracy. Give people the space to be creative and have them—allow them to fall in love with the mission actually, and have them understand what problems people care about and what impact the product has. I think Shopify has a lot of intrinsic advantages here, but the best way to attract talent is make it so that when people sort of try on the idea of maybe coming to Shopify, they come by, they meet everyone they will work with, and those are really, really impressive people that they will work with together.

I find it a little bit sad from a perspective of—I totally understand why individual companies do it. I just find it like, man, I'm so glad that so many people ended up buying Shopify stock for them. Exactly. It's just like—I can go basically anywhere and just meet people and they tell me that they just bought some shares at some point and it was really, really important to them. It's like, it's cool. I just want people to make money with Shopify because I'm very much not like the sort of one share, one vote kind of thing—the influence part. I'm like totally—what is that all about? Like just found it, let's go. But I think there are tickets to participate in something you believe in. I mean, look, my view of money is money is how you vote for what you want. And when you buy a product, you're voting for that product and everything that caused that product to exist. And that works with shares too. Like if you want to hop on a company because it does things that you agree with, then you can buy them and you can go to another company if someone else captures imagination better. And I think that's a wonderful institution. So I just also—I feel like it came from a meme. I think because I have lived the other side of this kind of easy to go public.

It's just like I don't—I mean—

Fair enough. I just never really saw these things as bad things. They induce a diligence and a data-drivenness and kind of a little bit of a set of responsibilities which I think are worth having, because you're responsible for thousands of people's jobs and an important product, right? So—

It helped us in a way that maybe is not always the case, especially as a Canadian company listing on the NYSE. We were just very small when Shopify IPOed—it was like a half billion dollar valuation, right? So—

Yeah, yeah. And therefore I find lots of people who tell me that was a good investment for them. It gave us legitimacy to then go into—we never wanted to go upmarket, but we always wanted to have a product that can work well for people at all parts of the scale. That just sounds like just words, is about 100x harder from building a product, right? Building a product that actually scales across the entire thing is insanely difficult to do and has been one of the most fun challenges to pursue. And so bigger customers—yeah, you're a public company, that's cool, that's different. And frankly, there's really, really good people who would only work for a public company. So that's also helpful.

Yeah. Yeah.

There's so many books I love. I mean, obviously I'm a huge fan of sort of short, incredible books like Parkison's Law and Lessons of History. And just like when people distill all of their knowledge into under 100 pages, I think that's great. I recently read a really good book called 'What Is Intelligence?' It reexplains basically all of biology from a perspective of how important prediction is and how it probably emerged just in a sequence. I'm like, this is like—felt existentially profound in a way that I don't quite know if that was because it felt—this part was just to support an argument later that wasn't actually that important to make. I think so, it's kind of interesting.

It's a job of the author to make you keep reading. Like this is an important mental switch, right? Like if you read a book and it just doesn't capture you, it's not because you're broken, it's because the author didn't do the thing that they set out to do. I sort of have an advantage—probably advantage at this point—of being out of cycle with my wife who also is a light sleeper. So I just go to bed and she wants it just way too early, usually at 10:00 or whatever, something crazy. I'm a night owl. I don't need a lot of sleep luckily. So I just have like hours of reading time, and the Kindle is not that bright, doesn't wake her up, and therefore I have a couple hours of reading time.

Yeah. Yeah. It's important to—like, I really, really am a huge fan of a Kindle, just not because it's great but because it isn't, right? Because it's such a limited but actually single-purpose device. So yeah, I find that wonderful. I just find books to be so remarkable.

This was really fun. Thanks so much for having me.