Tobi Lütke - Building Islands of Innovation - [Invest Like the Best, EP.393]
01:06:49 Share

Two fun facts about our newest sponsorship partner, Ramp. First, they are the fastest growing fintech company in history, reaching a level of revenue in five years that I can't quote exactly, but is eyebrow raising. Second, they are backed by more of my favorite past guests, at least 16 of them when I counted, than probably any other company that I'm aware of. A list that includes Ravi Gupta at Sequoia, Josh Kushner at Thrive, Keith Raboy at Founders Fund and Coastal Ventures, Patrick and John Collison, Michael Ovitz, Brad Gerstner, the list goes on and on.

These facts demand the question, why? Having been personally obsessed with the great businesses through history, one clear lesson is that the best of them are run by disciplined operators. These operators manage costs with incredible detail, and they are constantly thinking about how they can reinvest every dollar and every hour back into their business. This is Ramp's mission, to help companies manage their spend in a way that reduces expenses and frees up time for teams to work on more valuable projects.

First, on expenses, the average American business has a profit margin of 7.7%. This means saving 1% on costs is the equivalent of making 13% more revenue. The average ramp customer is able to save 5% on their expenses each year. Of course, every entrepreneur is looking for ways to grow revenue by 50%. They should just as seriously seek to save 5% on their expenses.

Second, on time. Unnecessary complexity is why most finance teams spend 80% of their time doing operational work and only about 20% of their time on strategic work. Ramp makes spend management very simple by handling your company's expenses, travel, bill payments, vendor relationships, and even accounting.

It's notable that some of the best in class businesses today, companies like Airbnb, Anduril and Shopify, and investors like Sequoia Capital and Vista Equity are all using RAMP to manage their spend. They use it to spend less, they use it to automate tedious financial processes, and they use it to reinvest saved dollars and hours into growth. At both Colossus and Positive Sum, my businesses, we've used RAMP for years now for these exact reasons.

Go to ramp.com slash invest to sign up for free and get a $250 welcome bonus. That's ramp.com slash invest. As an investor, I'm always on the lookout for tools that can truly transform the way that we work as a business. AlphaSense has completely transformed the research process with cutting edge AI technology and a vast collection of top tier reliable business content. Since I started using it, it's been a game changer for my market research. I now rely on AlphaSense daily to uncover insights and make smarter decisions.

With the recent acquisition of Tegas, AlphaSense continues to be a best-in-class research platform delivering even more powerful tools to help users make informed decisions faster. What truly sets AlphaSense apart is its cutting edge AI. Imagine completing your research five to 10 times faster with search that delivers the most relevant results, helping you make high conviction decisions with confidence.

AlphaSense provides access to over 300 million premium documents, including company filings, earnings reports, press releases, and more from public and private companies. You can even upload and manage your own proprietary documents for seamless integration. With over 10,000 premium content sources and top broker research from firms like Goldman Sachs and Morgan Stanley, AlphaSense gives you the tools to make high conviction decisions with confidence.

Here's the best part. Invest like the best listeners can get a free trial now. Just head to alpha-sense.com slash invest and experience firsthand how AlphaSense and Tegas help you make smarter decisions faster. Trust me, once you try it, you'll see why it is an essential tool for market research.

Hello and welcome, everyone. I'm Patrick O'Shaughnessy, and this is Invest Like the Best. This show is an open-ended exploration of markets, ideas, stories, and strategies that will help you better invest both your time and your money.

Invest Like the Best is part of the Colossus family of podcasts, and you can access all our podcasts, including edited transcripts, show notes, and other resources to keep learning at joincolossus.com. Patrick O'Shaughnessy is the CEO of Positive Sum. All opinions expressed by Patrick and podcast guests are solely their own opinions and do not reflect the opinion of Positive Sum.

My guest today is Toby Lutke, the co-founder and CEO of Shopify. I've spoken to Toby on this podcast twice before, first in 2020 and then again in 2022.

Needless to say, a lot has changed both about the world and about Shopify since we last spoke. And we start by talking about the biggest change of all, AI. But I'll remember this conversation for the next set of ideas that we discuss. Toby's very unique take on the very popular concept of founder mode, raising the temperature of an organization and the importance of building on an island that's distinct from the mainstream. Toby embodies the concept of life's work that I believe so much in, and this episode is a shining example of it.

Toby, along with nine other leaders we believe are doing their life's work, will also be featured in the first issue of an upcoming print publication we've been working on.

If you're interested in hearing first when pre-orders go on sale, head to joincolossus.com slash print. For now, please enjoy this great conversation with Toby Lutke. Toby, we're on our every two-year cadence. I'm so excited to see you again. This is going to be so much fun. The perfect amount of time, enough has happened in the world. And in this case, a lot has happened in the world, not just in Shopify's world, but also just in the world of technology writ large.

I don't know how I can possibly not start by asking you about how you are processing what is going on. With a two and a half year pace, we have a great thing is if you ever end up in one of those phases and we decide that, well, nothing much happened or we didn't much like what happened in the last two and a half years, all we need to do is wait for the next one. And clearly it's going to be a new world again at the current rate. I'm not going to say anything, I think, terribly different from what a lot of people say. Very humbled by just getting...

getting to be so lucky to be part of an industry that is constantly has all these phenomenal innovations that allow us to just have to re-derive everything. And like, I think that's, this was like the internet 2000s, Web 2.0, mobile. Over the next 10 years and now we get to reinvent everything on top of AI. They'll touch everything we do, everything we do in technology industry for sure. Much of what happens in the world is going to be augmented. And computers have gotten capabilities that are just unbelievably valuable now.

So this has incredible implications for how to build companies. It has incredible implications for how to build products. And what is software? I don't know. There's clearly more questions right now than answers. I think the map is foggy again, which I think is always much more fun to be a builder in times where the map is foggy and everyone has their own destination. I am trying to make more entrepreneurs. I'm building e-commerce software and I'm responsible for millions of small businesses.

I want them to thrive in this time. So that's a great task to get to think about every day. So the destination is clear, where company is here is totally aligned, everything's pointing in the same direction. And the path is unclear. The path is nebulous. And also the times for creativity, insight, research, maybe even helping to establish some of these patterns. I don't know, super fun.

I'm curious when something new like this comes out, like 01. I saw Noam Brown speak this week, and it's incredible how convincing he is that this new inference time version of scaling laws, another thing that we're going to stack on the pre-training scaling laws, and that if you just give this thing more time to think, it can conduct yet more magic.

When something like this comes out as a technologist that loves this stuff, how do you personally play with it to see what it's capable of? Because I think so many people are reading the headlines, but then don't go play with the thing, which is the only way to figure it out. So I'm curious how you do that.

Usually the second it gets announced, I'm like in some WhatsApp chat with someone saying, hey, enable my account. So that's step number one. Lucky that I've earned a right by sending good feedback to everyone, but I usually get the privilege of a beta flex. So then I don't think I'm going to go to bed that night. Like it's just too interesting. It's just, I literally cannot tell my brain to not. There's nothing that exists.

I almost have to schedule around these things. Maybe as a leader of an organization, I've learned that one of the best things to do, especially with high-promise younger engineers, leaders, is give them a reputation to live up to. And that's maybe slightly pushed from their current set of abilities, and then they raise the vocation.

I have a lot of these scenarios that didn't work with current models. I turn them usually into something called evals. So every time I get a new model, I run it against the evals of stuff that I couldn't get them to do before and see if anything changes. And then I calibrate from there. And then whatever is the new capabilities, I usually follow it to try to figure out where I'm running into issues or just using them. There's a kinesthetic learning aspect, which I think is maybe it's not universal, but it's the way I learn. And just playing with that.

I never find it to be very hard to come up with some random toy that would now be possible. But once you make this assumption that something would be possible now, given this new development, and then you test it,

It doesn't matter if it works. Your world model needs to update. You need to incorporate what is now possible. The most magical moment in the world of technology is when something goes from being impossible to being just very hard. That is the moment where all of the builders need to really converge on everything. That's where you get to very often, like, oh, industry is incredibly path-dependent. So this is where being first matters because you establish the norms.

I saw the talk that you gave, or I guess that the AI gave, and you were the medium through which the talk came into your team. Say a little bit more about that experiment and the implications.

What it made you wonder about, because I clicked on the link, I knew what to expect and what was coming. But had I not, it would have been pretty convincing that that was just the talk that you gave and that it was really tight and crisp and felt very unique, felt very human, felt very specific to you, not some generic chat GPT answer or something. So what was that experiment? What did you do and what were the lessons from it? Let me set the scene a bit because we do summits.

So here I sat down and just said, okay, cool. I have a conundrum. I have time here on my hands, but I set time aside to give a talk, but I want to teach something, but rather play with AI. So how can I combine these things? I came across these papers, which they're all making those cases that giving AIs personas and then having them argue a position and then even putting them into a chat with each other leads to just delightfully different outcomes than if you just prompt because they can question each other.

And given that you've given them different tasks, they can then argue against the current approach, have them collaborate on something. This has been done in a medical space, starting with symptoms, getting towards a diagnosis, not by just prompting it, here's the symptoms, but rather having saying, here's a bunch of specialists and someone at triage.

We give the symptoms to them. They decide, OK, well, you should ask with this kind of doctor. Then this kind of doctor is prompted to like wanting a second opinion potentially and collaborate for multiple turns to get a diagnosis. And they found that this leads to better results. So I said, OK, I can probably do this with a talk. And so honestly, this just worked incredibly well. And these different personas were not different specialists, but rather I had an archivist who had access to all my notes and all my prior talks.

I quickly realized that I needed a linearizer in there, so someone who understands the structure of a talk. And then there's others, the slides need to be designed and so on. So you have a little team. It's actually not terribly different from a team would look like that you put together for giving a talk in real life. So instead of doing this with people, I did it with AI, it's very collaborative. Most important insight was I really needed someone representing me in this, like a project manager who knew what I wanted. I obviously had to prompt it.

but not as a text, but rather as a job for someone to hold the other ones accountable to actually progress towards the goal. So this was actually absolutely a fascinating thing. They put a very good talk together, took 10, 15 minutes, and it was largely a riff on a talk I've given before, but they made their own decisions. They said a motivational quote needed to be inserted and they found one for my notes. They found a quotes file and actually found one which I thought was actually excellently picked. So there was a lot of interesting emergent properties of this.

So the talk I gave was then that, gave this, and then I told everyone that actually this was AI, it wasn't my talk. And then my actual talk was how I built the system that put the talk together. And I think that made it really fun for me. That was made it informative for everyone. And I think it was of relevancy. So learning is there though. At some point, you can't just give a talk. You want to show people how it works.

So I sent a message to someone I collaborate a lot at a company, Daniel Bosch, who during COVID, he built a 3D engine where everyone can hang out and do meetings while doing small activities. So I'm like, hey, can we just load my agents into this thing and just have this play out visually? So we did that and we built a little office 3D environment, which just made it visually much more appealing. But here's the crazy thing. Once we ran it through that, actually the results ended up being vastly better. And this is a big learning.

Because just the fact that the agents were now in a mock embedded 3D space and the combination of who was in the room with who just kept changing. We told them they had to move themselves around. They could make a decision to go to water coolers, these kind of things. And it changed the nature. And they had this happy little accidents of the slide designer talking with a project manager independently while the others were doing something else. Even that contributed to the quality. So all this to say, the

This was really fun, but also the 3D environment thing meant as a toy, but actually I think ended up being rather the discovery here. I can see people actually really doing book complex coordinated work in the future in this way.

Does it feel like in a couple of years we will have this computer mediated visual environment with characters in it and that as we engage with that, that world will eat all entry level white collar work that I will just go into my little custom room with my agents and assign it tasks and have it back in five minutes instead of five days. And if I'm thinking about any low level work like that, it's just gone.

I don't know. I remember the probably wrong story about there's actually many versions of a story of a logo designer just throwing something on a napkin of a lot of value and someone saying, hey, how much money do you want? Here's thousands of dollars. And it's like, you just drew this. No, I started 40 years ago when I honed the skills. I think a lot of things in the world, you can do them very, very quickly now due to your own abilities that you've studied and honed and also to a tool it's becoming very good. And I think people should not underestimate this.

I think AI is a tool. I've named all this my terrarium. There's 3D Engine, where agents run around. Check it out on my Twitter if you're hearing this, just because it's pinned post when we are recording this. I'll keep it so it's easy to find because it just visually shows better. And I've come to think of a terrarium as not a tool, but actually as an instrument. If you send the wrong instruction in, it goes totally off the rails, just like if you pick up a flute and hit the wrong note. But

But I know how to play it. I've learned it and I know how to instruct it and I know which agents to put into it for results. And I think this is a good way of thinking about it too. I think people will have themselves put together a support structure. And I think different people will get very, very different results out of the same systems.

I don't think they will replace the white collar jobs. I think they will definitely replace the crappy parts, the toilsome tasks that we don't want to do anyway. But I think it actually will allow people to just spend a lot more net time at their work using intangibles like judgment, taste. And I think that's actually really good because like creativity, they are technologically argumentable, but truly unique things that people do. And I think this is the way it's going to play out.

Another thing that's happened since we last talked that's gotten the technology world's attention is this notion of founder mode. And I haven't talked to many people about this, mostly because the question seems kind of boring. But I have a specific version of the question for you. I'm curious how your whole business, not just you, but your whole team and your whole business responded to that notion, which hit the Internet like a wildfire.

And by the way, what an incredible front row seat we are getting there for the power of labeling something that was in the air. I don't think it's terribly well-defined, but at least it's a category label. Let's talk about the category of what makes founder-run companies different from other companies. I think this is super valuable because there is an intensity. There cease to be founders or it cease to be involved in a company if they can't master the passion for the product, for the system, for the team, and so on.

But they are also doing a lot of important things. Like there is so much legitimacy to being a founder. Like when everything falls apart, when the world changes with like COVID, everyone had to really re-derive what matters because suddenly every business was just massively affected. If the founders were there for it all along and can just say, okay, it doesn't matter like what our priors say, here's what happens right now because I can discard our priors and can re-derive from principles what we should be doing given the change in the world. And that's valuable.

Then there's a couple of things that are just easier to do. If the legitimacy comes from a founder rather than legitimacy comes from a plan, because this is the core of the difference. If you're a managerial-run company, you're a custodian of a company, you inherited it. Very few leaders of companies, there are some, but very few of them could rebuild the company tomorrow if everything goes away.

What you have to do is you have to create consensus that a certain plan is the right plan. And then that plan is actually the source of legitimacy. That plan is being implemented in an entire company. You can campaign for an exclusion of a plan. You can talk to the board. You can't use your keynotes to talk about AI terms. You actually have to use your keynote building up the legitimacy for the next set of plans. This is why you see goalposts moving. It's much easier to just change numbers in a plan that already has legitimacy than it is to change the plan entirely.

So it's much harder to react. And so everyone, I think, generally gets this.

The plan is not sentient. The plan can't just say, hey, this company is spending too much time in meetings. The plan can't go and subtract all the meetings every once in a while because it's neutral to meetings. It doesn't reason. But the management and even the CEO might not have the legitimacy to do so because I tell you, as someone who leads meetings every once in a while, people are pissed. It's a Toby thing. So that's a good change management. It's nice to not have to deal with everyone being pissed. Everyone's just like, ah, Toby's just eccentric about meetings.

meeting a lot goes from people average five and a half hours to three and a half hours for the next year and the company looks better because the people who are actually writing code are actually getting work done. Any of these meetings are not important. So just as an example. So what's this founder mode thing? I think it's a label on top of something that people feel, especially people who have been working in a bunch of different companies, the relentlessness. But here's the way I would put it. I think what happens in most companies is they converge on, let's go with temperature. I think companies run at room temperature because

because that's the entropic equilibrium of temperature. That's not a problem if everyone else does it too. So in room temperature, everything is just so. But then there's people who inject heat into businesses. Brilliant product managers, brilliant people in any job do this. They're dissatisfied with status quo. They pump energy. The best leaders are literally exothermic. They are just like this complete wellspring of heat into the system. Every atom jiggles faster around them.

There's no chance for stasis. Nothing is frozen over time. However, people mean well, but they fight it. They do because everyone's allowed to try to conserve energy. Everyone's allowed to be an intelligent actor in the local incentive system. And there's a lot of heat in a project. Everyone go faster, do this better. And the best way out of it is to just politic that person away when the people will do it. At least someone will do in a large team. And if that works, then those people are gone. And then room temperature prevails.

Founders are not room temperature people by definition. The main act of starting a company is you're trying to inject heat into planet Earth. You're trying to do something like fighting against entropy. You're trying to make something better or something different. So I think founders are recruited out of a set of exothermic people to begin with. Here's the thing that founders end up doing. Not just are they injecting heat from the top into the whole thing, they also protect the others.

When things are going well, they make the politicking against them not work.

So if you do this well, I have a company where many of my reports, I think almost all of them, are the founders of the company. I literally collect them. During COVID, I had a lot of executive turnover. I went to the founder's Slack channel, which is people who over the years brought into Shopify for talent acquisitions and so on, and said, hey, we're in a crazy time and you all have been responsible for people's livelihoods before, and therefore you know what that's like and who can do what, let's go.

Let's dance. And I think this is a real thing. Companies have a clock speed, companies have a temperature, and it's just powerful to preserve it. And then it's powerful to be able to use all that energy and direct it into a mission where you can re-derive from first principles what the next best step is on this mission.

Especially during times where the fog of war has settled and the map is not quite clear, the mission gives you a direction and you need to figure out the best path through it. In a game of perfect information, in a time of perfect information, maybe this is sort of the 2017, 18, 19 era, this is actually not ideal for companies. I think you actually want potentially lack of heat in the system, lack of jiggly atoms, pick your metaphor here.

It's not fatal because actually the plans will work because the plans are not going to be invalidated by a changing environment. And therefore, it's way simpler to lead a company this way. All these stories about...

hire good people and get out of their way actually work, but they work accidentally because everyone's just following the plan. The plan is the thing that's telling people what to do, not the CEO. Especially interested because it resonates so much with my own experience. The act of protecting the exothermic hot people seems like where the rubber really meets the road here in terms of what your job is. Obviously, you got to get them in the seats in the first place. You're

We can talk about that too, but you can probably identify those people, even if it's just those that have been founders, like you said. Can you tactically tell me how you have protected those people and the steps that you've taken to not let the political machine oust them from the system and reject the organ? Look, the difference between the best companies in the world and the worst company of the world is probably that humans are political creatures and best companies in the world, people spend up to 20% of their time on politics and then worst companies is up to 60%.

There's a very small margin in which you can operate. You can harness politics too. It can be good, it can be bad. It's not even all bad, but it usually means that people have a different hierarchy of goals than the company itself. So the top goal is not the mission of a company. It's your own career or something like this. How do you protect people? Well, you just make the exothermic thing work. That's what you promote. Again, I pick people from individual crafter roles and put them in charge of large hundreds of people departments after they ran a code red.

For me, that just really worked. And they have the abilities to do this. But they skipped many, many rungs that other people think they should have really passed through on the ladder of a career. And I just don't agree with that. And the people that really, really need this to be all just so and everything being seniority based, I suppose. I think there's a lot of great companies out there that are structured so. And I don't have a product that they want to buy in terms of a way to think of a career. I think of merit and

I'm actually not arguing in favor here of the brilliant jerk and keeping them and building because I think that's actually can end up being toxic. I think also like the ego to merit ratio has to be enormous to just bother with people. I think life's too short to surround yourself with mean energy vampires. Like that's not a good idea. So high merit to ego ratio is important.

Some people are incredibly good at what they do, and they are allowed to be self-confident about them. I know some people misconstrue this as arrogance, but it's arrogance if you're not good at a thing. If you're actually good at a thing, you're just communicating it. And people might not like that, but that's just aesthetics at this point. It's a complicated balance, and you have to figure it out. The second way is malice in behavior. I'm out.

So there's a code of conduct, minimum requirements. I think that's a valuable thing to just break up in this context. So how do you protect them? Well, you do. You say, hey, we have a company built for these type of people and we celebrate going fast and what this place is all about. Can you have a team full of hot people, of exotropic people, or do you need some room temperature people no matter what?

I don't think it's possible to try that at any kind of scale just because it's just not that common. But I think, quote, absolutely. I think that's the story of the greatest agents of change. I'm sure the founding fathers had no passive players in them of the United States. This is why they're so fascinating. They're all exothermic in different ways. They just happen to agree on the broader picture and come together when it mattered.

It seems as though the most valuable potential actions of founders in founder-led companies, especially at scale, is aggressive redirection that if it was a manager would require crazy amounts of consensus and therefore often never get done. The last time we talked, you had tons of emphasis on the Shopify fulfillment network, and you've done one of these where you've really completely changed everything.

how you're thinking about it, what you're doing as a company. I don't think a manager, a professional CEO could have done that. Can you just walk us through that as an example of this power?

I agree with that. I think, again, that gets you back to what is valuable about having founders around because you can re-derive against new information. You can not just, what does this mean for the next step, but actually replay the entire history of the company, take this new piece of information, insert it at the beginning, run the counterfactual. Where would you end up? If that's different, now you know something that's true. Now the only question is, what do you do about it?

Do you have the courage to act on it? Is it the best set of trade-offs in the current situation to go make the change? But where does this piece of information come from? It usually comes from people who are there, highly alive player in the context and having a conversation around like, I don't know if you know this, but this thing is true. Or just what's better given all the things. In this case, the thing that changed, there's a couple of pieces of information. I'm not going to get into all the details, but one large thing was actually just also AI. Look, attention is important.

I think we talked about video game analogies. One thing you learn when you play a lot of good, something like StarCraft, your most limited resource is not actually the minerals and the vast team gas, it's actually attention. You win the game by zapping attention of other person. So attention management is one of the most important skills. We were in fairly stable times.

things weren't changing so much. And building fulfillment network was super valuable, but it's a big pivot for us with completely different skill sets. And my attention was extremely focused on building the skills that would make me a good leader of all so bad. The moment I came, it was very clear that this is all software has to be reinvented. Well, first of all, that's already in my wheelhouse. So the

So that's useful. And then second, non-fulfillment stuff matters. If Shopify ends up being the perfect solution to the types of problems that we had in the days before AI, it's like, that's not good business. So because of that, I'm like, cool, let's go deeper into the field that we already committed ourselves to and not go into the adjacency. And once that decision is made, my joking label for this is that I believe in the saturated sponge school of change management, which is not a school I made up. It's

It just comes from the analogy that if you put a sponge into a bucket of water, afterwards it's so full it cannot absorb more water. And therefore, it doesn't matter how much you submerge it after the point. So I think change management is best done as just saturate the sponge, load up all change in one go. Because people will mind, but they can't mind more than the saturation point. So all other change up past this point is free. So might as well do it.

I think these long drawn out go and just do a whole lot of stuff over a long period of time are incredibly incompatible with human psyche. It might seem more convenient in implementation, but I think it's bad for us.

I'm curious how the art and the science blend together in this. As the founder, obviously, you have directed the art of the business, the original product, how the world should be that it's not. You keep doing that on an ongoing basis. But now you've got a huge business. It seems like you've got a very talented executive team. And I'm maybe even especially interested in your engagements with your CFO and thinking about

I know all the things I know and the directions this could go and what I think we should consider. How do you then sit down and work with someone that's maybe more on the science, on the number side of the business to learn and make these decisions together?

My CFO, Jeff, is someone I've worked with intensely nine years ago when we took a company public. He was on the among-standee side, and now we get to work together every day, which is awesome. And look, the numbers, the science, as you call it, they are the facts. They are the most bold-faced points on the graph paper of trend lines, because most of you know for sure everything else is as good as the model and heuristics. So I think we talk a lot about these things.

Obviously, we talked a lot around the times that we were making changes to the fulfillment network because there's such consideration about how it's perceived. But I just think at the end of the day, short-term pain, if you know it's better, let's go. I don't over-index on these things. I just want to understand how people think about it so I can address their concerns. And then I personally admit this. I don't always have a full recall of all the numbers, even the top line numbers, but I know the trend lines.

So sometimes when someone asks me a very specific question about finances, I will go back to a checkpoint where I looked at and will run it forward and I'll be right about it. But I'm not in the factual numbers all day long because a lot of it is just I get better predictive information from a lot of machine learning model forecast systems and so on. That's sort of an interesting thing. We have an executive team. We come together every Monday and this is important.

Data and finance at Shopify work super closely together. And that was really important to me because I think many CEOs live in a split brain problem world where data is just forecasting and finance is just gap. And I think gap accounting is basically a dashboard for Wall Street on top of facts of a business. And I want those things to actually be extremely connected because otherwise things go very poorly.

And then who do you spend more time with ends up being how you think about the business. And if those things are not congruent, then bad things happen. So I think that's something that's gone really, really well in this company.

It's fun to think about Shopify, the business as this thing that came through you and your early team and has grown up on its own and exists independently, but you still steer and control. It's like the rider and the elephant or something. I'm curious if you think back two and a half years to our last conversation, Fulfillment Network, et cetera, and then forward to today with AI and everything going on and the way the business looks, how you would describe the difference in fundamental nature of the elephant? Right?

of the thing itself, of what it does, of what it's meant to do, of its key characteristics, and how, if at all, that has changed in this two and a half year period.

I think it's a much more self-confident company. It's a much more tension company. And in fact, it's a lot more similar to five years ago in a lot of ways. You might remember, we went public very early, a billion and a half valuation in 2015. I think a billion and seven first trade. So that's a very far distance here. And even then, I think we basically had all the money that we raised from financing around still in a bank at that time. Shopify is just a company that makes every dollar count.

It's like being operationally efficient, being efficient with money, being efficient with capital. It's like out of a fitness function of a company, things are tested. Performance waste is like physically painful in a very real way. I think the zero interest rates periods, really, they made all these characteristics of the business basically a liability because you cannot compete with everyone free spending on free money. If you're trying to make every dollar count, you can be swamped out of the market.

So I had to lull up not caring about this stuff for a while. And I think that just was not great. If I had a time machine, I would just, hey, let's keep tensing through the time period. But like, I don't think this was actually a valid play for this period. And this is how everyone got swapped up in this time. Also, it just turned out I was too exposed. Honestly, like I've been thinking a lot about islands, genetic islands, like

Darwin and his voyage of a beagle finding finches which are different and saying, "Hey, those are the same finches," and therefore coming up with a theory of evolution. Genetic islands are really important. I'm just making the leap here from genetics to memetics. Obviously, I'm not the first person to do this. Memetics works exactly the same. The exchange of idea just happens faster than the exchange of DNA. There are not many islands in the world anymore. Mainland is really big. The internet has connected us all. But innovation comes from islands.

Difference comes from islands. Literally visiting Japan is different than other places because it's an island. It has enlightened barriers such as water and language. And therefore, they can maintain difference in a way that is very hard to do if you're part of a mainstream. Mainstream thinking is incredibly fashion-oriented.

Sometimes it's fashion of this sort, sometimes fashion of this sort. And I think that this aesthetics even ended up impacting company building. And I think my takeaway is Shopify did not have language or water as enough to be able to make decisions that are significantly different from the ideas that are just floating around in the mainstream of this time. I think focus depends on more islands. I really do. Shopify is based on a programming language called Ruby, which hails from...

And I recently got to spend a little bit of time with the creator of Ruby for the first time. And it is remarkable because that language is different from all other languages. Optimized for totally different things. Ruby optimizes for programmer happiness. It implements the great pains to the implementers of a language, something called poetry mode, which allows you to just omit brackets here and semicolons. You just don't need them. The system does stuff just to make things.

engineers have. That's an aesthetic. That's a decision that Mr. Matsumoto made 31 years ago when he started this language and has been spending his entire life perfecting this language in his way. It's his. He makes all the decisions. He has

Lots of people were helping him, but like all good open source software, it's an island that creates gifts for everyone. It is not part of the mainstream. He's not compelling anyone to use it, but people who are using it love it. And then there's like Ruby and Rails, which use that gift to make a web framework, which I've used and I've worked on a bit. That is also just a gift. And David Hanneman Hansen makes all the decisions.

You would make another version of it because it's open source, but that island is yielding gifts. And I don't know, I think the memetic mainland has gotten so powerful. I don't think we have a lot of islands anymore. What new music has come out in the 2000s? There's all this music. EDM was over here and Berlin, Grave Caves, and then you had hip hop from abroad. Obviously, these things were around longer, but they were like fully formed products. And then what makes them as good as is distribution of ideas.

or even inside of a company, instead of becoming this monolithic mainstream thing, which is very good at distributing ideas, to be fair. How do you become more like an acapella of islands that all come up with products of merit and then trade ideas and give gifts to each other? This seems like a double-edged sword with AI, that in one sense, these large models have a homogenizing force to them. And anyone that's interacted with enough of the outputs knows it has this feeling to it, and you can prompt it differently in whatever. This is actually what I mean. I can

Academia was an island here in this metaphor and through the related peninsula, I suppose, of open AI ended up creating fully formed gifts that then were exported to everyone. This made it to mainstream was distributed fast, incredibly fast, fast growing product of all times. Again, because that's what the mainstream, let's call it the main continent does.

That's cool. But luckily, we also have Anthropic and their models just have a different tone. That's good. And then the incredible, and we should all be so thankful, Meta with its open source models and giving us incredible gifts in conformity to open source ideals here of models that we can train and can do different things. And now all these little islands can form that say, hey, over here, this is our job. And this is our leadership structure. Like this is who makes decisions. And we're doing something for a completely different purpose.

Then again, if you come find something that is of value, other people will do it. If not, no one cares. And that's fine too. The participants of an island are getting exactly what we want. They don't need everyone else to also use it. I think that's important. Can I ask about maybe two dimensions of this and get your reaction to it? So one is I'm obsessed with this concept of life's work. And increasingly, I have realized that the thing that attracts me to people is those who are doing what you say, like building their own island. And everyone's got a unique experience.

Well, maybe we could talk about that, like how different people's experiences are could be a problem here. But people do have unique experiences. And if they find a way to self-express into whatever it is that they do, a business, art, whatever, that's where real magic happens. And that a lot of that comes from separation. I always love that Joseph Campbell idea who studied the history of mythology and that he credits all of his findings with going into the woods for five years and reading books by himself and talking to nobody and just being separated. Being an island, yes.

Yeah, being on an island. So I'm curious, I want to talk about how we can promote more islands in general, all the ideas you have for doing so. My own idea would be this concept of life's work, plus maybe separation, which are different ideas, but really important, can be ways of doing this. Because I think a lot of people listening will find the idea appealing and say, yeah, I want to create one of those, but not know how to do so. So I'm curious for you to riff on that.

The two ones I now work extremely well is in the programming world, we have open source, which is always island producing. And I think actually the purest implementation of a market, open source is totally misunderstood by the way. People always think about this as the commons produces something that's better. Literally it's the opposite. Everyone can fork the project if they have better ideas and send gifts from there. And if those gifts end up being better than the previous version, then it's a marketplace that doesn't run for money. It just runs purely for legitimacy and merit.

And every good open source project is led by a very small group of people surrounded by a clear and highly legitimate leader. So it teaches us a lot about human organizational structures that work.

And then companies. Look, Shopify is an island from the beginning. And it's my life work. So you might be talking about my book here. But I was part of a real main continent, real world, let's call it. I looked at the software that was available to me and I just didn't want to use it. I built my own because I didn't think people were doing this right. I built it for myself. I built the software I was looking for that I couldn't find. I figured, man, I'll just build that because I think other people will find that useful. And then I invited people over. The amount of times I heard...

hey, this shopper thing, I like that you're doing something like this, but this will never work in the real world. It's basically every day. Someone like, this will never work within the real world for real retailers. This will never work in the real world for real businesses. This will never work with customers or enterprises or any of these. I've heard it all. And here's my response every single time. It's like, yeah, that's fine. I'm not building for a real world. I'm building for my own island. And I'll invite you over. If you think what I have is better, use it.

So I think that's important. This separation is important. And again, I'm going to give myself a B minus to C plus here, because again, when we went into, I think I could have survived 10 years of a boom cycle intact, but year 11, 12, 13 of zero interest rate phenomena appended to a 10-year bust cycle, I got washed. Someone built a bridge to mainland and I didn't notice. I got infected by some ideas. I would not cross with myself about this.

So my defenses weren't good enough to be permanently outside of this, but I still have some semblance of success here. And I think I've reestablished sovereignty. So again, we are not building for real good. We are building based on insights into the problems from first principles. And then if people like it, they use it. And then those trend lines we talk about, Jeff gives me the update points. I draw the trend lines between the points. And if those go in the right direction, that's really good.

But it's a consequence of making sound decisions for particular problems that we as a company, me as a patient zero, and then we as a company have just fallen in love with and give a shit about. Because at the end of the day, I know this is all very abstract. I'm saying it's all terribly confusing. So just say something a lot more actionable because the place you arrive at at the end of all this philosophizing is

Products are good when the people who make them give a shit. And because what they do is they fill up the products of giving a shit and everyone feels it on the other side. And that's just wonderfully simplifying realization at the end of the day. And I'm sure there's hundreds of paths into navigating a position where you can create products that give a shit. And the beautiful thing about products is they are one of these things which causes community. And community is one of those things which bootstraps

separation to some degree. And that is the substrate from which true lasting differentiation can be done. And again, we're 20 now in Shopify. And so it's gratifying to have done so. I aspire to what Matt's did with 31 years of a difference in the Ruby world. And I hope to spend my entire career soon. So there's this island, but let's just presume its existence. And then there's constantly attempts from the mainland to build bridges to it and infect it.

One of the things that's so fascinating to me as you're a big public company is that Wall Street always has lots of ideas for what you should do. And one of those ideas, if you go read the expert transcript libraries, which give you a nice window into what people are asking about, is why does Shopify not go monetize its end customer more? Why has it remained so focused on the merchant?

It has this massive network. Why is there not a Shopify Prime equivalent? Why aren't you figuring out some way to take this thing that you uniquely have and drive more monetization from the consumer side rather than the merchant side? This is like Island and Bridge to me and a tangible example to make it real for people. Walk us through that specific pressure from Wall Street.

I don't disagree with all those things. I just think people always made the value of an idea. An idea is worth one good bottle of scotch. Everything else is an implementation. And here's the thing, the implementation matters. For instance, it would be trivial to do this, but there's an understudied, underacknowledged thing that happens where your customers by a company stop being treated as a customer and are now treated as a crop to be harvested. That only works for a moment because this is not a sustainable long-term relationship you can have.

And I see our customers and people coming to our customers as all partners in a big project. That doesn't mean I don't monetize. Clearly Shopify makes billions of dollars. But we, again, earn money where we are of value and of merit. Very often Shopify is the cheapest software people have in their entire stack. People pay 10 times more for their CRM.

for the software that literally runs the entire business, which I think is hilarious, and I'm sure will change at some point. So again, because in some of these cases, this might sound too much of a pointed particular thing, I'm just making a general statement. So you can talk about metrics observability tools as well as CRM. And obviously, I've experienced this myself in a business at some point. I thought we were partners building something, you build awesome software, that means I don't have to put engineers on observability, and I get to focus on commerce, and you do this, and I pay you

sensible amount of money and that works for everyone. And suddenly I become crop. And it's very obvious when this happens. I'm just, holy crap, man, I know I have to replace all this shit. Anyway, so this is an experience that happens a lot. And let's just go and build long-term value stuff. And we are smart people. We find if you figure out and can share the attention, it's a priority enough for doing so. We'll figure out how to build something that people just love spending money on because it's

because something new just become possible. And that's the right way to think about monetization. And then revenue growth over a long period of time takes care of itself.

Can you talk about the impact that being in founder mode, founder-led, protecting the soul or the flame of a company makes marketing easier than trying to market a plan that's been built through consensus? I've learned a lot from Lulu Masurvi on this topic, who I know you know. And I'm sort of obsessed with people not realizing that if you just do it how you want and build on an island, marketing will be way easier. Can you explain, well, first of you agree, and then second, how?

I think this goes straight into the island versus the main continent, mainstream conversation. Look, the mainstream is much bigger. No doubt, if you want to grow as fast as possible, you have to convince people to switch mainstream narrative in your direction. If you are off an island and have a different value system, your direct message will not do the job here because that's not how you speak to mainstream. You have to love mainstream language. By the way, that's changing all the time. So maybe you accidentally speak the right language in some instance, but that changes tomorrow.

So what does this mean? You hire a marketing agency which speaks mainstream and speak mainstream to people and hopefully do it slightly more tweaked in such a way that mainstream pays attention. Again, that's not what I'm talking about, though. I think what we tell our customers, our merchants, is why does your thing exist? What happened to make you think that this needed to be done? What's your source of energy? What are you doing? You are exothermic, but again, Newtonian laws say energy has to come from somewhere. What are you converting into this exothermic energy?

Is it just some dogged desire to have a product exist? Is it anger at the status quo? Is it a counter? Have you been slighted by people and you want to prove them wrong? It doesn't actually matter. Human emotions are the only net adders of energy into these systems. Which ones have you tapped and contained and put a steam turbine on top to cause building to happen now? And then explain the journey and say this thing. That's a story for a founder, but that's speaking to your island, not to the mainstream.

So I've learned from Lulu and through Charles Nero, but honestly, I didn't take my own advice there and tried to always like, hey, let's frame Shopify and whatever. We placed our website, recently just put a video on top there, basically asked our internal videographer video team, people who've been with us for a decade, to say, hey, this story we always talk about, about why Shopify actually exists. Can you just pull this out of this company? Do your magic and just do that. And let's put it on Shopify.com and just explain this exactly the same way we tell everyone.

to explain why their products needed to exist. It's a little weird that that took us long to do, and we did. I think that's a better way to talk about it. Again, because we do not want to convince the mainstream that everyone should be on Shopify. We just want to tell people that life on our island here is really good. Things in the fitness functions are that we celebrate people reaching for independence, doing awesome stuff that's hard, that's overcoming adversaries.

building products, adding value, creating gifts that are sent. There's an exchange of product for value and a sense of money. And by the way, we have pre-tooled it all. We have a lot of infrastructure of our own on this particular island. Come join us.

I don't want to convince everyone to use Shopify. In fact, I literally want to do the opposite. I actually don't want people to do it, to use Shopify if it's not the right thing for them. But the people who want to pursue entrepreneurship will figure this out. I think people do figure this out about themselves. And for them, there needs to be a place like Island of Misfit Toys or other ways. I'm passionate about this. I would love to actually invite more people to think about this. To be different, you have to be different. It's an axiomatically correct statement.

Archimedes would be proud of me right now. But difference doesn't mean better. Difference just means different. But that's the cool thing about islands. You can have lots. And if something isn't working, you just go to a different one.

As long as there's no coercion there, you end up in a much more rich marketplace of ideas where lots of people can just bring ideas into the manifold. People can judge them. People can join them. People can adopt them and maybe pull them into the mainstream. But as long as something in mainstream is exported from an island that maintains its membrane separation, you will gain more and more and more innovation being sent. If it's integrated fully, the bridges, the highways,

The differentiation is broken. And I think what you end up with is something that can only be changed marginally, maybe slightly incrementally, but not in the richness that people have hoped for.

Can you talk about the, I'll use a big word, the glory of entrepreneurship? Because ultimately you are serving people who themselves are entrepreneurs and building something hopefully unique, hopefully their own little mini tiny island with a couple of hots on it to start and can get very big sometimes. Talk about that glory because that does seem to be the through line for why you give a shit. And that's like a renewable resource. You don't run out of that energy, it seems, at least in the time I've known you. So talk about that.

Exactly. This is so important. I love the term glory for this because it is a glorious pursuit. It requires courage. It's a journey. To me, I scroll through everything Netflix has to offer and sometimes I watch something. I don't think the stories are as good as the ones that I hear when I talk to my own customers.

This is the craziest and most awesome thing about Shopify and something I do bring up more frequently now, even when people are considering joining the company. You know how awesome it is to build software for awesome people? If all your customers are inspiring you, again, that's exothermic as well. And frankly, they are exothermic people. They are people who reach for a difference. They are trying to make something. I'm so unreasonably excited. I can scroll through my shop apps order list. And last night I found a Shopify store that's 100% specialized on making products

pirate ship treasure chest themed poker sets. I love that this exists because that's awesome. I'm taking family to the Caribbean. Clearly, we never knew it, but every poker game I've ever had with my kids and my family has been impoverished by it not being pirate themed because that's totally awesome. Someone realized that the world was not complete.

At least for me, maybe I'm the only customer here. Maybe I've been micro-targeted without this existing, but someone saw it through and made it exist and went through all the trials and tribulations of creating these products. And I love it. So I think that's just glorious. And it's glorious that people can do it. I hope they're doing excellent. And I hope they create lots of employment and lots of secondary employment in the supply chain. It's just net value creation. This is how value comes into the planet. Again, it never gets lessened.

There's no downer or no downsides. I'm actually almost humbled by realizing just how lot-bearing this is for all of economics and all of our standards of lives. Every time I think I've wrapped my mind around it, I realize it's actually more important than that and somewhat more fragile as well at times.

So I hope that one of the things that I've discovered with this terrarium, which I use as an inspiration for like make my own decisions about how products should evolve in the future, will be that way more people can go and put a team together of experts that they need, some human, some not, to help them along, inspire them, be thought partners, be nonjudgmental.

sparring partners if that's what people need or just deeply competent logo designers if that's what you reach for or just explain how to actually set up a domain and how to incorporate business wherever you are. People will get so much more support on their journeys that just more people can reach for it because

There's a lot of white space. There's a lot of pirate-themed poker sets to be done in basically every other intersection of multiple awesome things that can be put together. And I think my most hopeful belief here about this AI age is that a lot of the products, lots of the things that don't exist are

A lot of value that wasn't added to a global GDP happens because friction, confusion, lack of access to information and too much difference in naysayers who say reaching for independence is not the thing to do with a career. And I think that'll change over the next years. And it's already changed in the last 10 years. I think that's awesome. Can you re-articulate your product philosophy?

At Quorum, philosophy is humans are absolutely awesome and computers are absolutely incredible networks we've built and the international network we've built. The infrastructure that humans have built is awesome. And I want to put this stuff together in such a way that you give those awesome powers to as many people as possible and then let them have a go at it. At a different Zoom level of a fractal, we follow a philosophy of make the important stuff super easy and make everything else possible. That's why you can get a store in a minute that could power your business for life. You

You can dive into the miniature of a template language or compile a WebAssembly extension for Shopify to customize it just so you can do both things. So that's important. But all of it, I think from day one, the killer feature of Shopify has been that it's retained complexity into super approachable user interface that can be used by people with minimal training. And hopefully, in most cases, the significant power of a system only reveals itself to people in the moment that they actually have such problems.

And it's very hard to build software that scales as far as we do. And we're not talking about technical scaling. That's also hard, but super fun. And I'm never worried about the stuff that's hard and fun.

The stuff that's hard and not seen as significant fun is much harder to do because that's where the exothermic energy needs to come in. You know what unfun causes? Freezing. At some point you're at zero and then nothing ever happens anymore. So one thing that's unfun sometimes is this taming of complexity, which is building software that actually works for people that try to retrain independence and glory in their lunch break.

and still the same software works for multi-billion dollar businesses. Everyone's like, man, that is clearly the people up here in the market make all the money. They have money, they have sophistication, they have dedicated staff, they have training programs. Why don't we just build for them instead of everyone? That would be much easier to do. And yeah, the answer to that is 100%. And by the way, it would be very hard for a not founder mode company, like a military run company,

Someone to create a plan that the consensus will sign off on that says, hey, let's build software for people in their lunch break. That can also power billion dollar businesses. That's not a sensible thing to do, yet it's exactly the right thing to do. Because what is missing in this equation is, well, if you do that, you can help people become the billion dollar businesses fast. The billion dollar businesses on Shopify started on Shopify now. They've been around long enough to see the other side of something that was only a hope at some point.

So we can now take a very active role. If a business could be many million dollar businesses, I think the software can just push people to get there faster than they thought was possible. And we can reduce the complexity of the entire journey. And frankly, we are going to reduce the complexity of that journey absolutely tremendously for all the AI stuff that's coming. So I think we can remove so much friction that at some point business can just grow very, very, very fast. I think that's awesome.

So my life philosophy is like build the best possible software that anyone could possibly go find for who wants to do this thing. That is what our mission is about. And then compromise to the point of currently technological possible and then build that. But know what insanely great looks like. And then compare the changing landscape of technology capabilities to what we actually really, really want.

The ideal Shopify version of product is you have a quick conversation or email tell us about products that you have and want to sell. And then there's a beautiful website doing it perfectly for everyone. Maybe it's customized to every individual person visiting it or is talking through AI as an advisor to other AIs to tell them why you also need pirate themed poker sets. And then if someone orders it, we teleport it to people. That's the ideal state. Like this is everything working perfectly. But let's compromise because we haven't beaten physics on the teleportation.

But boy, have you gotten a lot of AI now. So this is what keeps it fun. And my product principles is pull as much value out of what the smartest people in the world of computer science and logistics and computer networks have figured out to be helpful on our entrepreneurial journey so more millions of people can do it.

I love the product philosophy of connecting people that want to use the tools to the most magical elements of what technology enables. I'm curious in the face of AI, how you think the way we interact with computers in the first place might radically change? Someone talked to me recently about a world post applications. What if the way we interface with computers is incredibly fluid and not in separate panes, not in separate windows or applications?

and that it just starts to feel more like a conversation. I'm curious if you think that will come to merchants, to the way that we shop, and bigger picture, just the way we interact with machines, and how you deal with that. I 100% agree that this is what's in store. Now, you're not moving from one thing to another. For instance, all of us technologists, especially programmers, have lived through many such changes.

We are all using a paradigm of hopefully nice high DPI screen. It has Windows on it. Most computer programmers will have somewhere terminal window and that reaches into the command line. That's for Unix. Unix is still alive, still sits below all this stuff. And the people who know how to wield it like an intricate, incredibly powerful instrument reach through the windows back into the Unix system. So it's there.

I think a lot of what we're doing on computers is going to take a similar form. It's just like programmers reach to a Unix command line. Our generations of computer users will appreciate Windows and applications. Many of these applications might exist only on our computer because it was bespoke created by an AI for us, for our purposes. But I think it's an aesthetic that works. We also then leave our desk and our meteorite glasses are going to superimpose whatever we want onto the world and

And it just seems extremely natural. And we will have conversations. I have conversations with my MetaRay bands right now. So I think you can already live in basically this future if you try it. I think about these things as vacationing in the future, like set myself, okay, for this week, I'm only going to work in this way. But I now is more representative of this Gipsy and idea of the future is already here. It's just not evenly distributed yet. I'm a tool builder. So therefore, I need to live in everyone else's relative future, the very latest version of it.

So you can do all these things already. And I think it's going to be a huge and. You are going to do all these things. And you're going to have a conversation about your business on the phone with a person who's then going to do things. And then you're going to have about the design for the next campaign that you just planned with AI that's going to send proofs to a window on your computer by the time you're at home, maybe through an email, maybe through whatever else. Maybe it's like something you're going to review on your phone, which is also still going to be around. It's going to be all of it. So this is the direction it's going to go.

Or shopping depends entirely on what you're doing. In the same way how I can right now do e-commerce or go to New York Fifth Avenue and I have these possibilities and there's going to be 15 other ways you're going to discover that are also valuable. In

In the future, that's not very far away, I will have a favorite AI that either through just being connected to my systems knows that we are going to the Caribbean, knows that the family is coming, again, because it's presumably my calendars, knows that I play poker with my kids, knows that I play Sea of Thieves with my kids. And therefore, as we are getting closer to the Caribbean trip, it says, any chance this is the thing you would be excited about? Because it knows me well. Now, it does this because it has a lot of data on me and I trust it. I've

I think AIs working together, doing different jobs, representing both sides is really what I think people mean with energetic flows, or at least I think the user observable consequence of all the talk about energetic future. And I'm here for it. This sounds so much fun. In a counterfactual world here, I'm listening to your podcast and hear someone talk about what I just said. I'd be so unbelievably, unreasonably excited to listen to your podcast.

to live in a world where there's a before and after on that, because that sounds fucking valuable. And I want that to be true. It sounds cool because I love computers and I love discovery and all these things. So I want to be in a world where this all works and the builders of our times are building to a future. The fact that I get to actually play an active role in this is so crazy unlikely to myself even, and I'm so grateful for. It just feels trivial to have a good time right now in this industry. So let's go.

My friend, Zach Frankel, has this interesting idea that AGI is already here. These things are already smarter and way more capable than the vast majority of humans out there. And that the problem is just figuring out what to do with it and how to connect it to the real world. And I'm really curious because you run a company that is so incredibly last mile connected.

That's literally what the business does for all these different merchants and your customers. And whether or not you feel as though it just takes time to digest this stuff, we can't speed up the rate at which we connect this new thing to just the real world itself.

the world can only accommodate so much change in a given year or something, and that we're already maybe in catch-up mode, and that we do already have AGI, and it's just the process of deploying it now. Does that resonate with you? What do you think about that? I think the AGI, I think you can get in the definition. I just don't think I have anything to contribute to the definition. I'm much more fatalistic about this. What's the utility that can be pulled out of it? Also, what is the total potential utility they can pull out of GPT-4, and what's the percentage that we have pulled out so far? And I think I would put that at single-digit.

So let's just say like GPT-4 is out, it costs per API call exactly what it costs right now. We'll never get a new model again, but we have infinite capacity to call it and it doesn't deteriorate in speed or quality. If this would be scenario, basically freeze frame exactly what we've had before O1 was released, let's say, then I think there's probably $10 trillion of value for industry that can be pulled out of just that for the next 10 years.

And that's not the scenario we actually are in. It's literally not the scenario we are in today because we have our own models and new models we drop on Hang Face every day. We have for first time models that are better than scaling laws, returns for training, trained by a startup. We are seeing incredible progress. So is AGI here? No idea. Does it matter? I think this is an extremely important question for a small group of people and everyone else gets to it. But basically magic.

I'd love to close and collect all of these ideas together into a final question. What's really standing out in my memory from the conversation is these exothermic people, the role of heat inside of organizations, the role of hopefully getting those people on an island somewhere to build something with variance, with uniqueness to it. These are beautiful ideas. Shopify is one of them. Maybe you had a little detour there for a couple of years in years 11 and 12 of Zerp, but you're back. So you've got one and it's big. And

and there's a lot of people there and it serves a lot of people. How do you build trust inside an organization at this scale? Because for this to keep going, that seems like the core element that you need. And I'm curious how you do it. Work with people deserving of trust and will repeatedly show that they come through whenever you trust them. Explain why this is non-negotiable and also explain that this is not everything has to succeed.

Actually, people build more trust with me quicker than they disagree with something and the right, or at least right based on everything they know and then articulate that. And then we can have a conversation about why this still isn't correct and what perspective they might not have missed. I love these conversations. I think you have to have a culture, but the important thing is what is your definition of right actually? Because I've learned that a lot of people's definition of right is actually the opinion that most people hold.

It's the right opinion. You're trying to converge on the consensus decision. I think it's deadly. I think that's a fitness function of the mainstream again. It can't be a fitness function of your island. Your fitness function of your island, I think if you're trying to be different and you're trying to be on the good side of different, aka better than mainstream, you need to make actual true correctness the thing. So I think when people are correct in their first principle sense and are willing to say so, even if it's inconvenient, what happens next is the thing that will shape your culture. If that is appreciated,

You're going to go places. If that isn't, then you're dealing with either wrong assumptions about the culture by some people or politicking. So that is very important that that doesn't work. So then it becomes a way for people to help themselves to having opinions that's distributed because again, is it true? Is the thing that's important? And that allows people to step out of the org chart and just put up their hands saying, I've got something here and know that this is going to be going well for them.

even if it contradicts potentially what the project has been doing before. It's also lots and lots of interactions. I do have an entire thing on the trust battery, which is micro-interactions too. I think the big one is what is the opinion about correctness, about truth, and about right that is at the core of a particular community that you are a part of? And again, if it's truth, then trust can be built across the company at a fairly high rate. And then the cost of failure has to be very, very low there.

You reminded me of Brett Victor's idea of inventing on principle. That basically sounds like what you're saying, that Shopify has a very clear principle for how things should be, has a set of standards around how that thing should be brought to life. If you do that well at Shopify, you're going to do well and the business will do well. And if you violate it, you're gone.

And Shopify is not a place for everyone. It's a place for the people that really want to be on this place. There's no compulsion here for being part of Shopify on the inside or on the software. There's a super, super small percentage of people are on any length contracts, mostly because they want it to be for their own purposes. You can walk away from Shopify at any moment. And I think that's actually an important part of keeping things very, very healthy. I found those sets of ideas, they're not better than other people's sets of ideas. This particular mix constitutes a

a particular set of trade-offs where individual parts of what makes up Shopify, the company, could easily be argued as being unoptimal for all sorts of purposes. And people could come up even in system why this is actually underperforming how this usually works.

But the holistic picture is like a set of trade-offs that work extremely well for Shopify to make Shopify and keep it Shopify. And again, this is why you got to be an island. You have to be a genetic island, memetic island of some degree, because otherwise you're losing those tenants and then revision to the mean will happen. And then you have a company that's like all others.

That's not what I would want to work for. And I want to run a company that I would absolutely join. That's unnegotiable for me. I would see that as career level failure, if not. Those are my centers of exothermic energy that I'm tapping into to keep building.

I think we've done a lot of this in our conversation today, and it's been more philosophical this time, very tactical last time, but I love this framing of how to think about Shopify. Given who's listening, I'd love in closing just to ask how you would address an investor interested in the business and how they should think about Shopify long-term through an investing angle. What kind of shareholder do you want?

I want people to listen to this part of this conversation, basically, because everyone else has given up and I don't really care what the investment thesis is. I just want people to be part of it. The beautiful thing, Shopify is a founder-run company, obviously. It's also a founder share company. It's not one vote, one share in this sense. I vote 40% of the thing. So it's not the traditional share company in that way, which I don't think was a lot bearing concept of the thing.

I think being well-governed is independent of one vote, one share. Judging if a company is well-governed is what you should be doing. So what it is, is a ticket to ride. It's a, hey, this is how I build a company. This is how all of my friends here at Shopify think about this place. We are trying to build something of tremendous merit for people that also are trying to be of as much merit. And you're taking a tiny little percentage of that, which millions of little lights make a sun.

The numbers are big. The opportunity is stupidly big. This way, we are lucky. We have a great position in this company. And I think trust that you talked about earlier actually is one of the most defensible things. From an investor perspective, so many things where gap accounting really runs into issues in my mind, almost every asset of this company can't be put onto a balance sheet or an asset statement. How much do our customers trust us to wait on their behalf? People outsource innovation to us. That's a

big, big ask in a time of so much change. Shopify keep me current, keep me cutting edge is what our customers say and we take seriously. Our ability to attract the best engineers in the world asset with billions of dollars of the market cap, how many I don't know. We have a system of record for our customers that is a different status compared to a lot of other software that people end up spending a lot more money on because it's just

more co-op than partner, I would say. And these things are of long-term viability. Let's say tomorrow the all-knowing AI comes around that can write all software custom build. Well, you still need a system that you trust to be a store of record of information, a system of truth. I think an investment thesis can be built from that. And then I would say also like look at the track record of... I talked about first principle thinking, figuring out what's true. We got into this topic in basically every subplot we explored here.

but more specifically around the Shopify fulfillment network. It was true that this was better, this move, and this was not popular. I think you want to invest your money in terms of significant change if you're willing to go for a long-term ride, which I think you should, into teams that are willing to do what's right and correct.

even if easier possibilities are available to them. Because as I always say, it's not a principle unless it costs you something. But once I'm excited about other people who buy the ticket to ride, who wanting to support and wanting to buy into a highly alive, somewhat exothermic, sometimes wrong, but always willing to do something about it and reserving the right to wake up smarter every day management team in a space that is, I think, excellent. That's very big, constantly growing.

And of long-term value, I think people will build businesses forever against the backdrop of any abundance level of humanity or superintelligence presence.

Our conversations are always so fun and thought-provoking. I'm really excited for the one two years from now because I can only imagine the stuff that will change in the core layer of technology and the ways in which you and your team will deploy it and many other companies for their customers, hopefully on their islands. Toby, this is awesome as always. Thanks so much for your time. Really appreciate it, Wes. Thank you. Thank you.

If you enjoyed this episode, check out joincolossus.com. There you'll find every episode of this podcast complete with transcripts, show notes, and resources to keep learning. You can also sign up for our newsletter, Colossus Weekly, where we condense episodes to the big ideas, quotations, and more, as well as share the best content we find on the internet every week.