Invest Like the Best with Patrick O'Shaughnessy - Howie Liu - Building Airtable - [Invest Like the Best, EP.375]

Episode Date: May 28, 2024

My guest today is Howie Liu. Howie is the co-founder and CEO of Airtable, a no-code app platform that allows teams to build on top of their shared data and create productive workflows. The business began in 2013 and now has use cases built out for over 300,000 organizations. As Airtable begins to integrate AI and the latest LLMs into its product, Howie has maintained a focus on an intuitive building experience, allowing anyone to build out their workflow within minutes or hours. We discuss the future of the platform in the era of AI, his perspective on horizontal versus vertical software solutions, and his crucial moments as a leader in building a critical component to the advancement of productivity. Please enjoy this discussion with Howie Liu.  Listen to Founders Podcast For the full show notes, transcript, and links to mentioned content, check out the episode page here. ----- This episode is brought to you by Tegus, where we're changing the game in investment research. Step away from outdated, inefficient methods and into the future with our platform, proudly hosting over 100,000 transcripts – with over 25,000 transcripts added just this year alone. Our platform grows eight times faster and adds twice as much monthly content as our competitors, putting us at the forefront of the industry. Plus, with 75% of private market transcripts available exclusively on Tegus, we offer insights you simply can't find elsewhere. See the difference a vast, quality-driven transcript library makes. Unlock your free trial at tegus.com/patrick. ----- Invest Like the Best is a property of Colossus, LLC. For more episodes of Invest Like the Best, visit joincolossus.com/episodes.  Past guests include Tobi Lutke, Kevin Systrom, Mike Krieger, John Collison, Kat Cole, Marc Andreessen, Matthew Ball, Bill Gurley, Anu Hariharan, Ben Thompson, and many more. Stay up to date on all our podcasts by signing up to Colossus Weekly, our quick dive every Sunday highlighting the top business and investing concepts from our podcasts and the best of what we read that week. Sign up here. Follow us on Twitter: @patrick_oshag | @JoinColossus Editing and post-production work for this episode was provided by The Podcast Consultant (https://thepodcastconsultant.com). Show Notes: (00:00:00) Welcome to Invest Like the Best (00:06:49) Exploring Horizontal vs. Vertical Software in the AI Era (00:11:00) The Future of Customized Applications (00:15:28) Perspectives on AI's Future and Enterprise Adoption (00:18:13) The Evolution of LLMs and Their Impact on Software Development (00:23:33) Harnessing AI for Business Transformation and Innovation (00:27:28) Reflecting on Airtable's Founding and Evolution (00:33:23) Airtable's Approach to Customer Engagement and Innovation (00:39:59) The Impact of AI on Platform Versatility and Market Penetration (00:46:00) Achieving Product-Market Fit and Initial Monetization (00:50:23) Scaling Up and Securing the First Unicorn Round (00:51:52) Rapid Growth and Organizational Scaling Challenges (00:55:00) Reflecting on Tough Decisions in the Business (01:02:55) The Role of Capital Allocation in Expanding Airtable (01:06:55) The Kindest Thing Anyone Has Ever Done For Howie

Hello and welcome, everyone. I'm Patrick O'Shaughnessy, and this is Invest Like the Best. This show is an open-ended exploration of markets, ideas, stories, and strategies that will help you better invest both your time and your money. Invest Like the Best is part of the Colossus family of podcasts, and you can access all our podcasts, including edited transcripts, show notes, and other resources to keep learning at join colossus.com. Patrick O'Shaughnessy is the CEO of Positive Sum. All opinions expressed by Patrick and podcast guests are solely their own opinions and do not reflect the opinion of positive sum. This podcast is for informational purposes only and should not be relied upon as a basis for investment decisions.

Clients of positive sum may maintain positions in the securities discussed in this podcast. To learn more, visit psum.vc. My guest today is Howie Lou. Howie is the co-founder and CEO of Airtable, a no-code app platform that allows teams to build on top of their shared data and create productive workflows. The business began in 2013 and now has used cases built out for over 300,000 organizations. As Airtable begins to integrate AI and the latest LLMs into its product, Howie has maintained a focus on an intuitive building experience, allowing anyone to build out their workflow within minutes or hours. We discussed the future of the platform

in the era of AI, perspective on horizontal versus vertical software solutions, and his crucial moments as a leader. Please enjoy this discussion with Howie Lou. You have this really interesting perspective in that you're an established large company, but you're not some massive fang stock. So you can move very fast still. But you also have a good amount of distribution and a brand and a workflow tool that's been used for a long time. How are you approaching both, I guess, the risk and the opportunity of this insane new technology into an established startup that when you started it, the stuff didn't exist. You couldn't use it. Now you can. What is your strategy in thinking for using this rather than getting killed by it? Yeah. Well, I think you're exactly right. We have to be as scrappy as the smallest startups.

If you're a big, big, big company like the fangs. If you're Microsoft, you make these multi-year bets on this. You can afford to be a few years late. But you're going to have such a distribution advantage by the end of it that even if you build not that great of a product and it's a little late, you'll still be able to win a lot of market share. I think for us, we're somewhere in between. And I think we do have an existing distribution base. I would say that's a big advantage. We have more than half the Fortune 500 as paying customers of some kind. We have around 100,000 total paid customers when you include the

SMB self-serve stuff and probably hundreds of thousands more organizations that use their table for free. So that's like pretty sizable distribution base versus a net new startup coming into the space, I would also say for us, the advantage of having actually a platform. So we have a product that basically is an app platform. The whole point about Airtable is you put your data in there, it becomes a system of record, we can integrate with existing systems of record, you can build a workflow layer on top of that. And it turns out, those are all the ingredients that I think are completely necessary to actually

unlock value out of these LLMs. I think it's really challenging to build a pure, for instance, a gentic LLLM experience that just does a bunch of stuff behind the scenes, with. without human input, without the ability to present data or to take well-modeled input data into it. And so in many ways, not just the distribution, but the platform we have itself was almost made to be able to now plug in all the awesome off-the-shelf and constantly improving models. And so to us, it's less about having to go and reinvent ourselves. I would say it's more about taking the platform and distribution we have and then getting really scrappy with going in and spending time with customers, figuring out what are the actual killer use cases

and making sure that we don't just give them a 5%, 10%, or even 80% complete solution. We actually give them all the building blocks they need to completely customize the use case they have. That's where I think it's not just what we have. It's also our posture or attitude and really getting in there and getting our hands messy with the details.

That's a lot of what I'm spending my time on now. So I'd love to talk about the age-old comparison in between horizontal and vertical software. Air table's horizontal in some sense in that people in lots of industries can use it to do lots of different things. And typically the horizontal players, Microsoft or something, won't go into very specific vertical niches because it just doesn't make sense to. How do you see that playing out in this new AI world?

How far down the specificity chain will a Microsoft go? And where does that get handed off to more focused players so that they can actually compete against someone with huge resources and CAPEX ability and distribution and all these cool advantages that the big tech companies have. So I would also describe it as three different buckets. You have these really broad but shallow product experiences, let's say the office suite. Extremely potent. Everybody uses Word, Excel, PowerPoint, and now Teams.

But individually, each of those products is fairly shallow, as in you're not going to run a really deep end-to-end process on Microsoft work. You're going to use it all the time for lots of different. word processing use cases, but it's shallow and broad. And then on the other extreme, you have these very verticalized solutions. When you obviously have ProCore in construction, you've got Viva in health, you can go and build a very specialized vertical solution, although they don't have to be deep. I mean, you have new companies like a dandy that does dental CRM, probably a little shallower

than to Viva, but it's very specialized. And I think the third bucket, which is the one we play in, is actually both horizontally applicable, but also pretty deep. And the way we do that is because we have a platform that is not just a productivity tool that is shallowly spread across lots of different users and industries, but we give people building blocks to create their own applications, and they're able to take those Lego pieces and build very, very specific use cases to them. So that's why we have companies like Netflix early on building these content production systems that are actually quite sophisticated. we didn't go and show them,

okay, here's how to engineer a content production system on our table, but we gave them really great building blocks that they were able to use and take their proprietary know-how of their process, their industry, their use case, and build that into the platform. So I do think there's that third bucket, and I think it plays really, really well into this current era of LOMs. And the reason for that is these LOMs are both broadly capable, but also deeply capable, meaning when I think back,

I was playing around with neural nets back in college. This was 05 through 09. Fiddled around with the Netflix data science prize at one point. I tried these very simple two-layer feed-forward neural nets. You could do very simple classification things with them. But they were pretty narrowly applicable. But I think what's really caught the imagination of the public and every enterprise now is that the latest generation of LLMs and obviously chat GPT was a big breakthrough in

recognition at least of these models' capabilities, if not actually a breakthrough in the capabilities themselves. I think what's happened is you have models that are actually capable of there's no other word to put it as but reasoning. It can actually reason about fairly advanced topics and it starts to feel more human-like and that humans are very broadly intelligent and we can solve a very wide variety of problems. And so the fact that you can take all these pieces of research about a customer and actually not just summarize them, but tell me what are the through lines that apply to how Airtable is going to sell to them. I think that's really profound.

And so having a platform that doesn't just wrap around the LLM and use it for one narrow application, that would be the equivalent of taking the beautiful human intelligence of a person and then telling them to just perform one road task every single day, although that kind of does sound like the office. But, you know, I think to really take advantage of just how broad and deeply capable these LOMs are, I think you want a platform that enables the customer, every customer, every line of business, every person within a company to design exactly how they want to exploit these LOM's capabilities in the context of their own work, not just as a disjointed,

shallow chat experience. What do you think about this vision? I've heard a few times that the future of applications is entirely customized and personalized, that because tools will be available to build software exactly tailored around one person's workflow for whatever thing, that everyone will just have. their own software and applications as a business is a bad idea because if I want a CRM that does very precisely one thing, I don't need to work with Salesforce or affinity in our case or whatever. I'll just get the exact thing that I want because software development will be so flexible

and so cheap. Do you think that that's like overly utopian? Definitely true to a great degree. And Salesforce is the case in point. So I think one of the eye-opening things about working at Salesforce. So I founded a previous company, went through YC, it was called ETACs, personal CRM product, and long and short, after about a year, ended up getting aqua-hired by Salesforce. And I specifically wanted to go there

because I thought it would be a really great experience, learning about enterprise software, B-to-B, just stuff that's hard to intuit from the outside, unlike consumer. And the big unlock for me, I mean, literally it felt like a light bulb moment, was realizing that Salesforce wins and has gotten to, I mean, now it's tens of billions, but at the time even, it was in the billions of revenue, was not because they went and built the best possible CRM,

all hard-coded into a prepackaged app, but rather that they built a platform. And with that platform, I mean, they had literally a metadata-driven application platform where every customer can customize the data schema. So you can add fields to your contacts, to your accounts, but also add new object types if you want. You can link between them.

You can create logic on top of that. You can create a custom page layout and interface on top of that. And so ultimately, Salesforce is a big platform that won against the predecessor, Siebel, or even positioned differently from SAP, because the prior art was, let's just build a fairly hard-coded solution and try to presumptuously guess, here's everything that you need for the best-in-class CRM or ERP. And Salesforce said, no, we're going to give you a template, a starting point that codifies a lot of the best practices and a lot of the right shape of this data model.

in process, but ultimately every customer is going to customize their own implementation. And so this idea of one size does not fit all. And in fact, every customer should be able to customize their own software has been true. And I think Salesforce has been proving that out. And I would generally say that most successful platform companies eventually get to that level of platform customizability, whether you are, I mean, Atlassian as well. Jira is actually also a very flexible, very customizable platform. Now, both Jira and Salesforce may be fairly hard. for the average layperson to customize, that's where we come in. But I do think we've been moving along this trend.

And I think that AI will for sure accelerate it, both because A, the way that you build apps can be accelerated with AI. So being able to come into Airtable and say, hey, here's my exact use case. Here's where I work. Have it go and build a use case that's actually based on both external understanding of that customer. So if I worked at, let's say, Nike, there's a lot already on the public web about Nike and how their business works and what they care about from 10Ks.

That's already, first of all, pre-trained into the LLMs. And second of all, can be prompted into the result. But you can do a lot to infer exactly how any given person within Nike wants to build out their use case and do it for them. That's category A. And then I would say category B that's really exciting is using AI in the actual workflow. So not only can you use it to build the app in the first place, but as you use apps, you can build AI into the app so that they're AI enabled apps, not just AI built apps. That's where the

broad and deep capability of LLMs really will come into play because I think you've got to be really imaginative to find every single place in a business process where you can introduce an LLM step, an LLM capability to really automate something that would have taken a human hours and maybe even deliver a better result. If you're coming to. coming up with like a new marketing campaign, maybe the AI can actually a certain step give you inspiration of the campaign strategy or even content. Here's a storyboard or like a concept art to represent the campaign that actually makes the human then go and take it to a better result.

Just to ground us, if you had to rate yourself one through 10, 10 being your rate Kurtz file and you could not be more bullish about the future of this stuff and one being a total skeptical or something that just thinks chat dbt is an idiot that hallucinates or something, where would you place yourself? So it depends on time horizons. On a 10-year time horizon, I would say vary on the extreme end of bullish. I think these AI capabilities are only getting better. How much better can they get?

Do we exhaust ourselves of real data to train them on? How well does synthetic data work, et cetera? But just barring that for a second, I think you could take even today's LLM capabilities, ask every AI researcher, every company developing new better models to stop and just freeze time and say these models, what we have today is what we have. GPD4, we got Cloud 3, we've got other models, that's it. I still think there are many trillions of dollars of GDP that can be created by figuring out how to better leverage those models in the context of real business use cases or even personal ones. First of all, very, very bullish. that being said, I think that actually one of the fundamental bottlenecks is going to be

behavioral change and the willingness of enterprises. I think S&B and consumer will actually be faster to adopt this stuff. And in fact, you're already seeing that rapidity of adoption with chat TBT being mostly consumer driven because you don't have to worry about all the safety issues and copyright issues. And frankly, an individual can just start experimenting with this stuff or even a company like Hagen, getting a lot of growth from small businesses that just go out and figure out how to use this thing on their own, I think Enterprise really, it's going to be a question of, how are we going to get these use cases actually implemented

in both a way that's very sanctioned, but also maximally useful. And there's just a lengthier, I think, both behavioral change and also process implementation timeline. So I don't think a year from now, everything is going to change everywhere. I think that it's going to be a very methodical approach to actually getting value out of these models. I was talking to Chathen from Benchmark last week about this, who's always someone I go to trying to figure out what's going on at the frontier. And one of the points he made was, look, it's just going to take some time for people developing applications to get used to a probabilistic

way of programming versus just purely deterministic. We'll get there. We'll figure out how to work with these things, but that is a feature of these LLMs that we just have to respect. How have you respected that so far? So maybe you could give us an example of an AI enabled warflow not just built, but enabled workflow in Airtable that you feel like is a good representative example of the power of these things. Well, so there's two lenses I have on that point, which I agree with completely. And I remember that was what tickled me with AI and like neural nets is the ultimate meta approach to software development is instead of having to go and literally code out to the specific

if this than that statement, what you wanted a program to do, just train it to do it. And I think what's really cool is these LLMs have been merged with emergent capabilities that maybe even the model researchers didn't fully anticipate with every successive generation of these models. But to your point, that also means that the frontier of even what the current models are capable of today and certainly what they're capable of in every single specific permutation of industry, function, use case, even at the team and individual level, ways in which they can be applied is still a relatively undiscovered frontier.

And so the first lens I look at this through is the no-code lens, which is the entire premise of our app platform, is we give anyone non-technical the ability to go and experiment with these Lego pieces. AirTable is really intuitive to use. We have this amazing community of builders who just go out and build their own workflows, build their own apps. And they do so in a very play-like way. It's safe. It's easy. It's extremely fast. You can go and build an app in minutes or hours with AirTable. And so now by introducing AI and our whole premise here is we're not trying to go and be the leading foundation model developer. In fact, we want to just ride the wave of all the best models that are out there, whether they're open source like Lama 3, whether it's Anthropics or Open AIs models, we're going to plug those in and basically use those to be able to actually empower our customers to build anything they want with them.

And so we get to have this petri dish, a very large petri dish of our customer base going out and experimenting with how they can use LLMs in context of all of their very, very specific use cases. I mean, we have everyone from literal cattle farmers managing cattle in our table to small nonprofits or legal firms, etc. But all the way up to really, really large enterprises doing very transformative work across an end-to-end process. And so we have seen some really interesting use cases there. I think everyone could get their arms around some of these primitive functions, collection, search, summary, synthesis. These are tasks that all humans do and very content specific. What do you think is the next layer of functions that either we begin to explore or get enabled by successive versions, GPT5, Cloud 4 or whatever? So first off, I think even just those functions,

form the building blocks of a lot of very advanced work. And the main difference between what a human can do today and what these LLMs can do is I think we can chain those functions together into a sequence and not get too drifted away from the original intent of the project. I think we all probably saw AutoGPT come out a year ago. Yeah, a year ago now. It created a lot of attention because in theory, it looked like how a human goes about doing work. Let's say the prompt is, go help me build a business that generates a thousand

bucks of profit per day. And so it literally will go like a five-year-old and Google like, okay, well, how do I start a business? How's profit calculated? Maybe all these different underlying steps. And it's funny because you watch it and it just progressively drifts more and more off course, which is why those completely automated LLM core agents were and still are pretty ineffective and actually solving real world problems. I think they drift too much. I think today a human, though, follows not that dissimilar a process for a very, very advanced research project. Let's say I want to do a big research project on some concept.

Blockchain. I'm coming into this 10 years late, really, but I want to catch up and do all my homework on blockchain. And, okay, I would break that down into first a few preliminary questions about blockchain, and then I would go and search for articles about it, read through those articles, synthesize them, combine those thoughts, come up with a thesis, maybe try to draft something, critique it, revise that. And so I think actually a lot of the capabilities of the LLM are there. I actually talked to somebody who is a senior product leader at one of these auto companies. And they were like,

I think we may look back in a few years and realize that we had all the pieces of a much more advanced form of intelligence with these LLM's day. We just didn't quite figure out how to chain them together to get the most value out of them. And so my view is, I don't think we need dramatically novel capabilities, although there are certainly some interesting ones, like true multimodal support where you can feed in a video or any kind of document and it's actually looking at those in native form, the LLM, and able to like answer questions based on those and to do so with larger and larger context windows at a time. I think there are going to be major advances. But I think the most special advancement, in my opinion, is going to be the

quality of the reasoning. And there's a difference between a five-year-old going and trying to answer that question of how do you create a $1,000 business or how do you do a big, grandiose research project on blockchain versus somebody who is a grad level student or even better than that, an expert at the topic or at least an expert in research. How do you frame this to your team? What speeches are you giving them about your responsibilities as a company for not getting left behind by something like this? Yeah.

So it's been a really important part of our internal strategy and also. the drumbeat of what we pay attention to. I mean, we have a number of different mechanisms within a company to elevate and highlight certain things across the company more than others. If we're going to talk about exciting parts of our product roadmap, it's not the only important thing. I actually think a lot of our other platform improvements, vastly improving the scale capability of Airtable to support tens of millions of records of data or even hundreds of millions of records of data is actually going to be a pretty foundational improvement that itself complements AI capabilities really well, because now we can allow you to do AI processing

or AI workflows on top of even larger core data sets. It could be from your ERP system, for instance. But I think the AI narrative has been really important because I really do believe that AI, and especially this generation of Gen AI that has these broad and deep capabilities to reason and perform human-like advanced work is a perfect pairing to, the platform that we've built with the data and the workflow modeling capabilities, because at least until we reach the true NSA, where you have AI that is AGI and can solve any problem completely autonomously, I think the human element of having an app where you can see the data that you're working with and see an interface that lets you perform a workflow,

it could be the review of an AI generated output for those marketing campaigns and saying, like, oh, let me edit this and tune it and then bring it to the next step or hand it off to the next person, that's still going to be a really, really important. In fact, probably the most important part of unlocking these LLMs values. So I think what we've really tried to emphasize is the important role that we and only we really can play with these LLMs because that's the big gap that I see in the market today is, again, these LLMs could already be exploited for a thousand or maybe 10,000 times the economic value impact that they are today. And the big gap is really around having that

data and workflows and human interface layer to them. You're a seller of software, but you're also a buyer of software for people that work at Airtable. What kinds of things are you currently most willing and excited to pay for that are AI first products? What I really want are products that are not just AI as a feature. So as an example, we turned on Slack's AI capabilities. It's fine. It's useful. You can summarize a threat, et cetera. Slack is a particularly rich data set.

It's a lot of also inefficient text. I mean, you have all these long back and forths in Slack that really could turn into much more compact synopsis. So it's actually a pretty good feature for Slack. But for me personally, that's a lot less strategically important to us as a company or interesting because what I care more about is what I would call process transformation. So I want to actually take the end to end of how we do something. It could be our own marketing operations.

It could be our product operations, which we actually run on air table and have introduced lots of different AI steps as part of that, including capturing customer feedback and synthesizing that into insights that then get triaged the appropriate part of the product roadmap, helping with the generation of product requirements, documents, et cetera, creating very beautifully summarized emails or weekly updates. of what's going on in product. But even if it's from a different vendor, what I care about is a company that's coming in

and actually thinking about the actual job to be done, the business process, the use case, and showing us how we can solve that problem rather than just give us a feature on top of their existing product, if that makes sense. Can we go all the way back to the founding of the business and talk about product philosophy? Because a lot of my questions so far,

and the ones I have remaining, are how we're going to use technology to build things that we couldn't before. And I'd love to hear what the founding product philosophy was. And then also how, if it has evolved, how it has evolved in the years since founding. Airtable is founded in 2012. So it's been a while. A lot of things about both the philosophy and also the practical execution path that have evolved. I think what's consistent is we founded with this view of most software.

can be platformized or can be reduced into these building blocks of basically data model, business logic, and then a interface or workflow layer. And in a way, it's like saying, well, all software can be built on a model controller in view, like NBC programming. So that's still true. And that was very much inspired by my own brief time at Salesforce where I got to see the power of their platform. And I think what's been a great learning opportunity for us is, well, that's good in theory.

But in practice, people sometimes want more than just a very, very abstract blank slate. That's not even literally a blank canvas where they can just type into a draw on it. If you give them very, very abstract primitives, building blocks, sometimes they get stuck and they don't know how quite to build the thing they want. And so I think that's a place where we've learned a lot around giving people templates, giving people an onboarding journey or how we engage with them if they're an enterprise, is very tailored around understanding their use case and then showing them rather than telling them how they can build towards their end state.

And then having this co-participant model of we may be giving them some predefined templates or we may be helping them if we're doing a more high touch engagement, helping them along the way, but also having very much the customer participate in designing and even implementing what they want. So I think the evolution there has been from pure Lego kit. We just give you a bunch of Legos. would be like if Lego the company only sold. I mean, I remember you could buy a big red bucket of all the parts. And it would be as if Lego never sold anything else.

And in truth, we're starting to go to customers and say, hey, actually, do you want the Millennium Falcon or this Castle kit? And we can help you build that. Or even here's a predefined template, a blueprint for how you would build that. And we are starting to do things around actually productizing solutions that like Salesforce and their solutions, sales, CRM, support CRM, etc. are actually somewhere in between a template and a full app. There's going to be a little bit more structured to it.

But at the same time, it's going to be built on the same underlying, completely flexible platform that is almost infinitely extensible. I think finding that blend of it's both platform and also application and solution, and the way that we engage with our customers is it has to be that flexible and adaptive as well, has been the probably biggest evolution since the founding. So if you were to take everything you've learned about building and selling software, pretend I've just forced you to be an investor from this point forward. You're not allowed to run the business anymore.

What would be your stack rank of variables or criteria that you'd be looking for in a marginal software business to get you excited to invest in it? So maybe going back to one of the earlier themes, I would definitely be a much bigger believer in generally horizontal but deep platforms as opposed to either vertical and narrow, or horizontal and shallow products. And to be frank, there's not that many of them out there. I think ultimately great businesses evolved there. I brought up Atlassian and Salesforce before.

I think those are both products. In fact, Salesforce started out as a not customizable product. It was a true product. And it was very similar to a lot of the SMBCRMs you see out there today popping up. But pretty early on, I think a few years after they're founding and getting some, in their case, PLG, I mean, they were literally one of the first PLG companies on the, web, they evolved to having a much more flexible platform foundation. So I think that would be a table stakes requirement from my standpoint, because I think it's just

so important in enterprise, especially, but probably also SMB, given just the diversity of use cases, industries, and company-specific variations in how these processes are run and how the data model even should be defined, that customizability that having a platform model at its core would be a requirement. And then the second would really be how deeply, you know, is AI integrated, not just as a single feature, but rather are they thinking about their product approach as almost becoming now a conduit for how do I make this product capable of deploying more and more AI into different parts of that workflow, that process? They have to be able to

introduce AI not just into a single place, but into as many parts of that process, or ideally give the customer the ability to implement AI into as many parts of that process as possible. Do you think that we'll see successful companies, let's just use the world of investors, because it's so familiar to me, where you create almost a platform-like approach, using the two criteria you just said. I approach this from the perspective of, I've got these primitive things that I know matter to investors. And then I've got these AI functions that can help me perform that array of platform-like

tasks. Yeah. Is that overly verticalized in your mind? Do you think that there's no point to do that because I'll be able to do that for every place with Airtable plus a couple other things? I'm just trying to figure out what you think about the future of vertical markets, applications and software when wide and deep becomes possible for the first time. Maybe I'll zoom out and share a framework that I've been using a lot as I go and both meet with our customers and think about how we can be useful to them. But also now I do think of myself as almost an investor in a way. What's exciting about my job in Airtable is I don't sell one thing, like a support automation

use case to customers again and again. I can go in and like an investor chasing Alpha in different industries, different companies, I have to learn what's going on in media, what's going on in retail, what's going on in FinServe, and figure out how that necessitates a certain process or operational change that then Airtable can be a fit for. That's been a really fun part of the job. But also to your question, I think I go into every customer and I start. start in that order. I first start with trying to understand what's motivating change. Because I think

the worst software businesses are changed for the sake of change. And it's like, oh, here's a shiny new object. It's pretty. People like it. But if there's no real strategic driver for implementing this new thing, I think usually it's not a very durable product. As soon as the hype subsides around it, it will go away. And so I think if you look back at digital transformation trends, there were a lot of dollars spent by a lot of very big companies on the vague promise of digital transformation, but not necessarily a very strategic thesis in mind. Here's a very specific application that is motivated by the changing physical supply chain. That's actually a good use case for digitization. But starting first with that perspective, and then turning that into an

understanding of, okay, what's the process change that needs to happen? So as an example, in media, obviously streaming has become the massive change driver across every media company. And every big existing media company is trying to launch their own owned streaming service for a number of reasons, including the fact that they don't get disintermediated and commoditized, but they actually have their own subscriber base. They also get direct data on those subscribers. And there's compounding advantages to having a large streaming service and trying to catch up against the headway that Netflix now has.

But that's an example of, wow, that requires a lot. of internal process change because if you think about the whole value chain of a large media company, the way that they produce content and then distributed it on linear TV is now very upended by the pace and the amount of content they need to put out on their own streaming product and not to mention building up the actual technical and product capabilities to deliver that new service. So herein is a lot of work that needs to be done to rethink operations. And I think that's where you see the greatest opportunity for a disruptor like ourselves, but it also could be a new startup. I think it's generally really hard to get companies to change for no good reason

when there's not a motivator. But when you have a motivator that's flowing from existential change or impact from the very top, the entire industry is being upended retail. The shift to digital commerce and Omni Channel and stores becoming more of an experiential place where you go to try products, but then maybe you transact online or vice versa, that's the big, big change motivator. And building the operations to keep up with that and support that, that is where these companies want to spend their time and energy and ultimately dollars. So whether you're a net new startup or air table, that's where we really want to focus, is the places where there was the great change and therefore we can capture the flow share

of new value created in the company and become stood up as the new operations platform. Do you see a concentration today of a couple use cases that are most dominant in Airtable? When you give someone a system that can do anything, it's fascinating to know the data of what people actually use to do with it. So how parado distributed is it? Is it super concentrated in a couple of things or is it really broad and fragmented? Yeah.

So what I've learned is it is concentrated, though not always by the dimensions that you would think. So the high level dimensions are industry. So we have a lot of concentration, especially in our large enterprise business, which now is far more than half of our total revenue and the biggest driver of growth for us. But that's really in media, in retail, in FinServe, and then in tech. And I think the reason for that is these are the industries that have undergone the most change. This is where I've derived my theory of where there's great change and existential

evolve or die type forces at play. Companies have to really rethink how they operate. And that's when Airtable is a much more modern, much faster way to stand up these 2.0 operations has a chance to come in and really win big versus trying to chip away at very small, very incremental team-level use cases within a company that just is not motivated to change how they operate.

The second area is where I think there's just a more dynamic nature to the work in general. So a lot of front office use cases where the actual process itself is quickly evolving. So it could be marketing operations, could be digital product operations, but also things like content production operations. Retail store openings has actually,

actually been a big one for us within retail companies that are now, especially post-COVID, going back into full, let's open lots of stores mode. And I think these are all very strategic use cases, but ones that require agility, like the way that you want to iterate on how you plan and manage store openings is quickly changing. And you need to move faster than ever. And then third, I think we actually have a psychographic dimension of fit for Airtable, which is I always tried to figure out why do some small businesses. fairly slow-moving industries like restaurants, for instance. Why do some restaurants go all in on AirTable and they're doing their employee staffing and inventory management? They're building very,

very advanced use cases or they're implementing on Airtable and a very advanced run their entire operations. AirTable is their OS way. And then obviously a lot of other ones are not because we're not saturated into the restaurant world. And what I realized, I know after spending a lot of time talking to these customers is I think there's just some, it could be a restaurant owner, it could be a lawyer who owns their own practice, especially in S&B. I think there are some people who just naturally have an innovative and operational mindset. And these are the people who might also be using other low-code products, Zapier for automations, and maybe they're building their own website on Webflow, et cetera. But they're using Airtable because they see it as this really

flexible platform and they know what they want as the N-C and they're able to use this product and get to that N-Sate. And they know that by customizing their operations, they can get an edge versus there are other people who don't want to innovate there. I mean, it's scary. And they just want to use something out of the box as simple as possible. And for them, maybe some verticalized operations platform for restaurants is enough. Or they'll just use it because it's less daunting than building something yourself. I do think to come full circle to AI, this is an area where AI and especially AI to help

gently bring along that person to show them how they can get very, very, very. very specific use cases relevant to them on Airtable without having to put in the imagination or effort required to literally build every part of an app from scratch is going to be a huge, huge unlock and may allow us, even on the low end of the market, to also in parallel, win the down market because I do think their AI is going to be the big tailwind for platform plays versus verticalized solutions. If you think about all the things that the next generation of these AI core foundational models could allow for that they don't yet, which set of new capabilities would most enhance what Airtable can do for its customers? I think it's going to come

down to higher quality outputs more consistently with easier prompt engineering. I briefly touched on earlier multimodal capabilities and what if you could feed in to one model, full video. I mean, there are specialized models for video processing right now, but what if you just had a single model where you could throw in video, long reports, documents, text, audio files, and output any of the above as well. And that is coming very, very soon. But to me, actually, that's a little less profound because you can piecemeal get some of those pieces of value today by using different specialized models and integration with the data and workflows you have in a platform like Air Table. But I think the reasoning and quality improvement,

is going to be profound because the way I think about it is think about when these LLMs were at 20% accuracy on certain human tests, let's say the SAT or the GMAT or the GRE, any of these tests. A 20% accurate LLM is not that useful. Think about a 20% or even if it's better than guessing, getting somebody in your company who can complete work assignments at a 20% accuracy rate or usually gets Fs on their papers, not that useful. But you get to 60% and you're like, okay, this is starting to become interesting. You get to 75, 80%, you're like, okay, this is like a decent, maybe average performer, right? B student is not so bad. And then finally you get to like A, A plus, like 95, 99. I mean, that's top of the class.

And so I think what a lot of people underappreciate is that even just doing better at the same domain of problems that we can solve today, but doing better and doing it more consistently well, And on the prompt engineering point, doing so without having to require fine-tuning or a lot of advanced prompting, because right now you have to be somewhat of a prompting expert to get the most performance out of these models. I think if the prompting gets easier where it will more consistently give you good outputs, even without you having to be overly clever or experimental with how you prompt it, will all of a sudden make it even more broadly useful. I think of it as if you take every person out there who could benefit from LLMs, which is probably everyone, or at least almost everyone who does any kind of knowledge work today. And if you gave them an experience today where for hours and hours, they have to play around with this stuff, experiment with it, to get any kind of value unlock. Of course, most people are not going to put in the time and energy to do it.

Now, you advance that and you make it so that in minutes or even seconds, like the first thing they try, they get this immediate magical value. And not just for novelty's sake, because I think we've all gotten that from chat to BT, like, write me a bedtime story based on our pet as the character. That's magical. But it's not very durable recurring value. I think the second you get a really useful, sticky use case, and if you can unlock that very quickly, I think we're just going to see a lot more real durable adoption of AI.

And so to me, that's the big unlock is quality, immediacy, and consistency of results. even for the same type of problems that you could already throw an LLM at today. I like Rulov from Sequo's idea that every one to three years, so you've had 12 years, so maybe you've had four of these or something. Companies go through these, he calls them crucible moments, areas of huge decisions, huge tradeoffs, huge risks or threats to the business. I'd love you to tell the air table story through the lens of the crucible moments. If you think back all the way to 2012 through to today, in the first 12 years, what have been the most important crunch time moments for the business? So the first three years were basically building

the product pre-launch. And I think we started around the same time as Figma, which is really interesting comparison because I think there was something happening in the world at that moment, both behavioral change on behalf of end users within companies, bringing in their own products to work like enabling more of a bottoms up adoption model, which really didn't exist five or 10 years before that. But then second of all, also browser capabilities. browser performance improving to the point where you could actually have a really rich desktop-grade experience in the web. A moment for us was we thought it was a good time to enter the market and to start working on this problem now. I mean, I do believe market timing has a lot to do with the success of an idea.

Instacart tried 15 years prior through Webvan, didn't work. It was probably the wrong time. But then obviously it did work when they came around. So for us, the crucible moment initially was just making sure that we got the market timing right and also that we got the product details right. because ultimately we had competition. I mean, there were other companies that were trying to do vaguely the same thing, and some that had raised more money than us earlier.

And yet, I think we just out-executed them on building a better product and actually shipping it. I mean, two and a half to three years is not fast by any means, but I think it was the fastest that we could ship a really good product that was thoughtfully architected, both back-end and in terms of the front-end product experience. Got some compounding growth. It wasn't run away at the slug. black level, but it was enough to show us there is real product market fit here. And importantly, there's durable value here. People are retaining. They're expanding their usage to other people

on their teams. And they're actually so happy with it. They will go out and tell other people within their company or across companies or even tweet about it in a way that drives this organic virality. So I would say that was one inflection moment. I would say the next moment was, okay, we built this product. We were pretty sure as a B2B product and it was solving real use cases. people were actually running their operations for either a group or an SMB on Airtable, but there's still this unknown of, well, people pay for this. And if so, how much? Because there's still a drop-off of you have a great product, but can you monetize it? Can you build a great business out of it? And so this next phase was about what is the business of Airtable? And we had

always advertised pricing plans. We didn't really try to monetize or convert people up to the paid plans until maybe a year in. And so once we started doing that, I think the next inflection was really seeing, wow, people actually do pay for this. Now, initially it was small teams here and there. But I remember the first time we got a $10,000 customer, which now seems small in hindsight.

But at the time, it was a big dollar amount, and we were both a smaller company, so it felt bigger, but also symbolically to me, and I grew up in a very modest family where $10,000 was a lot of money. I mean, it would be like literally winning the lottery. Wow, somebody is getting enough value out of this product that is pure software. I always still feel this immense sense of awe of software is so abstract and yet so powerful.

Somebody's paying for this service that lives in the cloud. They ship us their bits. We give them those bits, but wrapped in a really nice data storage layer and workflow layer, etc. And it's like valuable enough that they're willing to pay us real money. So I think as we started to see the on-ramp to monetization and, I think it probably took us three to six months to hit the first 500K, maybe only another few months to hit a million. So it was a really rapid acceleration. And then a million to 10 million was very fast.

I remember it was somewhere between a year and a year and a half. And so I think that next crucible moment was really, can we build a business? Can we make money? And I think we were very happy when we saw the answer is yes. And then probably the next phase was that led us to get to this point where, I mean, we were out. 10 million in revenue. We were scaling very, very rapidly. It was clear that we were going to hit 20 in less than another year from the 10. And so I think somewhere along that time, we started getting a lot of outreach from different investors. We engaged with some of them. We ended up

pricing around somewhere in between those milestones for our first unicorn round. That was Thrive, benchmark, Kochew, were the main investors. I think that was the moment. It felt like a phase shift in the company again because we went from being this somewhat indie darling that a lot of companies loved to use, it was almost like a cult phenomenon, to starting to feel like, oh, okay, we got to figure out how to make this mainstream. And that means taking our tiny team of maybe 30, 40 people, and starting to scale that up, becoming a real company and taking this business to the next level. And I would say that brings us to our next phase, which was 2020, I remember COVID hit.

And for a moment, everybody panicked. Everybody thought this would be the trigger of a downfall to the bull market. Initially, the first few months, it wasn't clear how the economy was going to react or be impacted by COVID and quarantines. And there were small businesses that were being really impacted. That's when the PPP program came out. God, I remember well. But we were worried because a lot of our customers were SMBs.

We also had some enterprise ones. And we didn't know how they were going to be impacted. And in fact, we did see some churn from those S&B customers who literally were scared. I think actually more often than not, we would ask people why they were downgrading, and it was we're shutting down our own business or we're laying off our employees. We don't need your product because I'm firing my 10 employees and now I'm one person. And it was obviously very sad to see so many of these small companies get impacted. From a business standpoint, we were also just very apprehensive of what's going to happen to us.

We ended up raising another round in the midst of that just to give ourselves a little bit more certainty, stability, and we wanted to always be able to be very long-term oriented. So thrive, led that round. And then obviously the world fully came back, where at least the economy did, surprisingly. And there was actually yet another strong bull run. In fact, what we saw then was an inflection in growth as it's almost like every company then wanted to invest more into especially collaboration software that could help them digitize and hybridize their work. Every company had to become a hybrid or fully remote company overnight. And so they started exploring initially the basics, Zoom for video conferencing, Slack for chat. But then I think the next wave was really

Airtable as a one of a genre of products that enabled more structured collaboration. You need some kind of structure to how you collaborate. And it can't just be emailing back and forth or docs or Excel. I think that next moment of what's going to happen to us and then pricing the round was about, okay, how can we now adapt to this new, almost like demand pull forward environment for us? How do we basically really scale up to keep up with this and play into it? And I think we did a pretty good job of keeping up with that demand. So for the most part, we were able to keep up with both the physical scaling demands of our servers, so a lot more growth of top of funnel, just overall growth of usage. And then scaling up our go-to-market motion, be able to hire a lot more

customer-facing people in addition to our R&D to engage with those customers, I would say the thing that I'm most neglected in that moment and what leads us to the next final crucible moment and the current phase that we're in was how do we build the right foundations so that when this demand pool forward environment subsides, we have both a repeatable growth model, but also organizationally, and it turns out it's really hard to scale organizations and management structure and company context super rapidly. And during that stretch, we grew headcount and more than 100% per year for multiple years. So we went from low 100-ish employees to over 1,000 in just the span of a few years. And I think we hired a lot of individually very talented people

during that time. And again, during the growth years, we were keeping up with all this amazing customer demand. We were seeing very high growth as a business ourselves. We had like triple-digit growth years and the enterprise business was really flying at that time. And then we get to the last phase, which is we obviously saw the economy shift over, or at least interest rates really shifted from ZERP to not ZERP. The actual customer demand, we felt shift the last calendar year. So it was a little bit lagging versus the equity markets. But in that new phase, customers went from a default spend and then ask questions later about why am I using this? How is it differentiated? What's the business value to ask those questions first? And in fact, like, re-evaluate every

seat, every vendor that they have in place. We weren't the only ones that this happened to. Even hyperscalers have the year of compute optimization. These companies were all going and saying, hey, do we actually need to use all the capacity, all the compute, all the seats that we're paying for, or can we actually optimize some of that? And fortunately for us, we were embedded into sticking enough use cases. I mean, these are like really important processes, high value data sets that we saw very little logo churn during that period, but we did see we ate some contractions in terms of just the seats and the revenue from some of these customers. We still ended up clearing the year with a pretty healthy growth rate, 40 plus percent revenue growth rate for the overall

year. But it was definitely a big mood shift for us as a business. And I think, To me, it was the moment where we had to really figure out what is a repeatable go-to-market engine that aligns with our platform or product vision. Now we go out and have a much clearer point of view, whether it's what are the industries and use cases that we have a story around how we provide value and or also just how we figure out what are the right product capabilities to build. How do we deliver those to the market? We don't just wait for the customer to come to us. We need to go, and whether it's a high transactional velocity, efficient mid-market business, or on the strategic side of things, really going out and spending time with our biggest customers and figuring out how their business works and being able to come up, show up to them as almost like a strategic consultant and understand their business, understand a really large media company.

What's going on there? And you're spending $5 million plus with us. We better understand your business and where we can go from here if we want to become an even deeper and more expansively deployed platform. So I think that has ultimately resulted in a lot of progressive foundation building over the past year. In that entire story, what do you think was the hardest decision or biggest tradeoff that you made intentionally? This may be recency biased, but we made a tough decision to actually do two riffs. And now we're done with them because we're cash flow positive and growing. And actually now we're really investing into the business in the right places to execute.

on the AI product roadmap and then also how to bring that to market. But I think in the moment, we had this dilemma where we had a lot of capital still on the balance sheet and still do. In fact, now it's growing because we're cash flow positive. We had almost a billion on the balance sheet. And so theoretically, we could have just never made any changes. We could have kept going as a cash flow negative business for a while and then eventually evened out.

Instead, we made the very, very difficult decision to do these reductions in force and to actually significantly downsize the company. And I did so because, A, I wanted to be able to continue hiring throughout, meaning I think philosophically, it's really important for companies to always be hiring because it would be arrogant to assume that all the skill sets you ever need to execute, even as the market changes, the strategy evolves and add new skill sets, like if you believe that you were immune to ever needing to hire new people. And then second of all, I think, It also is a really useful mechanism to have some amount of cross-pollination of external other companies' perspectives, external market perspectives.

It's almost a little bit like doing a funding round for companies that aren't public and aren't constantly fundraising. You only get these few and far between moments where you really get to- How we doing? Yeah, check out and see how you're perceived and test your story, test your business quality. And in a way, I think keeping that recruiting flywheel going is similar to that in that you're constantly checking, hey, how are we showing up as an employer brand, but also can we constantly get some fresh perspective into the business? So by doing those riffs, we were able to continue hiring versus if we didn't then, we'd already be overscale on cost relative to where we wanted to be. And then we would have had to make a hiring freeze decision. I think the second reason for it was we ultimately thought that we actually could move faster in certain ways with a smaller organization. This is obviously something that now has been talked about by lots of other companies,

but it's obviously really, really tough when you're talking about people and jobs. But as a company, and if you did step away and make a very objective decision for the company, I came to a very strong belief that in this phase, we actually need to execute in a much more agile way. And agile meaning we got to, like you said at the very beginning, act more like you said, act more like a startup when it comes to executing on AI. And I don't think this is a problem that throwing thousands of people at it will solve faster. I actually think you need smaller, tighter knit teams that work more directly with customers to really understand what are the use cases, what are the jobs to be done, and figure out how to actually make it real very, very

quickly. So iteration cycles need to be very, very low. And I generally think that happens better with smaller tight-knit teams. I mean, there's the famous Amazon two pizza box rule. I think there's a reason why a lot of the best software or even hardware products like the Apple Macintosh were developed by small, almost renegade teams. I mean, Mac was made by 50 people when Apple, the massive behemoth, put thousands of people on, I think it was the Lisa computer, which ended up not being nearly as innovative or as good of a computer. It's a fascinating phenomenon. If we were doing a big draft of companies and the idea behind the draft is the companies that are going to most matter in the future as you see it have the biggest impact. matter in people's lives. Where do you think your draft order

would be the most out of sync with the norm? What company or companies would you draft the most high relative to what you would expect others to do? I think some of my top picks would include the big hyperscalers, although Microsoft and Google actually get a very top pick in my mind because of the role they're playing in AI specifically. I think that there's actually going to be a lot of these low draft picks

seeming upstarts. And for instance, if you're thinking about an industry like media, when you have that much upending of the value chain of how content can be generated, I mean, let's say in the very, very near future, you can replace a lot of CGI with models like SORA from opening eye. But in the not too distant future, you can actually stitch together enough of these scenes and actually compose entire end-to-end movies. In fact, there's already pioneering filmmakers that are doing this.

I actually think what's going to be really interesting is that it's not just the pure tech startups that are creating the technology to do that and sometimes selling them to the big incumbents to be able to innovate internally and evolve so they don't die. But I also think you're going to see, and we haven't yet really seen a lot of these, but I think it's going to occupy a really big chunk of the draft roster. So this is more of a hold the space for this new set of companies where you're going to see really aggressive industry operators. So imagine in media, it's not going to be like a pure tech founder who only knows the technology. It's going to be some person who comes in and is willing to

be aggressive, entirely foundationally oriented around how can we get a lot more leverage in our efficiency and how can we produce better, faster, cheaper content. We're still primarily about the quality of the content. And it's like the Pixar. Pixar was a fusion of both technology and filmmaking and storytelling, and you couldn't have had just the technologist. Pixar without the storytelling, without the ingredients of still making a beautiful movie, would just be a cute technology play. But I think it was really in the combination of those pieces. Same with ILM and Lucasfilm.

There's going to be players who come in and really figure out how to innovate with the technology. And I think there's going to be some of those in every industry. And I'm really excited to see it's going to be a race. for a while, everybody thought the new D2C brands were going to be the thing that took down the big existing retail companies. I think the retail companies fought back well on DTC ended up not being quite as big and as rapid a disruption to the core business of these large retailers as we thought. But I do think there's going to be AI first industry-specific businesses that probably get founded by a combination of a technologist or technologists. And somebody really understands the industry and is ready to be a very, a very, very, very aggressive executor on building the next X of each of those industries.

So they don't exist yet, but I think there's a lagging nature to talent flows into new technologies or new markets. And we've seen in the first talent flow, a lot of people who are focused on the foundation models, maybe because of people who are closest to models, new at the best. And now we're starting to see a really interesting pure tech place. So like SaaS veterans, SaaS founders who now have left their prior companies, maybe they wound down a previous company they were operating, maybe they're hard pivoting to do a AI first or AI-oriented software play. But I think the third wave is going to be the real business

operators who know that industry, but are also savvy enough and forward-thinking enough to really use AI as a technological advantage, but ultimately to win as a best-in-breed media company, retailer, pick your industry, construction firm maybe even. If you think about the role of capital allocator that you now play. So sometimes I think about in the early days of a business before they get to profitability, you're a product person, you're a growth person, you're a founder, now you're generating cash. Now you're a capital allocator. What companies, people have you studied, what books? How are you thinking about this role that is new for you? I hesitate to use that turn to wholeheartedly because capital is fungible. And so it makes it sound like this is purely

a spreadsheet game and you just put some numbers towards this initiative into these teams and give this person this much headcount and then your job is done. And I think in practice, the best operators at scale that I've observed are much more detail oriented. They still very much embrace the craft of what they're doing. I mean, you could take SpaceX as an example or Tesla. This is not just about capital allocation towards, okay, we're going to invest this much into the factory group and whatever. There's a reason why Elon slept on the factory floor is there's a reason why Elon slept on the factory floor. a reason why he's been very involved in even the design details of every product they build. So I think there is something to that detail orientation. And like I said before, it's not always the

case. In fact, maybe more often than not, it's not the case that the biggest best resourced team wins, especially when attention to detail moving quickly, having a very agile approach where you have very fast feedback cycles, that may matter the most. And so I see my role as almost a little bit of capital allocation, but I think it's also about maybe executive producing, where a movie that just has a massive budget and you throw the best talent at it might still be a flop. But I think probably, and I'm just speculating here, probably the best producers, they're a little bit more hands-on. And they're really making sure that the teams are operating in harmony. And they're really focused on making sure that the talent is just jelling perfectly. If I were acting in that role,

I would also be a lot more involved in the thematic design of this thing. Are we getting the story right? Are we getting the feel right? I think it's a much more craft, detail-oriented role, even at scale. And I think if you look at somebody like a Mark Zuckerberg, he's still very intimately involved in the details. I remember even before this recent AI wave for them, I would hear he would actually go and spend time individually with Oculus engineers when that was his big focus and literally touch and feel that product experience. I really admire the operators and especially founder operators who have had to scale themselves and involve their own skill set from.

Like you said, the early days where writing code, building a product and a team of one or two or five is very different from doing it at a much larger scale for us, hundreds of millions, for many other companies like billions or tens of billions, but to still have that level of finesse, attention to detail and somehow zoom into the micro in the areas that matter, but also have the whole forest in view because you can't only look at one part of one patch of the playing field. If you could take over any business in the world, no matter of its size, and run it for the next 10 years, what would you pick and why? Hmm. I am very happy with Airtable. And so it is a little hard to answer the question. But if I

took the business hat off, I think there are probably a lot of non-tech industries or businesses that would be probably a lot of fun to just work on for the sake of it. If I could take off, let's say, just the theme parks part of Disney. I know that comes with its own operational challenges and political challenges and so on. That's a great one. I think it would just be fun. It would be so different. And to get to like imagine your experiences for people and kids all over the world to come to and enjoy it.

It's a platform of its own right, where their platform is to enable people to create their own happy experience. Yeah. exactly. So yeah, maybe that's it. I love that answer. Well, I think you know my traditional closing question for everybody. What is the kindest thing that anyone's ever done for you? There are two really important moments for me that I'm cheating a little bit, but I want to. The first is I think my parents have just both been so incredibly supportive of my journey as an entrepreneur, even though I don't think they fully understood what I would entail. We grew up in a family and a dynamic where we didn't have a lot of money to spare. We didn't.

didn't have a lot of direct exposure to people who were entrepreneurs. So I think the fact that coming out of college, both my parents were just extremely supportive of me taking the leap, even though it meant giving up a potentially stable source of income to at least provide for myself and taking the risk that, like, I might have to come to you at some point and ask to have your help or crashing your home and even giving me what they could spare to help with that initial journey. And then second and relatedly, so I actually did this very brief stint at a company called crowdflower that was ultimately like a precursor to scale.

AI. And the founder, Lucas, I think I got recruited to the company when it was only three or four people. I became a wear all hats intern. And this lasted three, four months. I learned a lot while I was there. I ultimately wanted to start my own company. And I remember when leaving, Lucas asked me, what's stopping you from going and doing

your own company? And I basically told him, well, what I'm worried about is my savings because I basically have two and a half months of runway on a ramen budget. And so even just be able to get to something that allows me to either raise money or keep going, I feel like I'm threading the needle here. And he basically wrote me a small personal check, didn't expect to have it paid back on any particular time frame. And it wasn't just the thought that counted, even though that was powerful.

It was actually a pretty significant thing to me that I got to extend my ramen runway by just enough to where I teamed up with a college friend of mine, who I'm still friends with, went through YC, got accepted there. and they gave you very little money, 15,000 at the time. So it wasn't exactly like we hit the riches then either, but it gave me just enough more runway to then make it to that next runway extension, to then raising our first round, to then getting the first exit from Salesforce. And I'm always deeply grateful for all the incremental experiences and people who have

helped me get to each unlock because it really has been milestone to milestone for me. I would encourage everyone listening to go try to build something with the Airtable. It's like a general purpose computer or something that isn't constrained by your technical abilities. And we talked about it obviously a lot today, but this notion of the simple primitives of what makes a piece of software or a program we all live with, maybe people don't realize how accessible that is today through what you've built. And I'm excited to see all the ways that that grows and changes with the addition of AI to everything that you do. It's been so much fun talking about the history of the business with you. Thank you so much for your time.

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