Auren Hoffman (00:38.436)

Hello, fellow data nerds. My guest today is Guillermo Rau. Guillermo is the founder and CEO of Vercel, a cloud platform for building and deploying web applications. He's the creator of several major open source projects, including Next.js and Socket.io. He's also a dual threat CEO. He runs a solo VC called Rau Capital. Guillermo, welcome to World of Daz. Really excited. Now, Vercel's kind of at the forefront of a bunch of major trends in

rauchg (00:59.639)

Hey, thanks for having me.

Auren Hoffman (01:05.996)

software development and programming right now, like edge computing and serverless computing, where do you think like the industry is going?

rauchg (01:13.322)

Yeah, there are a lot of trends that underpin a change in how we're building and deploying applications. And to me, the thing that relates to my personal history is we're switching toward a front-end first way of building applications. So I have a background as a front-end engineer. I grew up in the JavaScript world, later the TypeScript world, Node.js, a lot of projects that had this dream

Can we write things in one language and run them everywhere? I actually wrote a book, uh, many years ago called universal applications. And kind of this idea of you can write code that gets deployed to the cloud, to the edge, to the client. And you can have all these kind of potential touch points with the customer experience. So.

Auren Hoffman (02:07.068)

Because it used to be there'd be like a JavaScript kind of front end thing with a front end team. And then there was like maybe the Java or, you know, guys on the back end or something like that. And yeah, maybe they wouldn't even talk to one another or they'd look down on one another.

rauchg (02:15.142)

Yeah, and you had to coordinate them. Yeah, so like literally what we had to do, yeah, what we had to do was like, the Java team needed some templating language that was kind of ad hoc. And then there was a JavaScript engineer team and they had to like also figure out how to re-render that thing. Because what's happening over time is a lot of the value in new groundbreaking applications.

is in the user experience, right? Like, ChadGPT comes out. What is it? It's a great user experience to interface with an AI model that is actually just a small twist on things that they had been trying in the past in the case of OpenAI, right? But like, they just nailed it with the user interface. They nailed it with the product. Overnight, you know, tens of millions, hundreds of millions of users. Funny enough, ChadGPT is built with our framework.

Next JS. Okay, so funny enough, chat GPT is built with our framework Next.js. And they create this awesome user experience that has captivated folks, right? Like, there had been other AI products, dashboards, things that they put out. But this particular combination of an incredible

Auren Hoffman (03:38.492)

Heh. Heh heh heh.

rauchg (04:05.362)

revolutionary AI technology with the right user interface can scale and travel the world so quickly. So what we did with this framework and the Versel platform was, what if your starting point in how you build and deploy software was the actual user interface, the thing that you're gonna deliver? And then you work backwards into the backend. And what's really interesting is that this kind of revolution is in juxtaposition of another revolution, which has been

the cloud in itself and the rise of the API economy, right? Like a lot of the things that we used to build from scratch with that Java team we were just talking about have become off the shelf API services and or really well optimized data platforms.

Auren Hoffman (04:45.74)

Yep. There's some sort of like, and it's some sort of, it could be an internal microservice or, you know, in many ways it's an external service like Stripe or something.

rauchg (04:57.23)

Correct. So this decoupling had already been sort of happening. We're missing that technology stack that really emphasized, okay, like we can't inhale the microservices part. What about the front end part? It used to be that the front end was kind of like this like last minute, like I'll scramble something. But how do you like justify that in juxtaposition with the fact that the best products in the past 10 or so years?

have been from the companies that have created a incredible set of innovations on the front end side. So one of the things, one of the secrets of Next is that under the hood, if you like opened the hood of the car, you would find an engine called React. React is this UI library that came out of Meta that powers some of the most engaging and interesting user interfaces on the web.

of the past few decades, right? So the newsfeed, the idea of this auto-scrolling experience that comes up with like near instant recommendations of content, their notification system, their chat system, all of this is predicated on putting a lot of investment into the front end infrastructure. The rest of us, the mere mortals that didn't work at Amazon, Meta and Google, were kind of...

not seeing this matrix of I need to prioritize the front end. And I think part of that was that again, like there was just so much work to do on the infrastructure. We're starting to decouple the monolith of the backend into microservices. We're laying out the data infrastructure foundations that we didn't get to the front end piece, I think. We didn't get to the front end chapter. The other really interesting thing is

And I think why Versel has seen so much traction with e-commerce in particular, there is a cohort of folks that are very incentivized to hop on the front end bandwagon. And those are the folks that have been instrumenting their businesses. So e-commerce, there's very clear data that the old fashioned data I'll share is for a hundred milliseconds of better performance that you're delivering the

rauchg (07:17.222)

impression of the page, Amazon saw 1% fluctuation in sales. That was like the canonical like study that everyone shared. Most recently, Deloitte and Google put out a follow-up that was instead of focusing on desktop, like Amazon did, was predicated on mobile devices. Now for a hundred milliseconds of improvement across travel, luxury, marketing and beyond, they saw an 8% lift in conversion.

So the decoupling of backend and frontend and optimizing the frontend experience is already kind of the standard in e-commerce. We call this the composable architecture that you can sort of make Shopify not a monolith. Shopify doesn't keep track of your entire frontend, your plugin, your liquid templates. You just use it as an API, as a headless GraphQL backend. So that's just one example.

Auren Hoffman (08:13.167)

Yeah.

rauchg (08:17.342)

one vertical that is seeing incredible ROI on, I'm going to start really emphasizing this front end thing. So my thesis and what we're seeing with Vercell is that this is not just like something you optimize. You can actually say, I'm going to start building front end first. So think about the world of AI right now. It's very hard to compete with open AI on putting out an

a better LLM than GPT-4. So what's your innovation opportunity? Velocity of integration of the API. Experimentation. Getting your first chatbot, your first support center, your first, you know, AI recommendation system. Getting it out quickly, iterate on it, and then connect to the right backends instead of kind of reinventing the wheel from scratch.

Auren Hoffman (08:47.64)

Yeah, very expensive.

rauchg (09:13.046)

So at Versel with this idea of the front end cloud, we're kind of going all in on like, hey, front end is the profit center. And the back ends that are sort of under differentiated, which are not all of them of course, are your cost center. And so far it's been an interesting journey.

Auren Hoffman (09:31.756)

And there's so many businesses where if you just like that you interact with, whether it's travel or, uh, commerce or banking or government, where just the UI is just atrocious, it's horrible. It's so hard to use. It's constantly breaking. It's really, really slow. Even things where like, even sites I use all the time, I would throw in like a Salesforce or LinkedIn or something like that. They they're kind of, they're kind of atrocious. They don't really value the user.

rauchg (09:46.478)

Thank you.

Auren Hoffman (10:01.156)

very well. Do you think those are the businesses that are ripe for disruption or do you think they're just going to come around and they're going to actually start investing in some of these UI things?

rauchg (10:09.91)

It's both. It's both. I like the LinkedIn example because LinkedIn and you and your podcast audience know this even better than me. They had to be burdened with creating a lot of the foundational projects in data in search engines. I mean, yeah, they kind of like how to like invent the cloud. So major kudos to them. So I think what you see with some of those products is that, yeah, like the front experience does lack. I mean, I'm not going to sugar coat this. Like every time I open it.

Auren Hoffman (10:23.364)

Yep. The back end. Yep.

rauchg (10:38.322)

it greets me with a spinner and it waited for quite a while. And the argument that we've been making is again, each 100 millisecond of spinner. And that's why we talk so much about one of the trends that you called out is like edge compute, server rendering, the global deployment strategy as opposed to just one place. A lot of that is predicated on like let's render less spinner, more content, more action, more transaction opportunities.

So yeah, so I think two things to answer your question concretely, like one is they will have to adapt and sort of integrate the newer ideas and tools. But also I do think there's lots of disruption opportunities in the market right now. And for, I don't wanna sound like a parrot on AI, but there is opportunities where you're saying like, I'm not just gonna make a faster product, I'm gonna make a more intelligent product.

And again, if you're composing services, you can iterate your way to get to that product market fit, I think, much faster. So we end up talking a lot about iteration velocity. This is kind of like one of the big lessons of our cell so far. We started out on like, fast, fast wins. And the thing that we hear a lot about is like, but iterating faster brings even more value to my business.

because I find the right thing to build sooner.

Auren Hoffman (12:07.812)

How do you think of, there's kind of an art and there's a science of UI and UX. And for a while, it seemed like the science was taking over. It was like, we're going to A-B test everything versus the orange or bright red button or something. And how do you think that is evolving?

rauchg (12:23.627)

Yeah.

This is a great question. This is a great question because you open up a Pandora box. Some people like, uh, I believe like folks, uh, from Airbnb, the CEO of Airbnb was just at, uh, the Figma conference and saying like we're just AB testing too much. Um, I also see some people that just don't experiment enough because they're a victim to the limitations that they're stacked. They're like experiment what? Like, what, what is that? I'm

Auren Hoffman (12:51.184)

They can't, it's just too hard to move, yeah.

rauchg (12:55.102)

I'm over here just trying to make sure that the website doesn't go down. You would be surprised about how many companies I talk to that are saying, I send out a newsletter and I literally have to cap the audience size even though I could be sending it to millions or tens of millions. I cannot because the site goes down. Imagine going and talking to them about A-B tests and whatever. So we have that cohort of customers, right?

Then we have the customers that are literally limited by their infra. They want to AB test more and they know that maybe one team can, but they cannot. They just didn't have enough of the infrastructure, like platform engineering teams, attention, time, budget, whatever, to provision that more complex infrastructure that enables more complex workloads, more dynamic rendering instead of like putting things in a static site and things like that.

everybody in the world needs the ability to roll things out with different strategies. Even just today, we announced our integration with LaunchDarkly. LaunchDarkly is a feature flag management provider. One of the selling points is experimentation, but the other one is that if you want to launch a thing, you should be able to control the rollout process. You should be able to roll things back instantly.

So it's about not just experimenting, but also having the confidence to ship things and not having an engineering and product team that are terrified of messing with a machine and touching the different knobs and really empowering them. And believe it or not, a lot of this is predicated on the front end itself.

Auren Hoffman (14:45.752)

Now, when COVID hit, I decided to get back into programming. And I was really surprised at how hard it was to set up my local environment. It was just like, it was a mess. We're really just talking about three years ago and then recently kind of redid it and it just seemed like it was way, way easier. You have everything on the, you have a lot of things more cloud-based now. Do you think that frustration that I had three years ago was kind of quickly becoming something of the past?

rauchg (15:15.262)

I think one of the biggest selling points of Next and Vercell is that we're obsessed about the local development experience. And this is very interesting because I mentioned earlier, like, why is edge computing exciting? Why is accelerating and deploying a global front exciting? Well, right now I'm in Switzerland. I can go to any of the storefronts that deploy in Vercell and I can have a personalized experience that knows where I am.

who I am, what I've visited before, what I've ordered before, and the right currency while I'm here, which is Swiss francs or euros or maybe it's even dollars because it knows that I'm logged in and I'm not from here. So to do that at scale is extremely hard. It's literally a globally distributed system. If you ask a computer scientist, they'll tell you one of the biggest challenges in computer

rauchg (16:14.93)

and reasoning of a globally distributed system. So how does that fare with the concept of, I want a good local development experience. So one thing that Next.js does that's really clever is, when you work locally, it feels like a monolithic experience. I can touch any part of the system and kind of like develop and work on the feature that I need to work on. When I deploy, especially on platforms like Vercell,

it becomes a global distributed system. So we create the infrastructure from the analysis of the application. So we're kind of like, we found a very strange mythical, like best of both worlds thing. For the dev, great developer experience, especially in local machines. For the visitor, a global system that tailors and runs compute all over the globe, just in time. And all of this without having to kind of staff.

platform engineering team with lots of SREs and Kubernetes experts and all of those things that you outsource that developer infrastructure to Vercelli instead.

Auren Hoffman (17:25.948)

There's this kind of new wave of, let's say, developer tool companies, for lack of a better word. And I put Vercell in there or Replet or some of these like really cool companies that are really kind of changing the game. There's a scenario where you guys could change the game so much that we could unleash this whole new wave of innovation and we can get a whole more layer of productivity and really cool things that are happening in the world.

Like how do you think that is going? Like where are we on that, on that journey?

rauchg (17:59.986)

Yeah, I think we're part of this new code of companies that are saying, you know, first of all, power to the developer, right? So the developer experience is a key part of a modern, successful product led and product centric organization. So how does that translate into the traction we're seeing with Vercell? Well, every company is looking to become

a digital product organization. So I was speaking recently with a CIO of a very, very large product company whose product you've heard of, you love, you probably have in your household. And what they were saying is like, we don't want to be disintermediated. Art.com is just as important as the hardware that we're known for that we put in, that you have in your household. But if you, going back to your perception of like...

Why do I go to some websites and they're so slow? The average airline booking experience, even the average e-commerce experience, why are all these front ends so clunky? So that's what we're looking to change. And again, the tailwinds are in our favor because the people that are going to make your.com great turn out to be these developers that care about this developer experience that are not the manufacturing and hardware engineers.

that you became known for that particular product. Right? So as you look to complement your skillset as an organization, my bet is you're gonna look for comprehensive platforms that solve a wide range of problems for you. So this is the interesting thing about Vercell. We took on a burden, which was to actually create a framework like Next.js. When the traditional wisdom was like, platforms don't...

have frameworks. They don't even write the integrations. Think about AWS. AWS just sells you the primitives. They're like, well, the tools are someone else's problem, right? You come to us when you exactly know what you need and where in the world you need it. In fact, for those that have used the big public clouds, you know, the first question they ask you is where in the world do you want to put in the workload? How do I, how do I...

rauchg (20:22.67)

make that gel with the fact that I talk to these CIOs and they tell me, well, we're expanding in North America, in Mexico. And do I tell them, well, go to the AWS console and select the region, US. Yeah, yeah. And it's not even called Oregon. It's called like south, south west, north, like UDP one, three. Like it's hard to figure out the cloud. So I think we're entering this phase of

Auren Hoffman (20:34.192)

put in Oregon or something.

rauchg (20:49.306)

not to dismiss the incredible work that they've done on building those primitives, but the average company in the making or even established will not be shopping for primitives. They will be shopping for the vertical integration between, think of it as the vertical integration between operating system and laptop that the MacBook and Apple sort of pioneered. What does it look like for the developer ecosystem? And that's what, for example, the integration between Next and Vercell,

will return accomplished there.

Auren Hoffman (21:21.668)

And if you think about even like, if you think about the UI, like some of the worst UI places are like the developer tools, like AWS, like every time I log in there, it's just like a mess. I can't navigate. I don't even know what's going on. It's like, it's like they kind of expect everyone to command line it or something. And like, they're in some ways, they're like not treating their customers in a way that they, they're not treating those customers in a, in a way where they really respect them.

rauchg (21:41.654)

Yeah, it's rough.

rauchg (21:52.19)

Yeah, I think it goes back to that idea of like a front end cloud. Like if you go to the Versailles site, like, you know, we, we want to make our product do the talking, like pages are snappy. The transitions between pages are snappy. Like, cause again, we're, we're prioritizing the front end experience. So I think the, the way that, so the obvious question is, okay, I've been, I've been going all in on cloud for the past decade or so, right? I'm a big company. What do I do with all that cloud?

And what do I do with all my plans to get more of that cloud? And how do I not fall for the trap of like, I end up shipping a site or application that doesn't live up to the standards and frankly, like, even in more crude terms, it doesn't live up to the sticker price. Like I go and look, like I go to my like, uh, you know, statements and, and billing centers and say like, Oh, we're paying all this much. And then I open our application. Do I get what I'm paying for?

So the question would be, if I'm in that situation, how do I move forward? So the way that we bring Next and Vercell to market is very incremental. So we don't tell you to move your data to us. We connect to your data. We connect to your GraphQL gateway, REST API. We connect to your microservices. We launch the protocol Vercell Secure Compute.

that allows you to run our rendering functions for the pages in a secure environment, dedicated, VPC connected into the cloud. So we actually connect to AWS. We build on top of their primitives. We can connect to other clouds. So our go-to-market positioning is very incremental. Look, you already have a backend cloud or you're looking to transition to one. Here's how you actually ensure you have a great...

end user experience.

Auren Hoffman (23:50.104)

If you were advising like other founders to build successful companies around popular or up and coming open source projects, like what are some of like the do's and don'ts?

rauchg (24:02.686)

Yeah, that's a great one. I'm happy to report that I and my colleagues at Vercell have been sort of starting to put together a framework for how you would go about doing a good dance between open source and a successful infrastructure business. So one of the concepts that we came up with is what we call the framework to find the infrastructure.

So for those that have been around the industry for a while, you might remember like software defined networking. So that's kind of where the term gains its inspiration. With framework defined infrastructure, the open source framework, when deployed on our platform, defines the infrastructure that we create. So by definition, the open source project gets amplified with a service that is fully managed, or we would call serverless today.

where the value add on top of the framework is very clear. Like, look, I can go and operationalize this project myself. I might even be able to get it into a single region Kubernetes pod. But in the world, in the context of Versel, I'm looking to actually optimize and deliver a great production experience for visitors all over the globe. So we found that good balance between open source and giving you the...

primitives or the ability to orchestrate those primitives yourself or the sort of infrastructure as a service solution that also comes with an additional layer of what you would call like workflow and collaboration tools. So the way that I explained this, yeah.

Auren Hoffman (25:43.584)

And so is there any like core checklist for a founder to kind of like think, think through? Like they've got a cool open source project. It's got, you know, 500 stars on GitHub. It's starting to take off. Like how, how should it, how should they be thinking about those trade-offs?

rauchg (25:49.922)

Yeah.

rauchg (25:58.014)

Yeah, so one thing to think about is, again, the service has to provide a true power-up on what the company would otherwise do with. And again, I don't like to withhold features because if I'm selling to developers, it's in my best interest, in fact, if they use the framework to its full potential. I want them to be excited about the framework. I want them to love the framework. So that's a very clear...

Auren Hoffman (26:09.804)

Yeah, not withhold features from the open source. Yeah.

rauchg (26:27.086)

part of the framework there. The other one is the experience aspect. And I'm obsessed with this just naturally, but the biggest advocate within the organization will be a developer for whom your project makes their life easier. What I would add to that is, can you think about how your service makes the lives of their colleagues easier as well? Whether it's, you know,

IT, security, DevOps, do you have answers to their concerns? So I think a lot of entrepreneurs might be after a local maxima of like, well, I'm just after the experience. I think this is honestly something that I see quite often, which is you might not even be designing with the right inputs. One of the things that we do that...

we've become known for it in the space of e-commerce, for example, is when Black Friday comes around, we guarantee that you thrive during Black Friday. So we had a company that was an early dot com company, like the companies that were like their CEOs or CTOs were like visionaries and they were like, this inner thing we should be selling there in an early mover and they were proud of that like a badge of honor.

rauchg (27:52.823)

One of those companies recently moved to Versel and what they told me is in our 20 year online history we had never not gone down during Black Friday and the e-commerce season. So going back to open source versus platform, I'm giving you a very concrete value out there. Look, you can go and figure out all this infrastructure by yourself. In fact, it's open source.

Auren Hoffman (28:06.128)

Crazy. Yeah.

rauchg (28:18.706)

It's built on standards. It's built on like, it's an HTTP web server at the end of the day. It speaks standard protocols, et cetera, et cetera. Now, I have engineered an infrastructure that has been not just even load tested, is the primitives that we've selected are the ones that can actually scale to like really, really demanding workloads. So that's another example of like, there has been open source projects that have actually developed an anti-reputation.

because everyone loved them because they were open source. And then the mythology kind of started about pitfalls in prod. So that was kind of like, it's funny, like each generation of entrepreneurs learn from the challenges of the past. So I was obsessed with that problem. I was like, I cannot make open source be the downfall of the reputation of like, what happens to this thing at scale. And I suffered from that quite a bit with Socket.io, which was one of my open source projects because

There was only so much I could do to like kind of transfer the best practices of how to scale it at a massive concurrency. And like I could write a readme, but that's, that's as far as open source can typically go and like where. Services and infrastructure can start.

Auren Hoffman (29:35.128)

Now you're a dual threat CEO. You're both a successful CEO and you're a great investor. Now being a CEO almost certainly makes you be a better investor, but does being a good investor make you a better CEO?

rauchg (29:48.198)

Yeah, I love this. I love this tweet from Naval that came out a few weeks ago, which he said, give advice and then apply to yourself. Because like a lot of folks, I think I tend to give I think what I would consider okay advice, I add a lot of caveats. I say like, look, this is what's happened to me. This is what I'm serving out the market. So like, that's one of the things that I do that again, like I think folks tend to they like to hear it from an operator.

it's harder to hear it from a venture capitalist, especially in...

Auren Hoffman (30:21.452)

Yeah, or even somebody that used to be an operator, but they last operated seven years ago or something like that. Yeah.

rauchg (30:26.73)

Yeah, things move so fast, right? Like there's operators that build incredible, you know, dev tools or infrastructure companies, but it was pretty serverless, pre edge. It was pre DX and developer experience is not a thing. Like so many things to your point, like so many things change. But then there is the other one that to your point is about like, Hey, I'm, I'm telling you about advice that I want to apply myself right now. That's how in the trenches I am.

Auren Hoffman (30:53.068)

Yeah, yeah, yeah. You don't wanna be a hypocrite. Yeah, yeah, yeah.

rauchg (30:57.99)

So that's why I said I love that part of my job or job that I added to myself because I learn more about myself. I learn more about our journey. I rarely have the time to even think about the past and reflect on what we've done. So it's a great opportunity to be like, oh yeah, we did it that way. You should consider that or you should not consider that because it didn't work for us. So

I think that's one part of it. The other one, quite frankly, is because we're in an ecosystem business, we are part of this huge cloud and Vercell is part of this front-end cloud, which is getting to become a huge movement. So there are a lot of entrepreneurs that can add an integration to Vercell or think about how to best connect this to all the modern frameworks.

and they'll see a lot of, you know, perhaps customers, interest, whatever. Like because I'm interested in seeing this ecosystem flourish, I'll give you an example. There's companies that are doing front and first off, front and first experimentation, front and first, you name it, CMS. So there's a huge ecosystem of companies that are just so exciting to invest in because

They're hopping on the right waves at the right time.

Auren Hoffman (32:29.5)

We're starting to see a lot more dual threat CEOs. Do you think they'll be taking a greater and greater market share from the traditional venture capitalists?

rauchg (32:39.346)

Yes, in the early stages, I find it hard to believe that again, that entrepreneur that is looking to like go from zero to one will say like, oh yeah, you really need to partner with this particular individual from a firm as opposed to I'm going to partner with again, folks that have very concrete hands on experience on the most pressing problems that I have at hand.

I do think it's super healthy for the ecosystem that investing is getting democratized. Overall, I would love to see more diversity in the space, not just, again, from there's five firms that do all these Silicon Valley deals, but there are a lot of great individuals that have great taste, that have great experiences, that give great product feedback.

women, men, there's just so many awesome people out there that can be empowered to help create the companies of tomorrow. So it's a whole, it's a societal thing the way I look at it. It's exciting that this is happening.

Auren Hoffman (33:53.796)

Now, Steve Yegge wrote an interesting piece recently discussing the significance of like data motes in determining the future of success of these AI companies and AI enabled companies. This is a data podcast. Like what's your view of the role of like data in an AI driven world?

rauchg (34:10.13)

Yeah. Where'd I start? It's so exciting. So thanks for asking. Number one, data in motion in data in the hands of your customers is supposed to like internal data is what Vercell Plus your data systems enables. So I'll give you one quick example. We built a lot of amazing data products at Vercell, like Vercell Analytics, Vercell Speed Insights. I'll give you an example of Vercell Speed Insights. It tells you in real time.

how your pages are performing from an end user point of view. We give you a score and we tell you, look, in China right now, you have this score. In Switzerland, you have this score. So in another parallel universe, that would have been an internal dashboard. Right? Like, because I care about that. Actually, this is one of our key customer success metrics. When an enterprise partners with Vercell,

Auren Hoffman (35:01.924)

Yeah, yeah.

rauchg (35:09.714)

I take it even as a personal responsibility that your stuff is fast. So we pipe the data, securely, privately, etc., from the front ends into your hands and into the hands of your account manager, for example. So again, there was a world where that data only solved and fit that goal. Now, I actually have a very successful self-serve product-led growth motion.

predicated on the fact that I'm not just giving you next year as hosting. I'm going to give you a comprehensive platform with built in real time analytics. The time to market for this product for us was extremely short. Why? Because most of the workload was on the front end and we're front end experts and we dog food of our cell. So obviously like your mileage may vary, but we piggybacked on world-class data infrastructure.

We partnered with this company called Tiny Bird, and they use ClickHouse as their engine. So if someone in your audience is thinking, okay, I have a lot of data here sitting in my analytics colon or store. I have a lot of data sitting in my warehouse. I have a lot of data sitting in S3 files. Well, what about how quickly you can turn that into product? That is one of the value adds of our front end cloud. So that covers the data part.

And that's just one area, which is creating data products. And turning...

Auren Hoffman (36:41.392)

from like the actual data like that these LLMs are trained on, whether it be just crawling the internet or YouTube videos or, you know, different, different question and answer sites and stuff like that, that's out there. Like how do you think these, like, how do you think these, whether these, like whether it's the broad LLM models like open AI or, you know, maybe, maybe more specific models.

that are like specific to a very, very like whether it's a coding thing or legal or whatever. How do you think those are going to be thinking about like the truth behind the data?

rauchg (37:21.898)

Yeah. So onto that next category, which is broadly like AI and its relationship to its data. I think all of the data that you mentioned is most relevant at training time and at tuning time and sort of the life cycle of evolving your model. Where Vercell comes in is that eventually you deploy that model as an inference endpoint, which is once again an API.

So just like Vercell can get content from your CMS as an API, one of our larger e-commerce customers uses Salesforce Commerce Cloud over an API to get its data. You can actually also incorporate this LLMs. So there's a couple of trends that we're seeing that are very encouraging. Number one, there's just a tremendous level of integration between like Open AIs, APIs and Vercell Frontends.

We actually just announced the Versel AI SDK to make it really easy to connect to these LLMs and create frontends from those. So you can open actually in your web browser, chat.versel.ai. So what this is, is your own chat GPT with the world-class frontend, open source, deployable with a couple clicks. So...

In Versel AISDK, which is what sort of brokers that connection between the LLM and the front end is actually provider agnostic. So this is onto the second part that I'm excited about. We're going to see the rise of models that are open source. We're going to see the rise of models that are deployed in your own infrastructure with your own data sets that are trained on your own sort of experience that are fine tuned on your own.

terms with your own data and labelers. So for this models, you can use platforms to facilitate its deployment. And you can use a Vercell AI SDK to again, integrate it into an actual real world product. So because there's so much going on with AI, we saw an 800% increase in AI products deployed on Vercell since the beginning of the year. We just launched our Vercell AI Accelerator.

rauchg (39:46.626)

to actually put in touch creators, developers, companies with all these AI backends, a lot of which they don't even know they exist. So I'll tell you another one that I'm really excited about. 11 Labs just announced a round from Andreessen Horowitz, Nat Friedman, Daniel Gross, and I participated as an angel. I'm really excited about this company because it's voice AI. And I could already picture once I heard about

I was like, holy crap, the API is so easy to use. Imagine all the products that are going to be created on top of this. So I think we're going to continue to see this vertical AI companies. They specialize in one problem domain, and then they empower developers to create experiences on top of it.

Auren Hoffman (40:23.128)

Yeah, amazing.

Auren Hoffman (40:35.312)

Now, a few personal questions. Like, you really kind of started working when you're like a teenager and just like kind of got into it. I don't even know if you've ever even finished high school or not. You just kind of like started. Like, give us a little backstory too to just how that happened. And do you think that was just like completely unique to you? Or do you think there's just like, there's millions of other people who kind of follow a similar path?

rauchg (41:01.258)

Yeah, I'll start by saying like, I always share my story saying like, it's really unique. I don't necessarily recommend it to everybody. So don't go and quit high school and things like that. I didn't finish high school. I had maybe like another quarter of studying left. And it was because of my open source work, I got recognized by a company in the US from San Diego.

and a company in one of their clients actually were not too far from where I'm visiting here now, which is in Lausanne, Switzerland. So by this time I had already been working like through the internet for quite a few years, but I kind of had an opportunity to like literally like launch my career, travel the world, present at conferences, develop more open source software.

and then eventually moved to Silicon Valley to start my own company. So yeah, it's not the story of like, you know, I dropped out of high school because I had no clue. Like, I wasn't interested in like, I wanted to pursue my dreams and they were like, kind of like not very well defined. I started

Auren Hoffman (42:16.612)

Right, you already like, you were making money, like you had a job, like actually doing these things.

rauchg (42:19.946)

Yeah, yeah, I was supporting my family financially. We grew up in Argentina in an area of the outskirts of Buenos Aires that is not particularly rich to put it nicely. It was pretty dangerous even growing up. Dollars, it was insane. It was insane. It came with its challenges. My dad even started relying a little bit too much on it. My ankle was not very reliable because I was going to...

Auren Hoffman (42:33.22)

And you're getting you're getting paid in dollars, I presume, too, which was nice. Yeah.

rauchg (42:48.918)

first elementary school and then high school. And it was like, you know, I have to focus on my work and I have to focus on school. And like, it was, it had its own tricky parts to it. But overall, the idea that open source and like the web and my fascination with front end sort of unlocked that journey is pretty amazing. And we're hoping that, you know,

there'll be many stories of folks that I'm not necessarily saying like, don't go to high school, but like we are in the best time in history, I think to really teach yourself anything, um, Chad GPT, uh, AI auto-completion in your terminal, AI being integrated into every piece of the stack. Uh, I mean, everybody, everybody can really, if they put their, their minds and, and effort into it, anybody can learn how to code.

anyone can learn how to create an AI first business. So I'm hoping that this does inspire people, but again, the high school part is kind of tricky.

Auren Hoffman (43:59.82)

Now for people following on video, they might be able to see that you're super fit. Like what's your, you're kind of known for that. What's your fitness routine?

rauchg (44:08.846)

I don't know if I'm not for it, but I mentioned if you put your mind to it, I think the one thing I would, this is not advice for entrepreneurs, this is like axiomatic, I think that the quality of your mind and your thoughts is one of the key inputs there is the quality of your nutrition, the quality of your exercise, the quality of your posture, the quality of your meditation.

whatever it is that you do to accomplish that state. So I always remind myself, it's hard to balance all of what I have to do and go to the gym for an hour. But I remind myself, I'm just so motivated by, when I was a developer, I was motivated by solving problems with code really quickly, just thinking well. And today I don't code nearly as much, I code maybe during a weekend.

But I am motivated by like, how quickly can I attain the right thought, the right result, make the right decision. So yeah, I love ironing that out and literally ironing that out. I recommend it to everybody.

Auren Hoffman (45:20.048)

Is there anything you do that's non-obvious on the health side?

rauchg (45:26.142)

I do. Yeah, I've painfully learned that most people overdo their fitness journeys. And I'm like, victim number one, like we could be in a support group and like I'm Guillermo. Not even the injury, they just like they go overkill for like a week. And then or maybe even if they keep it up for longer, it's just so intense.

Auren Hoffman (45:38.652)

I mean, get some injury or something like that or.

rauchg (45:55.702)

that ends up being detrimental. I lucked into this Instagram reel from a actual running coach that was basically sharing that the way that he trains his best athletes is by encouraging to make most of the runs easy. Route G, fascinated with speed and performance, was like, I have to make every single workout the last workout of my life and go all out and like...

Auren Hoffman (46:23.501)

Ah, okay.

rauchg (46:25.19)

every mile has to be the fastest mile I can do. So I switched my mindset.

Auren Hoffman (46:30.916)

So you're taking like interval training to kind of the extreme.

rauchg (46:35.134)

I switched my mindset to like, actually, I'm going to make most of my fitness not easy in the mental sense because actually the bigger challenge that I find that I think dissuades most people from continuing is that it's the confrontation of like being at a place where you're challenging your body for a long amount of time is a mental challenge that I think is even bigger than the physical challenge. So I think what ends up happening is that

because you overload the physical challenge, you are now, like you have, worst of both worlds, you have the mental challenge which we all have, which is a, you know, you're fighting against inertia and gravity. And then you add the, every workout is an extreme thing and you burn the candle on both ends. And you think, so my overarching advice would be just optimize for longevity, like,

I'm going to be working out every single day of my life or the vast majority of them, hopefully, until I die. So work backwards from there. What do those workouts look like for me to be 100 years old and reflect and like, oh, look, you know how Apple plots all the little circles? Imagine just spanning through your life and seeing, oh, look, mostly they're complete circles. Of course, here I took a day off. Here I had a weekend.

Maybe here I had a trip, but I scanned through my life. So when you work backwards from that, you realize you can't just be doing freaking like 300, this is asparto, whatever, every single day of your life.

Auren Hoffman (48:15.48)

Now, two more questions. What is a conspiracy theory that you believe?

rauchg (48:23.582)

I believe that we... There's so many thoughts here. So one thing on conspiracy theories in general is like, I get taken aback a little when people get really annoyed or infuriated by the idea of people questioning things. I think, you know, to each their own. If there are people that want to like speculate on what aliens are doing on YouTube.

power to them. They can have their own alien conversation. You know, like, so I think first of all, I would say like, I'm a huge fan of exploring the boundary of mind of what's possible and empowering people to think different. This is kind of like Apple's motto back in the day. And a lot of great ideas were first met with a lot of resistance and questioning when they came and I'm talking about history.

Auren Hoffman (49:21.724)

like Galileo or something like that.

rauchg (49:23.702)

Yeah, it's crazy, right? Like, it's, I mean, a lot of those things literally sound crazy the first time you hear them. I think Darwin, like, I don't know exactly what the exact reaction was. I should look it up because I'm kind of fascinated by that whole story. But I can't fathom a world where the idea of we evolved from what is considered to be this lower animal.

To not be seen as like well that has to be a conspiracy theory. That's nonsense a lot of things that turn out to be true have very negative reaction, so Now onto what I consider is theory that I believe

rauchg (50:11.562)

This is more on the things that people believe when they talk about conspiracy theories. I do believe that we are not, especially in the United States, we've over-indexed on what pill I can take, what magic solution. This goes back into the longevity thing that I was just talking about. What can I buy off the shelf to make my life better? Is there a magic cure?

to everything I have, every element. And I believe that I really believe in setting yourself up for success over the longterm. And what I found just through my own history is, again, like there's just, there's gotta be some pain. There's gotta be some sacrifice. And I think, not to say that this is a conspiracy theory, but like, I do think that...

as a society with overindex on those magical solutions. And then sometimes when folks bring up that they're not fully on board with that, we're like, well, I don't agree with you. That's a conspiracy theory. And that might mean that sometimes you talk to people that believe in things that are just lived experiences. I believe in meditation, for example.

And I don't know if I can point to a specific paper for why I would recommend it for your life. And again, there's a lot of sharp corners around all these topics in conversations, but I encourage everyone to have an open mind.

Auren Hoffman (51:49.36)

Alright, last question we ask all of our guests, what conventional wisdom or advice do you think is generally bad advice?

rauchg (51:57.482)

Um...

rauchg (52:23.948)

I'm going to say something from our own personal experience.

I think that a lot of folks give product people the advice that they should validate everything with data and that they should run a lot of user research and that there's a justification for every action and everything that you put into your product. Goes back to your A-B test question.

rauchg (52:59.054)

products because there is an abundance of belief and identification with the creators behind the tool. I see this a lot with Apple where it's almost like a reality distortion field even in the way that people perceive the products. People used to say like Steve Jobs is a reality distortion field around their teammates. I see a reality distortion field sometimes with people.

where you believe so much in the promise of the brand and the journey that they've taken their customers on for a long period of time, that you're actually able to gloss over what are objectively really bad things. So I think companies then tend to underestimate the power of building that high value brand connection with their customers. So...

Maybe the advice would be that I hear like, well, don't focus on your brand. It's too early. Ship some like janky page with plain text. Like iterate your way to product market fit. But what if your product market fit is predicated on that connection with the customer and the thing about it from this point of view, right? Like you are assembling, I really believe in this, you are assembling an airplane after you jumped off the cliff.

Auren Hoffman (54:04.505)

Yeah.

Auren Hoffman (54:07.856)

Yeah.

rauchg (54:27.842)

There's just no way that you cannot because I'm a guy from Argentina with zero connections in Silicon Valley. I am 19, 20 years old and I have nothing. I have just knowledge and open source projects and a community of people that I knew as engineers and so on. How can you actually possibly build a company with, even if you raise a seat around, like it's just...

It takes years. It's like such a catch 22. You have to get to something that's really good, that inspires people, that solves businesses problems. And that takes just an enormous amount of time. It almost seems like an impossible problem. I think that a way that we have inspired people to think about the long-term vision of the product has been in telling that story about what could be possible.

even if it's not there yet. But delivering, of course, and delivering quickly. And then going on to the next thing. So we invested a lot in the brand, we invested a lot in our connection with our community. We've invested a lot of things that are not measurable. Maybe a CFO would tell you they're horrible, horrible ideas. Because how do we measure that? And like, oh, it seems like we're just wasting money and whatnot.

Auren Hoffman (55:45.444)

Yeah, it's hard to know.

rauchg (55:52.37)

And yeah, there might be some casualties on the way in terms of like, oh, like, maybe I did some marketing thing that it turned out to be more expensive than I thought. And like, maybe they didn't even think about it as marketing. So that would be my advice is that there's a healthy in some ways, but there's a lot of skepticism about like, what are the actual attributes behind the success of companies? And I think some folks underestimate.

just how good the products are that you think are universally regarded as good and how much more is compensated through the storytelling.

Auren Hoffman (56:29.34)

You said something interesting, you moved to Silicon Valley, but that was pre-COVID. Do you think that Silicon Valley is both going to attract the best and brightest from around the world now? And do you think that it makes sense for the best and brightest to come to Silicon Valley?

rauchg (56:47.754)

Yeah, what I feel 100% bullish about is that the United States, and I've been to many countries, many lovely countries, the United States is the most welcoming country to immigrants that I've experienced. That is not just welcoming in the sense of like, here's a path to get a passport and a house. It's, well, here's a path, if you're willing to walk it, to actually create an incredible company.

to invent new things, to team up with other people, to say things, to get a platform to say things. Like, there's just no comparison. And I point this out because there has been a challenge to this narrative by a lot of different people. And I always come back to like, I think it was Warren Buffett, I don't want to pull the classic Woody Allen quote. I think it was Warren Buffett that said, everyone's criticizing the US so strongly.

I invite them to relinquish the passport and I think there's all these countries that are having to trade their passports and like boom, here you go. So what I'm in and I'm kind of like, I would say I'm almost like fanatical about this because again I've been to so many countries in like the US is definitely really special. Now I'm also kind of like that about San Francisco just because I love the environment. There's just so many good properties about SF.

and Silicon Valley, part of it is rational, part of it is emotional, to be honest. I just love how symbolic a lot of the inventions that have happened. Take one of the best typefaces that has come out in the last couple of years called San Francisco, Apple put it out. There's just so much that inspires someone to build and create there. Now what I think will happen because...

I think this is all downstream from the US as a system. There will be other Silicon Valley in the US, and I think that there are a lot of countries that will also follow the steps of what they're learning from watching this grand experiment happen, which is, as a reminder, all of these things have happened in a very short amount of time. What does that mean? You can't take any of this for granted. In second, there will be countries paying attention and copying and exceeding some of these properties in the global...

rauchg (59:11.574)

competition for talent and for the innovations of tomorrow.

Auren Hoffman (59:16.496)

This has been great. Thank you, Guillermo, for joining us on World of Dazs. I follow you on Twitter at RAUCHG on Twitter, and I definitely encourage our listeners to engage with you there. This has been a lot of fun.

rauchg (59:25.356)

Awesome.

rauchg (59:30.05)

Thank you, Aaron.

Auren Hoffman (59:31.672)

Alright, amazing, thank you, this has been great, it's perfect.

‍