AI Agents Will Build Your Next Billion-Dollar Business?

>> It's really sweet to have you all here. like I think this has been a really really wonderful event. You can let's uh let's um uh thank the organizers here for the wonderful um system here. Team has absolutely cooked and thank you so much for joining. Um it's great to be here. Um we had on some of those notes earlier than um uh with with Harley already, but like um again like I I I said this before if a web would be invented today um uh and someone would propose it to um you know Apple and Google to try to get it launched in the app store, I would just be rejected, right? Like it's like this is not a thing. Um so we luckily all live in a counterfactual world where the web's open and it's available for everyone. the uh I launched my business um building applications within the web. You are building um incredibly inspired infrastructure for helping other people uh launch their web businesses. And um uh I think there's like two comp like like of our two companies are um uh at the sort of f definitely in a set of companies that someone would name for in in uh like the top few companies fighting for open web making sure it's there and making sure it's good and so um I I was looking forward to this conversation uh because like I think um first of all um you know this is under doumented and under discussed but also um uh there's no better to discuss us where things are going and what's now possible. >> Thank you. I I've always felt like our

companies are two sides of the same coin, which is helping people publish to the web. Um, and what I've always admired about Shopify is that it's so ambitious in that you've helped people not just publish, but actually build their businesses uh and and grow and scale their businesses. We've sat on the infrastructure side building the frameworks, the tools, the vast networks that help the internet be really fast. And I'm I I'm here sitting here at a

potentially the most exciting time of my life because we've always been about making it easy to create on the web. And with Nex.js and React and things like that, I think we made it pretty damn easy. Like we you know like I was obsessed with like can we make building and deploying one command like I started Verscell with that idea just one CLI command next dev you dev on the thing and then Verscell uh which is called now at the time asite now it's I had this like German like nmanure detour but uh one command you deploy on the web uh and then can we make that as accessible as as possible to a huge number of people and I think we sometimes forget because the world has changed so much in the past two years that you know the stats for how many web developers Verscell could address when I would make investor presentations or whatever was like 10 15 million like something like that 10 to 15 uh and it was a number of JavaScript savvy developers because we're like okay can we bring it down to the to like what is the lingua frank of the internet it's not Ruby right it's it's JavaScript and and is the thing that sort common is on every device is on every web browser. It reaches the highest number of people possible. Um but that was like 10 to 15 million. Now fast forward to today and I keep having this epiphany every day with a with coding agents uh with this uh you know whether you call it vibe coding or agentic engineering. we have now the ability to create at the moment of inspiration at the moment that we have an idea and we're opening up the the market and the total addressable audience to hundreds of millions or billions of people that can now build with natural language. Uh and obviously we can get into the details of like you know what does it really mean to build and and how deep do you go with those tools but I think it's the most bullish I felt about the web since I started forcell. Yeah, because again the the you know the web started as a academic uh docnet exchange right like and um you know at some point um um you can find stories um uh very easily about like the invention of image tech and so on like this is all documented which is kind of fun um in especially even even real time in the news groups um and um so the it's sort of been in this slow conversion of like um uh you know like what do out of all the uh duads that are required to you know create the modern web application in a it's it's I'll say it took the it's been the most roundabout possible way like it's really really it's been surprising um but like and and and along you know because it's just not it's not independent in the like obviously the web browsers and there's multiple ones and they had to sort of figure out how to agree and even how to pass the markup files but like really um Um there's also like you know regulations came came in you know unfortunately what started out extremely simple um with you know just some you know formatting tags um uh to to to make it actually work in a world of different device sizes responsive uh you know this like accessibility laws and so on. It's just like like there's this beautiful thing that we all need to be able to create uh you know documents, stories now, artifacts for uh and of course uh businesses, websites and online stores and everything. Um but it was kind of like um um buried um against this backdrop. Uh you know, React uh came out which helped a lot. Um uh it at least like wrestled everything into a functional direction which was clearly the right direction to go. Um when when did you counter react? What's what's your moment of discovery? Yeah, I was

working at a company that was fully monolithically PHP uh which was has always had a special place in my heart. >> Yeah. >> Because PHP >> Because PHP uh was the OG file system API. You create a folder, put a bunch of files in it. >> It's funny how to think back of this like now we like Yeah. You know what's really good? Give file systems. They like file system with folded structure and you know what like um you know opening tags in the middle and just like saying what you want. It's like it's it's kicking it old school, but it's like really on point. >> I was just reading the dogs on the uh liquid improvements. My first idea for Nex.js was actually to create a liquid style syntax for React. I actually haven't shared as much. It's like like breaking news like uh the earliest prototype because I I loved this idea of seamless dynamism that could be sprinkled over markup in documents. And what was missing at the time in the JavaScript community was that you needed to become like almost like a web server expert to take advantage of JavaScript on the server side. You know the whole like fast CGI and uh worker processes and demons and then if you get if you go further you start peeling that on and you get into cluster Kubernetes etc. So, it's like, wait a second, PHP only required for me as a young kid in Argentina to create a folder called web root and throw a bunch of PHP files. Imagine if you're an agent for the agents in the audience. You're salivating because they love when this coding CLI, they just love operating on the file system. Um, and it's so intuitive. It's also intuitive for humans. And so the idea for NexJS was like, can I bring that ease of use and friendliness to the React world, which React was also going down a little bit of that path of like, oh my god, I need to take a PhD uh in order to create my first React website. It was kind of nuts. >> Yeah. Yeah. And like, you know, it's I mean, obviously we can dine out on the PHP story. It is really really funny because like I honestly I I don't know like every time I'm feel like I like I I I see that deploying shop before it takes like 40 minutes in CI and like it then like there's an entire amount of work and um you know waste heat emitted into the universe by um just getting a new version up uh is pretty remarkable. Um, and like I just like you know back in the 90s here's what we did [laughter] like >> I I this is a this is a topic for me

because I'm so obsessed with deployment suite and you know one of the things that has made Verscell such a good uh counterpart to agents is that we started the company with just a CLI. I mentioned one command >> is incredibly inspired now like I mean it's like uh you know you couldn't have nailed the landing better on that like CLI tools are now the absolute meta. Yeah. And uh to your point about like how the web emerged so much another thing that gets me really excited is that we're going back to like the fundamentals of Unix. >> Um the the idea of processes, small programs that uh that plug in together open protocols, open standards, file systems. Uh I mean even the terminal, right? A lot of us I think are using agents on the terminal which is an invention like the VT 100 terminal goes back to like >> 1970s right and so there's a return to fundamentals that is extremely healthy for the ecosystem obviously with Shopify pioneering or pushing a lot of the new standards for agent eiccommerce and things like that uh MCP for example MCP is a great success I know that it's somewhat controversial because CLI sometimes are seen as being in competition with that. But anything that helps the broad ecosystem click together and democratizes and decentralizes power, I think is extremely uh good for for this new era that we're going into. >> So the area we're going into here is now the browser is incredibly uh mature. it like the um APIs it already has are uh extraordinarily underexposed partly due to this like complexity like it's like you know how is um uh like the thing you're doing relating to again the you know accessibility laws and all these kind of things and uh you know various um governments especially in Europe feel like they're web designers and tell us exactly what the check out button needs to say um and so um we we're managing But like um uh there are there is a there be dragons in some of these places. However um at the same time uh like this sort of calcified everything to a world where like templates reign supreme and like um we made very very configurable templates and Shopify kind of like bent this too. The thing the thing you read which we had on stage yesterday is that you know we going back to basics on like hey you know templates are good and easy to reason about for humans. um and for for for for agent and can be modified. And the reason why this is important is because um like again I think with our additions that we launch we try to like just we those are meant to be a love letter to the web. We we we we pay a lot of attention to make these engaging and fun and and and also um you know just delightful around new new technologies. But then um the next set of technologies that are coming right now that are becoming available like HTML uh like uh in canvas and these kind of things this doesn't even like all these words don't compose to something terribly exciting but when you actually figure out what HTML and canvas can do um it's mind-blowing to a degree that like no one's prepared for. So and then we're just sort of getting these gifts from the from from the platform. So this

is an amazing time because you have an agent um uh your stack makes it tremendously easy for the agents to reason about. It's like etched into the depths of the neurons like every like there's so much react code out there and there's so much RLF to uh make everything from their book. >> So you guys are adding Tailwind support uh to liquid templates that's really exciting. Um we similarly have a standardized ontailwind for all new create nextjs apps. You you touched on the power of APIs as well. One of the exciting things about even I mean even before agents the magic of Nex.js JS was that and this was super controversial because we a lot of the people that were coming from Ruby and Rails were saying >> I know people [laughter] >> they were saying why is next not like Ruby and Rails like why doesn't have an opinion like active record why doesn't have an opinion like O like are you versel guys out of your mind like do you miss half of the you know requirements and it was hard to answer that because we were living in a different world of composibility I mean headless or composable commerce was one of the things actually one of the big big growth engines of Verscell before AI happened. You know, one of our uh greatest growth years, much like you guys, was COVID where the business 5xed year-over-year and a lot of people discovered that they can move really really fast if if Nex.js and React as act as the front-end layer and then you compose a bunch of back-end services and that's why we're saying, you know, not that you couldn't build a full stack application with Nex.js. you could at the time. Now we've made it a lot easier. But hey, if you plug in the Shopify API, now you've gone from like trying to reinvent and boil the ocean to the best of both worlds, literally the best backend and the best front end together. And so fast forward to today and now you have agents love to operate on this level of abstraction. And um yeah, I'm just so excited about like next plus Verscell plus Shopify can do uh in terms of furthering the web and and growing it. The other thing since you touched on HTML if you all been watching closely what's happening with you know what do agents love to output when you ask them a question what what do they do? You know we've we've been very fortunate that a lot of them love to create applications with Nex.js JS, but they also love to write HTML. Simple HTML artifacts are uh you know, you could argue that the web browser is the world's most successful sandbox. People are excited about sandboxes as well today because agents get so much smarter if you give them a computer. Um, but an agent plus a web browser is basically unstoppable. Now, you know, it's it's hard to now predict what the web is going to look like because things like HTML and canvas, WebGL, Web Assembly, and agent is just becoming more and more and more competent. I do think the web is going to play a central role because of this popularity and widespread adoption. Uh but but it's uh evolving extremely fast >> right now. It's like it's so much is happening like I'm sure you have that experience too but like um what are some notable like you you wake up you see a thing you go holy um moments that you remember in the last it's basic for me I don't know it's pretty I would love to hear as well probably on a weekly basis uh the one that was just selling Vanessa uh uh in the green room cyber security massive topic uh I mean news breaking left and right about agents capabilities with cyber security. We run a lot of benchmarks to understand what are our agents really really good at and how do they compare to others. So the most recent like mind-blowing moment was seeing what models like Kim A3 and soul can do in terms of finding vulnerabilities in code. uh when you launched uh the auto research example for optimizing you know the like rendering performance of liquid or or the virtual machine or whatever like those those tasks that are so freaking hard for a human being. Um I just yesterday I reverse engineered a binary because I was I was asking one of our developers like why is the binary so big? >> That's a good right of passage for public company CEOs as well. Yeah. [laughter] And you could you could go to the source and you could look at all the linker compiler. You could look at the options that it was configured with or because agents just understand everything. It can look at the binary. It start dissecting it. It understands ELF just and it gave me like three big ideas for how this binary could get smaller. So those are the moments where like I'm holy crap. Like the fact that like me as a normie human have to so I I find reading binary impossible. >> Yeah. Well, >> uh, but >> uh, but >> I think that's a good thing, [laughter] >> but an agent just like has it for breakfast. Uh, oh, and sometimes I I I

we're talking also about like I think we underutilize agents every day. I'm asking myself how much am I carrying the inertia of the world where I grew up in and I lived in over the past few decades because one subtle one is they can ingest so much information, right? is like one million content switch like three Harry Potter books whatever people say but like you can just like paste vast amounts of content a and and I think sometimes we still treat it a little we we've taken the metaphor of it's it's a digital coworker a little bit too hard because a digital coworker I can't go into a slack message and be like I paste three Harry Potters give me an answer right um and so so I I'm just astonished by the level of capability that we now have at our disposal And the other thing that we were talking backstage, imagine a world where your NexJS and Shopify applications and websites self-improve. One way in which I'm convinced we carry the inertia of the world has been is that we think that everything we must be prompting. We must be the ones that sit down, bust open our terminals. Hey, can you make the button a little bigger? No. The agent can say, "Hey, I made the button bigger. I configured a feature flag on Versell. I run two permutations of the test. I'm seeing checkouts per minute go up. What do you think? I think we should ship it." And proactively makes improvements to your applications and infrastructure. I think that world is possible today. There is versions that we've already shipped. For example, our AI anomaly detection system. It can tell that your application is receiving, you know, anomalous traffic. A lot of you all are creating u storefronts for businesses that are based in certain geos. There's no reason that overnight Staples Canada should have a massive spike of traffic from Jamaica. And so today, we rely on people noticing these things or [snorts] handwired alerts. Verscell can tell you proactively that something anomalous is happening with an AI investigation. So the thing that I keep thinking about all day every day is how do we make these things more proactive so that it frees up our time to do the interesting parts of our business. >> Yeah. Yeah, like I I mean this is is such a good um way to think about it just because like your example especially on like observing the log files like it's clearly possible every like everyone on Shopify knows how to um you know go go add the data but like there's too much of it to to to watch it's extremely toilome um and um I mean this is really like just like automate

automate the toil so that you can work on the things that are actually making a difference like we want we want the merchants on Shopify make the the next uh version of their products with more of their strengths and few of their weaknesses. and um uh you know so that they can find the best possible product market fit and like anything that can sort of array itself into the um uh into the uh background agents, partners uh like like many of you here and uh you know Shopify and uh you know like all like your infrastructure layers, intelligent agents in all add all these layers like it just it can make it so much more fun to build businesses Because you know like sometimes you know like I don't know like I I've talked with um entrepreneurs who sort of when when you ask them hey what would like 16year-old uh uh you think about that um what you get back sometimes is something um which is actually surprisingly hedged. It it often takes the form of like 16year-old myself would be unbelievably proud of what we have accomplished but kind of dismayed for how I spend my day and um you know I think we can uh like make we can hone this to a much much much better point and just like like hey focus on the stuff that made you fall in love with solving that problem initially. I I do think there's something fundamental about like push versus pull like we all kind of like go and get information from the world and and and I think logs is such a great example because if you think about when people think about AI and agents um the co-founder of notion had a good zit today saying a lot of the world's still thinking about AI as search >> like outside of the San Francisco bubble like when you go and like talk to executives whatever they think of AI as like some really cool search engine. And so we have to actually also, you know, like realize that our own utilization of AI wherever you you sit can probably be a lot greater. And I think this pull versus push thing is quite fundamental because we're all used to this idea that we go to systems and we pull information out of them. But some of the best experiences in the world are literally when you get something good sent to you, right? Like I think this is why people love buying stuff with my wife loves shop pay. Like we love getting things uh that are going to improve our lives. Uh sometimes we buy good things, sometimes we buy bad things and we waste money, whatever. But like uh agents can give you this gifts. >> I gave my uh I gave my Hermes like a budget and the Shopif API and it buys gifts for me. It's awesome. Like I get like random stuff in a really dialed at this point. Like it's I always like it. It's really cool. I discover new merchants with it, too, which is even cooler. So, I I mean, again, it's like I'm not super representative here, but like um because I probably don't have to like necessarily look at the checkout screen so carefully. Um um but like it's like it's it still is a delightful uh experience. So, I I I think the amount of like there's trillions and trillions of dollars of uh enterprise value from turning some things which are currently pulled into Porsche. >> Yeah. Yeah. I I think I think also about

it in the sense of if you think about the very definition of intelligence like a lot of people have asked what is the definition of int is intelligence an IQ test I think the one that I've been most persuaded by is that intelligence is all about prediction >> exactly so if you're Toby young young Toby and you say like ah this web thing is going to be really big and this commerce thing is also going to be really big uh I'm going to place a bet here it's like you're basically predicting the future in the future more more and more and more of the commerce of the internet is going to shift to the web and to the internet and more and more merchants are going to come online and I think we have we now have to also have a humbling moment for ourselves that models like fable are extraordinary and superhuman at certain kinds of prediction. One such prediction could be what is the next product that I should buy? You know what is the next action I should take on a website? In fact, one of the reasons that we created NexJS was the frustration over a static web. The web the web only scaled until recently until Verscell and others have sort of helped solve this problem because we cached the crap out of it, right? So like going back to like the beginning like we we had amazing things like PHP. We were querying a database or a data source or whatever and then we were printing dynamic content out and then we put it on the front page of Reddit or Hacker News or DIG and then it shot the bed immediately and so we said okay what should we do? cache control. And so now we took this amazing gem that was by definition generative, right? We all get excited about generative AI, but the web that as it went more dynamic, it basically became generative, but then we cached it. One of the reasons that both Shopify and Verscell got excited about React server components is that we do believe caching is important. You should be caching certain component roots within the page because it has to perform. But the orchestration layer has to be generative. It has to be predictive. It has to be intelligent. And if you think about the the commerce at large today, it's not quite there yet. I know. I'm curious your perspective. I I don't think it is. >> No, I agree. And and and honestly this um I had a lot of uh work inside of Shopify is like you you saw like again change of direction with liquid but we also putting enormous amounts of um uh work into extremely fast liquid runtime. It's deployed to all the edges. um this like DIY the cell basically going on there which you know maybe I should um um be a bit more um inquisitive about um but like it's um we we go and um like we

are preparing for a world where no two people see the same website right because so like like caching is um like made the web work but basically was like also ended up being we ended up being significantly better at caching than we were at making fast systems partly because we could because the demand for dynamic systems actually wasn't surprisingly um low during these times and sort of we basically externalized a lot of um I think skills issues um uh and turn them into aesthetics and I think that's uh exactly what we now need to undo because >> it is just it is the greatest medium and we now can all build for it again without having to have a PhD in I guess uh German button gazettes. I I'm a big believer in the crawl, walk, run journey as well. I'm also very I try to be very grounded and realistic. There's two wolves inside of me. There is the SF AI psychosis CEO that like like vibe codes into the night and then there is also the let's look at commerce at large and realize that a lot of merchants just need to walk first. >> Um my my wife texted me a link the other day. a screenshot of a very very large nonshopify merchant where she literally could not complete a transaction. We're talking about latest iPhone San Francisco >> error page and so that merchant needs to walk. >> We could be talking about science fiction futures of aic covers and whatever but their website is broken. And so I think you know we all we are all responsible I think for just delivering really high quality experiences and meeting customers where they are. A lot of companies need to walk. Some you know some are ready to run. Some are like they've overcome the caching problems. They've over they are personalized. They're ready for adopting the next frontiers of you know and pushing the frontiers themselves. >> Yeah. Um good news probably is that someone in this room is working with fat merchant right now to fix this. So, so that's uh that's the nice thing. Um let's like uh I think I'm sure we have like a lot of uh common customers in this room, right? Like um you know like a lot of uh people here uh like like in rewards yesterday I think um uh half at the very least uh deployed to Vel. you have a very very deep um uh uh you know set of infrastructure. A lot of it is open source which I think I I particularly love because there's some wonderful the the open source and web thing. Uh um I it's just such a power combo. Um um and so love to um get your thoughts on uh you know how you make decisions to open source uh things and then you know just like also what parts of a stack are like underrated for the people who work on commerce. I think

open source we bet on open source before I guess before it was cool in a way uh you know it was very hard for me to raise venture capital in the beginning because I would talk to most investors it's like there's no way to build a business on top of open source and my answer would actually be very contrarian I would say no there's actually no way to build a business like this without open source >> because open source is a way that we distribute our software into as many practitioners in the world is how we can help them also contribute. There's a massive ecosystem play. So to talk about underrated things, the fact that we have a Shopify integration within Forcel, the fact that because NexJS is open source, every engineer at Shopify can extend it, make sense of it, their agents are capable is a perfect bread and butter. When you start using XJS, you don't just get a piece of software, you get an ecosystem, you get expertise, you get education, you get knowhow, you get the agents knowing about it. I've talked to people in the commerce world that always tell me would always tell me, I can build my own XJS. Some tried. they eventually say, well, why why am I building an island only for myself over here and then I'm actually trying to convince customers to join that little island that is separate from the rest of the world. So open source it really is uh the lay of the land in my mind as a developer. Personally, I would never imagine becoming attached or or uh blending my my identity to a tool that I know is governed only by a set of folks and I cannot read or write h what its future will be. I cannot fork I cannot introspect etc. Now that bad ended up being amazing because then agents became experts in our technology but it predated it. Yeah, actually there's a fantastic way to get into this because now just that I think about it like like the sort of island onto itself is of course incredibly good business for the people who own the island, right? Like so just this is something we saw very very heavily um even at uh predecessors of this conference. um that you know at some point it really became extremely fashionable to do uh headless um uh like commerce and um that sort of created you know like there was competitive pressure because like you know some people said like okay we're going to go straight bear we only going to be headless which really is funny to sell with the absence of an you know like a front-end system but like um like obviously people should compete and um um uh it it was then sold as like hey you can do anything in in the headless which is also is a good pitch but like there's a lot of really bad things came out of this right like because um uh so like replplatforming to a system that then had problems with scaling but more importantly these the websites themselves ended up like commerce is a unless you're extraordinarily lucky a visit visit rarely website right it's very different from uh Gmail or whatever you know so like having like to download over especially on Android phone over like over over 3G network um like 9 megabytes of data um to of JavaScript bundles >> a lot of people went headless really poorly because they use for example client side rendering >> we've literally seen business go bankrupt over it's actually really tragic and >> yeah so you know like exactly so so so unfortunately it really matters and um so you know the reason why we did hydrogen By the way, I saw this yesterday. I Googled I think it was a product and literally the description on the Google search result was this website needs JavaScript. >> Oh, like even in 2026 and and Google has a capability for crawling and whatever, but like I think a lot of people just got the infrastructure wrong. The headless idea is extremely good, but if you don't have the right infrastructure for it, uh it literally meant that your business could die. >> Yeah. like headless for a retailer um like is possible but like also it's possible to write any piece of software with a chip and a magnetized needle right like it's like that doesn't mean you should so it's um uh you know anyway so we end you end up in this bad situation we built um hydrogen uh initially just to like okay we need like a lot out here of of of of tools that like at least give you a large pit of success right and um um and then uh you know Now this is like we we have fully integrated partly because like of Vestal's commitment to open source right like this is like the reason frankly why we didn't go straight to um uh uh Vessel back then is like well we're trying to make people not build islands and uh we have to wait and see here a little bit. So thank you very much for being such a like have such a strong commitment to open source and launch gift after gift because you you run a very very AI pil company clearly it's actually remarkable like I I was telling you like you know agent browser is such a good tool look it up if you are the kind of person who works with terminals um they like just like they sell uh you know open source gifts that are in a lot of Shopify software because they're just like on merit like incredibly um inspired implementations of the kinds of things that we happen to exactly need for what we working on. I think this is this is awesome. >> So Mitchell Hashimoto, the founder of Hashior Corp, joined our board recently and he wrote this incredible article called the rise of the building block economy. I don't know if you saw it on X. Um, his argument is that agents can

write any and all software, but you don't want them to like reinvent every wheel and chase down every rabbit hole. Like, they're going to spend the infinite tokens and they're going to take forever, and they might not even do a great job for what it's worth. Like, imagine if the agent says, "Oh, yeah, you just asked me for a to-do list. Let me build a rendering engine from scratch and a kernel and an operating system." And like 10 hours later, you get your to-do list. No, it needs building blocks to work with. And so we've been fortunate to have uh spearheaded the the development of a lot of these modern building blocks. And I think as the world of composable commerce continues to develop, we're going to realize that what the future is is the combination of very high quality building blocks, all of which have to fight and survive in the agent economy for the right to exist. Today agents love Asian browser. If in three months that block is needs to be superseded by something else is to be replaced, should be swapped, you know, so be it. We've seen this in the commerce world with you know so many of the greatest commerce opportunities around content. So when content became composable, people were started to create amazing experiences and now we have you know agents that can combine these best of breath blocks. Uh and so we will continue to open source the the best ones we can find. you know, our product development philosophy is very simple actually. Like we solve problems, we create building blocks in the process and then we export them back out into the world. Uh, and you know, you all can can use them, fork them, and enjoy them. >> Specifically in this morning, we talked about um you know, just like Shopify's um uh like we we have River, which uh you know, builds a lot of software inside of Shopify. Um uh it it it's like incredibly well used. It's um uh it's the kind of thing like every company will want a river uh like thing. Um the um so we had some talks which were completely overflowing here. We we shared some articles and some philosophy around this. Um I I think very notable like you um like you for from building blocks upwards um you have an an AI gateway that allows sort of like the structured access to many uh models which is a really good piece of serverside infrastructure. You have SDKs that are open source that allow people to use this ideally but also go direct if that's what you want. like you I find that it's rare that like it's it's one thing to make open source that like is hyper specialized to your own stack if if it allows you to use whatever you want on top of it that requires like a different set of level of courage which I appreciate and um because like I mean best the best mode in the world for any business is simply merit right like this is what you really really want but following the stack upwards like then you have like a like harness system and it all ends up with like uh Eve, which you launched recently. I don't know how big of a deal that is, but like Eve is essentially you could you I recommended go to I think it's eve.dev. Um >> yeah, eve.dev. >> Yeah, it's a starting point with one init command, you get something that like is not very far away from having your own river. And so like now you you took the ideas of SDKs, the idea of CLI tools and the idea of composible tools um as different layers on top of each other all the way to a place where you're like okay here's I mean that seems pretty Ruby and Rails to me like like you you now have added the conventions, the file systems, the files, the um the structure and I I don't know like just like I I would love to you for you to talk about like maybe a little bit Eve because I think it's relevant and then also what you know what's how how do you lead a company to uh um compose things so excellently? >> Yeah, there's a couple things there. I'm sure you spend a lot of time thinking about this 0ero to one entrepreneur journey. >> Mhm. >> Mhm. >> Like what does an person with an idea need in order to be successful in this world? I spend a lot of time thinking about startups. Gan and I love supporting and working with startups much like the work that she did at Stripe. And I always ask myself, okay, what is the first thing I do if I'm starting a company? What is the very first thing you do? You know, and and I ask this question to people a lot and some say, well, I incorporate in Delaware. I pick a name. I register a domain. Registering a domain name comes up a lot. >> I don't know if you have a collection of domain names laying around. I register a lot of domain names. Um I recently bought rouch.com which I'm really excited about because there's a juice company in Austria. You know them like there's a conglomerate called rouch and I got rous.com. I'm really excited that I beat them to it. Um I love domain names. Domain names is one of the first things you do when you >> waterfront property of internet right? you know, [laughter] >> you you uh you might start a you know, I used to set up IRC servers uh because they need to chat with people. So, you might start a Slack workspace, something like that. Um but I believe that in the future a company will become synonymous with the agent that runs the company. And I'm not saying run in a Skynet sort of way like there's no CEO, there's just a robot. But literally like building a company is building institutional knowledge, indexing data, uh uh having business intelligence, dashboards, information, connections, leads, information, connections, leads, opportunities, sales, all of the Shopify backend operations connected through my FCB channel into my agent. And then I realized in theory this is possible for the but for the overwhelming majority of the world. It's really hard to create that Jarvis that helps you run the business. Some of us have the privilege that we did it at our companies with like large teams and lots of AI engineers and whatever. And so I think that by the way this is so profound of a shift that it might predate the introduction of Nex.js JS meaning like I might think about building my EVE agent to help me run the business before I create the website before I create the in fact I might tell my agents hey I need a storefront can you go build and deploy it and so this is where we're now basically spending a large percentage of my brain is going into what is the most delightful experience to do this least friction Like literally like I obsessed about the friction of you know time to first page for an XJS. I'm now thinking about like basically time to first token or time to first agentic interaction. I'm thinking about the openness of the web and the internet. Like for me uh the fact that you can bring in any model through the AI gateway absolutely non-negotiable. In fact, I would even argue that if you're choosing a particular lab or agent stack, it can end up being detrimental to your business, right? You lose uh negotiation power over the ability to get the best price per token. You lose out on innovations that other models can provide. Uh the model lab might, you know, sort of there might be a uh uh they might enter your business or your space or your segment or your industry or whatever. So freedom to own the agent that runs the business um and and customize it. By the way, also again grounding myself on earth, uh you have

to make these things secure, right? If they have access to data, uh privacy, compliance, confidentiality, security, you have to control the token flow, you have to control the ACL, you have to run this. So we we have a product called sandbox so that the agent can be guart rail. We just heard about is an agent that broke out of containment and hacked. I was I was wondering if you're gonna like because you you brought this up earlier like this um um this happened at from open AI right like >> the news broke that uh an agent >> it feels like a bit of a spot neck moment for cyber security but this this is like just like hard to follow potentially when you read this as an outsider but holy >> I'm completely unsurprised because I'm so in the weeds but I think the world will have their aha moment with this uh I mean we've been literally training models to escape containment. >> Yeah, >> that's what a post-training reinforcement learning environment does. If you're not familiar, we train, we pre-train models and then we put the agents in Docker containers and we say solve this problem at any cost. And sometimes there's like punishments that you can give the agent reward signals and counter reward signals and whatever. But like agents are extraordinarily competent at hacking. And so investing in the guard rails and security and sandbox so that you can get the best of both worlds. You want to run the business with an agent but you also don't want the agent to to sort of again like do do things that are counter to your interests. I think the the I mean my um I mean whenever I talk with people from the labs and they're like asking me what benchmark we are looking at it's like like I I tell them that the benchmark that they should look at is like you know make a million dollars with a new business on the internet like that seems like the best way to think about uh training these models. maybe like uh overfitting them to find zero days in Docker containers is like like has downsides but um you know so the funny thing about all this is like this is the world we live in now like it's like we we like the world is so much more cyber punk than I think the cyber punk models I grew up with and uh you know we have little uh aentic processes that kind of uh string together um exploits to just download the benchmark it was supposed to do well at instead of actually doing the hard work it so you know Like again against this backdrop um um we are all in the arena in terms of how to take the incredible capabilities and channel them into uh helping people rather than uh you know helping the cyber security teams uh uh you know um uh stay busy and um uh you know just building helping them build their own businesses because I think this is the the I think we are fairly uh maybe not uniquely but like uh The first step in our corner of the world, in the retail world that people often do when they uh want to build businesses is um uh that they realize that their plan B for their career might actually be better than plan A or somehow plan A is over because they you know something's happening in the company that they're doing and they say you know what I always don't try to do my own thing and um the number one time slot for signups on Shopify is always the lunch break of a current time zone. So it's uh you know people are often this is their entry point into the world of entrepreneurship and so um you know we have a unique uh perspective there. our AIS love often are used to figure out how to incorporate a business or um uh register domain and so this is what makes our corner so much fun because we get to be there right from the beginning and it's also a big um obligation and um I think therefore Shopify like what we celebrate is when people uh build things uh that are meaningful and are shared and um somehow make other people's lives a little bit better um by encountering them and so it's uh wonderful to uh you know get to you know work expertly about these topics if you're here. >> Cheers to that. Thank you Toby. Thank you. [applause]