Skip to content

By playing, you agree to uListen's Terms and Privacy Policy and YouTube's Terms of Service.

a16za16z

Why HubSpot’s Co-Founder Is Betting Against the SaaSpocalypse

As frontier AI labs expand into applications, where should startups build? a16z Partner Joe Schmidt joins HubSpot co-founder and CTO Dharmesh Shah to explore why the software opportunity is growing, and how founders can stay off the…

Dharmesh ShahguestJoe Schmidthost
Nov 10, 202659mWatch on YouTube

Summaries, chapters, transcripts, and FAQs on this page are made by uListen, not YouTube.

Great conversations run long. Their best ideas don't have to.

uListen is free on iPhone, iPad, Android, Chrome, and Claude.

See how

Transcript

  1. 0:00 – 0:47

    Intro

    1. DS

      ChatGPT and Claude Cowork, they're effectively the new OS.

    2. JS

      There's plenty of room to build software outside the path the frontier AI labs are racing down. Since then, HubSpot's been busy.

    3. DS

      The actual thing that's valuable is the work that's happening. The interface is a means to an end. Building an ecosystem is a lot like building a product, but you have to think of it that way. We had a very broad base of marketing applications. The reason we chose to do it all is the only way for us to solve the actual problem customers had, which is they're intimidated by this fragmentation. Everything's all right-- we had to do it all in one offering.

    4. JS

      The architectural decisions that you made when you originally built HubSpot was for a human to be navigating all of this information. Now we have these unbelievably powerful models that can go look at things in very different ways.

    5. DS

      What is the framework we need for the next twenty years of HubSpot now knowing that we have this AI world? Couple things. One is-

  2. 0:47 – 6:46

    The Best Time Ever to Be a Builder

    1. JS

      Welcome back to the a16z podcast. I'm Joe Schmidt, and today we're joined by Dharmesh Shah, the co-founder and CTO of HubSpot. We're picking up on a conversation that started with an exchange on X about my Yellow Brick Road essay, the idea that there's plenty of room to build software outside the path the frontier AI labs are racing down. Since then, HubSpot's been busy. They've launched uSpot, a CRM for one-person businesses, and they've been rebuilding their own platform around agents. Today, we're gonna talk about what's, what's worth building, what an established software company has to change, and what happens when AI becomes the way customers use your product. Dharmesh, thanks for joining us.

    2. DS

      Thanks for having me. Hi.

    3. JS

      Yeah. So excited. Dharmesh, you know LeBron, you know, uh-

    4. DS

      [laughs]

    5. JS

      ... Madonna. You know, the, you're, you've got the one word, you know the one name, uh, one word name, so pretty, pretty exciting. Um, this is a fun conversation for me because I wrote that essay, uh, you know, a few months ago, really in response to this, like, nihilistic, nothing matters feeling that was going on in Silicon Valley, and I feel like it's actually still going on. So it's gotta be strange running a big public company, successful pub-public company through this time period. Like, what does it feel like? I would love to maybe just start just with a, with a market check.

    6. DS

      Yeah, it's, uh, on the one hand, it's, uh, you know, the old best of times, worst of times. Um, it's the best of times because, you know, we have, um, all this kinda energy and excitement, um, that you have when you have this kind of massive transformation. I think back, HubSpot's twenty years old now, and when the company started, um, the kinda genesis of the company was, "Oh, there's this new thing called the internet that SMB should be benefiting from to drive growth and things like that." And they're not yet because it's overwhelming and the tools were fragmented and all that, and that's where HubSpot started. Now fast-forward twenty years, there's this new thing called AI that we think small, medium-sized businesses should be benefiting from, but it still feels intimidating or whatever. So it feels like this new, um, kind of, uh, revitalized mission, uh, in the same, yeah, same customer base, uh, just new technology. On the other hand, um, you know, it does create a lot of, uh, the fog of war and uncertainty, um, that, you know, we try to, for the most part, as a team, um, kinda stay focused on the customers, whatever the markets will do what the markets will do. Uh, we have our own thesis in terms of how things are, uh, evolving and, and what bets we're putting down. But, uh, I, I focus on the positive side. It's I've, you know, I've been building commercial software for thirty-plus years now-

    7. JS

      Yeah

    8. DS

      ... like literally, like every day. Um, and this is the most fun I've ever had, um, in terms of building things. It's, uh, this-- what a time to be a builder.

    9. JS

      Yeah. No, I totally agree. And it's, it's, it's so strange that the perception is negative when literally everything is working. It's like you fall out of the boat, [chuckles] out of the boat and you hit water right now. Um, maybe if we go back to... I'd love to talk a little bit about some of those format, you know, formation experiences, obviously with Pyramid and, and, um, and then of course HubSpot and, and how you guys got there. W- I, I, I'm sure back then people were saying some of the similar, you know, things about, "Hey, it's gotten so much easier to build software" and, you know, it's not, it's not... What, what are you even building now? I'd love to hear your commentary on maybe, um, you know, the, the similarities and the differences in terms of like how you're actually building companies and, and what you're seeing today with, uh, you know, with new startups.

    10. DS

      Yeah. I'll give you a, a slightly geeky response because I'm a geeky guy, is that, um, you know, having been in the software industry since prehistoric ages, um, a hundred and fifty years ago, but, you know, the story of software, um, and software development has always been this kinda increasing levels of abstraction, right? So w- when I started, we still had, you know, machine l-language. We were running on mainframes and things like that. Then we had 3GLs, uh, we had C and C++, and we had, you know, 4GLs. And so it's, what we've done is to said, "Okay, can we get to whatever manifestation you have in your head of whatever problem you're trying to solve, can we make that easier for you to do?" Right?

    11. JS

      Yeah.

    12. DS

      So back then, it was really, really hard, and then it got progressively easier. Uh, and we've had some jumps along the way in terms of, um, you know, nonlinear, uh, kind of increases in the, in the level of abstraction. But, like, at, at its base, this is the thing that kinda excites me about AI, is it gives us a new toolset to work at yet higher levels of abstraction.

    13. JS

      Yeah.

    14. DS

      That's the thing that's s-sorta changed, right? Um, and so in my mind, like the, the problems are still the same, you know, that, uh, humanity is dealing with and, and businesses are dealing with, uh, but now we have a better set of tools, and I have a s- you know, kinda software-oriented bias. Uh, that, that's my thing. But, um, so, so in, in some ways, it's sort of the, the same. And, and, um, we'll go off on a little bit of a tangent here because one of the things that's happened recently is, um, you know, we had the Jev model, uh-

    15. JS

      Oh, yeah

    16. DS

      ... uh, you know, released, uh, by the, by the TypeSafe folks. And what that really takes it back to is like, oh, well, it is really just software.

    17. JS

      Yeah.

    18. DS

      And here's another kinda dimension of development that we're having that's not a transformer, but it's actually kinda opening up. And, and that's what it does, is we're just putting more tools in the tool chest of builders, and we're expanding the scope of who can be a builder, which is awesome too. Um, but, like, a lot of it is still the same. We're still dealing with, uh, you know, kinda similar issues, and it's just nice to be able to work at a higher level of abstraction. Um, and one of the reasons I think this is a great time, it's always a good time to do a startup, uh, but it's a great time to do a startup, is that, um, because we're working at such high levels of ab-abstraction, AI lets you iterate so much faster.

    19. JS

      Oh, yeah.

    20. DS

      So you can learn the lessons that might have taken you six months, a year, or two years to do. It's like, okay, well- Just try it, you know, build something, get something out there, test it, you know, with actual users and customers and, and see what happens. Uh, and that just wasn't possible before. So it's, uh, but yeah.

    21. JS

      Oh my gosh, I, I totally agree with you, and this is-- I, I wish you could see, you know, all the things that we're even building here at, at our, at our venture capital firm, right? I mean-

    22. DS

      Yeah

    23. JS

      ... and so it's, it-- even being on the front lines is so much more fun. This is actually leads into a question I wanted to talk to you about, which is I think historically, and if you look, uh, especially at the businesses you've built, um, you know, you would start in some very narrow product, you know, surface and for a, for a very narrow type of customer. Um, how have you observed this changing with that fast iterative process and-

    24. DS

      Yeah

    25. JS

      ... the ability to build so much more, create so much more software so much faster? I'd love to hear you talk a little bit about that.

  3. 6:46 – 10:32

    How to Actually Pick a Startup Idea

    1. DS

      Yeah. I think this, this, um, you know, I've been in kind of the startup world, uh, directly, indirectly for a long time. I still think of myself, um, as a startup guy, is that, uh, like, picking the ideas has always been, like, one of the hardest things-

    2. JS

      Yeah

    3. DS

      ... um, especially for, like, you know, first time, first-time founders that haven't been through it. Um, and even now, after thirty years, I think I'm marginally better at it than I used to be. Um, but one of the things I think the, the common mistakes, uh, entrepreneurs make in terms of kinda idea selection is that think, they think it's a matter of, like, rigorous analysis, that we're gonna have this kinda eighteen-column spreadsheet that's gonna give me all the coefficients and how it's gonna spit-

    4. JS

      Yeah

    5. DS

      ... you know, the best possible idea. Uh, and-

    6. JS

      Yeah, the number of times I've seen the, like, US census file, uh, like-

    7. DS

      Yeah, exactly. It's like, oh, there's-

    8. JS

      The number of em- of people in-

    9. DS

      ... eighteen thousand lawyer law firms between two and five employees.

    10. JS

      I'm like, okay, this is not relevant. Um [chuckles]

    11. DS

      Yeah. Um, and, and just, and, and what I've experienced, and then, uh, not just me, it's, uh, not just a data point of one, but even, you know, talking to other founders, um, is that, uh, it's almost impossible to know a priori what the quality, quote-unquote, of an idea is or its kind of probability-

    12. JS

      Yeah

    13. DS

      ... of success, um, or what the outcome's gonna look like. And the only real way to test an idea is to actually, or know if an idea is good, is to actually test it. And what ends up, even if you're successful, the thing that you end up with that succeeds will oftentimes, or more often than not, not resemble the original idea you started with, right?

    14. JS

      Yeah.

    15. DS

      So what, what you get to that people now think of as, "Oh, that was a successful business," sometimes it doesn't resemble, um, you know, will have some similar, similarity. Maybe the, you know, the, the gene pool of the original, uh, you know, concepts are still there, but, um, but it's different. So the, so the kind of byproduct of that thinking is that, um, since you're unlikely to know, um, whether a given idea of a set of ideas maybe you're exploring is going to be good or not, um, then the, then the, like, the, the smart move is to say, "Let's get the idea out there as quickly as possible and learn as quickly as possible as to whether this or, um, is or is not a good idea, and should I be pursuing something, something else." Um-

    16. JS

      Yeah

    17. DS

      ... so I think that, that still holds today. Um, once again, back to the kind of AI point, we can iterate faster now. The one thing that's still hard, and I would argue maybe even harder now than it was, is distribution, right?

    18. JS

      Yeah.

    19. DS

      But that's, you know, because it's so much easier to build software now, many more people will be doing it, so there's going to be m- many more options, and it's gonna be hard to sort of cut through the noise and kinda get into contact, uh, with a customer. And, you know, one of the things I'm a big believer in is kinda starting out with, um, and I'm not the only person that has, uh, said this by far, but it's sort of to narrow the aperture in the kind of early days. Just kind of focus in, uh, because that's what increases your odds of-

    20. JS

      Yeah

    21. DS

      ... of, uh, you know, kind of getting, kinda getting through and learning. And the, you know, mistake number two that, uh, entrepreneurs make is that they feel like they need to start with this massive, you know, TAM where it's like, "Oh, this can be a hundred billion dollar business." Like, that, that's great that you have the kind of ambition and vision that this could be a hundred b- But, like, tell me what you're gonna do in the first thirty days.

    22. JS

      Yeah, yeah.

    23. DS

      Like, what's the thing that you're actually gonna focus on?

    24. JS

      I think it's a one dollar revenue first.

    25. DS

      Exactly, yeah.

    26. JS

      Then, then let's get to a hundred billion.

    27. DS

      Um, and, and I, and, like, even now for me, it's like as I explore new ideas, uh, you know, within, within HubSpot, both for HubSpot Next, which is our innovation arm, or even, um, new ideas within the core platform, it's like kinda narrowing that field of view. So like, you know, can we somehow... Like, I think the, the people that win are the ones that have the shortest iter- uh, iteration time and the cheapest learning time. So the, if you can sort of shorten that loop, uh, and learn less expensively, uh, whatever the lessons are going to be, uh, over, over the fullness of time, you end up, end, end up winning on average, obviously, yeah.

    28. JS

      Yeah. No, and I, I still think, though, I, that this moment is the best possible moment for high agency people.

    29. DS

      Yeah.

    30. JS

      And the reality is, yes, there's, it's way easier to build software, but there's only so many people who are sitting alone on, you know, ten forty-five PM when we exchange an email on, on Saturday night,

  4. 10:32 – 14:40

    From LegalSpot to HubSpot: The SMB Bet

    1. JS

      you know, hacking on their computer-

    2. DS

      Yeah

    3. JS

      ... and just go be that person and go build something cool. I, I wanna go to, back to, like, what I-

    4. DS

      Yeah

    5. JS

      ... I believe at the beginning of HubSpot you started with, and correct me if I'm wrong, uh, a legal product. Um, uh, maybe there was a legal, a legal product in there somewhere, and somehow that has now become this CRM for small and mid-sized businesses. I mean, it's unbelievable what you guys have, like, expanded that wedge into. I'd love to hear how you guys thought about, you know, the-

    6. DS

      Yeah

    7. JS

      ... of course, that evolved into SMB marketing wedge and so on and so forth, but would love to hear how you expanded from there to platform.

    8. DS

      Yeah. Um, so when we first started, you know, we were in business school. That's where my, uh, co-founder and I met, um, uh, here o- on the East Coast. And the, the thing we sort of alighted upon was first the, um, the kind of beauty of the SMB as a market. Um, it's like it was underserved. Like, and I'll give you the quick story here because it's, uh, it's interesting. So we had both been in software for a long time, Guy on the kinda, uh, marketing and sales side, mine on the kinda product development engineering side. Uh, and, and we had both done enterprise software before, uh, so we sorta knew that. Um, and the challenge with enterprise software was it's, uh, completely non-fun, um-

    9. JS

      Yeah

    10. DS

      ... as a startup entrepreneur. Sales cycles are long. You get revenue concentration.

    11. JS

      Yeah.

    12. DS

      You got a small set of customers kinda impacting your product roadmap and your vision and things like that. Um, and, and so the, so there's that. Uh, on the other end of the, the spectrum, you have consumer markets, right? It's, uh, consumer companies in the software space. And the issue with consumer markets is that a lot of them have very bimodal outcomes that says, okay, if you're successful, it's gonna be this massive outcome. Um, and a lot of times you're doing kinda, um, advertising subsidies or figuring out some other way. You can't really charge for the product, uh, because, you know, don't be a chump kinda thing. And the, the beauty of SMB was, um, just from a pure market analysis perspective, this is, you know, two guys in business school at the time. It's like, oh, well- SMB has, like, the beauty of enterprise in terms of you can actually charge money for a problem that you're solving. It's like there's nothing wrong with charging money, but it has a scale of consumer that you're not gonna have this revenue concentration. You can have tens of thousands, hundreds, thousands, millions of customers, and so you sort of have the kind of best of both, right? And, um, so that was our kind of original thinking is like, oh, if we were to do a startup, like... Uh, and so then our original idea that we took to the, uh, MIT 50K, uh, competition was, um, uh, like ERP software, like an Oracle for small business.

    13. JS

      Yep.

    14. DS

      Um, and the way, and this going back to the kind of narrow the aperture is like, okay, well, that seems like an interesting thing because, you know, we're going to this kind of transition from client server to, to the web. It's like, like someone should create like an Oracle for, for small business. That's everything's integrated, kind of works together. It's kind of built for SMB. Um, and when we went to the business plan competition, uh, we said w- that's, like, too big. Let's narrow the aperture, and let's call it LegalSpot. Uh, the HubSpot company name had already existed, but the thing we wrote the business plan around, um, purely as a way to kind of test the market, uh, early on is to say, let's focus on ERP for law firms, just as a narrow and, you know, back to the, your original act. I was like, ah, there's, you know, all the US Census Bureau things. Oh, there's lots of them out there.

    15. JS

      Yeah, yeah, there's lots of law firms.

    16. DS

      We can multiply X by Y.

    17. JS

      Yeah.

    18. DS

      And it makes for great, uh, business school kind of spreadsheets-

    19. JS

      Yeah

    20. DS

      ... um, and, and PowerPoint decks. And so, and so we did that, and, and, uh, you know, we have a kind of propensity for kind of testing the waters. And one thing we learned early on, uh, in that, uh, while we're still in, in business school, kind of noodling on the HubSpot idea, is like, that's not a fun market to sell into. Uh, and as we talked to the law firms and we talked about, you know, ERP and what the solution, you know, would or could be, uh, the, the marketing and the website, all that is, was the part that was the most interesting to them.

    21. JS

      Yeah. They're like, "Wait, hold on."

    22. DS

      That's-

    23. JS

      "Can you help me find more customers?"

    24. DS

      Exactly, yeah. It's like we got, like, time tracking and all that. That doesn't really matter. Can you help me find more, uh, more clients that are the ideal fit and things like that. And so we were sort of, like, down to those two ideas, like, oh, do we kind of stick with the kind of ERP for, um, for SMB, or do we do this kind of marketing for, for SMB? And we chose-- We almost could have flipped a coin. We almost could have. Uh, we, we didn't. We were, you know, a little bit more deliberate about it, and we did extrapolate from a, a minor set of data points, but it was like, life is short. Uh, as it turns out, lawyers, uh, are the second hardest market to sell into, uh, software developers being the hardest. Uh, it's like-

  5. 14:40 – 20:29

    The 60% That Weren't Using a CRM

    1. JS

      So now we're in, uh, you know, na- navigating the SMB kind of marketing, uh, product.

    2. DS

      Yeah.

    3. JS

      And of course, that then morphs into, you know, more of a, much more of a platform and a full CRM.

    4. DS

      Yeah.

    5. JS

      Was that thoughtful from day zero? And I, the reason I'm asking is I think there's a lot of folks out here that maybe are using some new AI medium, whether that's-

    6. DS

      Yep

    7. JS

      ... you know, commonly early on it was voice, voice products.

    8. DS

      Yep.

    9. JS

      And I actually still think that's a huge opportunity for so many folks.

    10. DS

      Yep.

    11. JS

      Um, I'm sure you're exploring that at HubSpot. Uh-

    12. DS

      Yeah

    13. JS

      ... or maybe there's, there's some sort of support kind of chat, you know, interface. And everyone, everyone always wants to end up being the HubSpot for whatever use case, right?

    14. DS

      Sure. Yeah.

    15. JS

      So maybe talk a little bit about how you, you navigated that product journey and created the right, the right loops to get to more of a platform from this initial wedge.

    16. DS

      Yeah. Uh, so we started off as a, um, a suite of marketing applications. And, uh, to answer your kind of, um, y- the first question, this was not a time T equals zero, we think we're gonna be a CRM platform-

    17. JS

      Yeah, yeah

    18. DS

      ... someday. That was not, um, it was not in any of our PowerPoint decks. It was, um... And, and so yeah, we, we didn't have that. But as the company evolved, we sort of, and this may be a little bit, uh, kind of idiosyncratic to HubSpot, but I don't think so. I think the pattern still exists even outside of a, a CRM. You sort of have the need for, um, kind of product number two, right? Like the, the second act or however, uh, one tends to phrase it. And we sort of hit that. So a couple of things happened simultaneously. Uh, we were in the marketing software business, and one of the things we sort of recognized is that being in the marketing software business, um, you know, was very, very close to, you know, sales and, and CRM. That was the kind of the adjacent category, so that kind of made logical sense. Uh, the other thing that we sort of learned, um, you know, once, you know, we had a bunch of customers, is that we did a survey of our customers. Um, this is, you know, maybe a couple of years before the IPO, um, so maybe five, six years into the company, uh, and just to see which CRMs they were using. And we discovered, um, that 20%, um, were using salesforce.com. Uh, another 20% were using some other CRM other than salesforce.com.

    19. JS

      Yeah.

    20. DS

      Uh, but 60% were not using a CRM at all.

    21. JS

      Yeah.

    22. DS

      And that was insane to us. It's like, wait a second. So you're using HubSpot software, you're paying us money, so these are not like random, you know, companies. These are people paying us money, um, to generate leads. What's happening to those leads? Like if they're not-

    23. JS

      Yeah, where are you putting them?

    24. DS

      Like what... And so that was, so the, the thing that we sort of aligned upon was like, okay, well, like these companies should be using a CRM. So the, you know, we're, uh, my co-founder and I are very, um, weirdly sim-simple-minded. It's like, okay, well, why are they not using a CRM? It's like it's the categories exist. It's not, and there are great companies with, you know, great products out there. And we came down to, it's like, oh, well, it's, uh, overwhelming. They're just intimidated by the idea of a CRM. Uh, it's too expensive. They think it's gonna cost too much to kinda get going with a CRM. Um, and so we made the decision that it was a nice adjacent thing. We need a second product anyway. There was a strategic value to saying, okay, well, over the fullness of time, um, the existing CRM players are likely gonna get into the marketing space, uh, and we're gonna have... So even if we wanna stay in, in the marketing software category, it's unlikely that we're gonna be able to do that forever, right? So the, all th- those things sort of converged a-around the same time period. And so when we launched the CRM, um, you know, we sort of addressed both of those issues. Like, okay, so we're gonna build a CRM, and we're gonna address the 60%. It's like, okay, for those people that have not picked a CRM yet, why have they not done it? Too hard? Okay, let's make the easiest CRM that's ever been conceived, you know, at that point in time, and we're focused on SMB, so that was a core strength anyway. And we're gonna take the, it's too expensive off the table because we're going to make it free.

    25. JS

      Hmm.

    26. DS

      Not freemium, free. It's like, okay, well, w- uh, and this, this has been sort of the, um, kind of philosophy at HubSpot is, like, create customer value before you try to extract it. It's like if you can just get out there with something that's useful, and we can obviously subsidize it with the kind of the marketing software business that we already had, um, but that, that was the idea. It's like, okay, well, then we've sort of removed the obstacles, uh, and then we're just like, "Well, let's go out there." And then, uh, and then as we iterated, we sort of learned, um, you know, that in, in order to win in the kind of CRM space, it needed to be a platform, right? It's like it can't just be a, uh, because CRM is almost by definition now, um, and even then, um, are platforms. So that, that was the journey. It's just going from product one to product two, expanding out into a strategic adjacent thing. But then, and this is the other thing I think that carries over to, um, the entrepreneurs listening is that do what you can to solve, like, the whole customer problem. Uh, like, how-however you define it, whatever circle you draw it around, it's like what's the actual problem we're trying to solve? So even when HubSpot started, um, we had a very broad base of marketing applications and which goes against every piece of advice you get that, and that I've given. It's like, do one thing really, really well. We did all the things. So we had content management and SEO and landing pages and forms or whatever, and they were great companies with great products in each of those categories. The reason we chose to do it all is the only way for us to solve the actual problem customers had, which is they're intimidated by this fragmentation. Everything's all right. We had to do it all in one offering, and this, we sort of carried that idea with us for 20 years, and we still do the same thing now. It's like, what's the actual problem we're gonna solve?

    27. JS

      Well, it's so interesting because I think, you know, with hindsight, everyone always talks about, okay, well, you know, from a go-to-market strategy, there's, you know, the brownfield approach of you go after all these people who are on every core system already, and you rip and replace it, and hopefully you're in the middle of a cycle like this where everyone's willing to revisit that. Or you go after greenfield and you, you try to win all these new businesses who have never adopted software before. And I think, uh, the standard thinking on HubSpot is like, "Oh, well, it was just like a very thoughtful from the beginning. Let, you know, let's go after the greenfield." And the reality is it's like, actually no, like all of our customers just didn't have it, and we built it, and, um, it's totally fine if you do it in that order, and you can still be a massive company if you, if you nail

  6. 20:29 – 23:33

    Building an Ecosystem Developers Trust

    1. JS

      it. Um, so interesting. How, how did you think about, you know, another part of this is, you know, building the right ecosystem around that initial offering, though, right? And I think that's something that HubSpot has done such an amazing job of. Maybe what are the key learnings from the early days of building the ecosystem, and then now that you, you know, it's such a different world today where that ecos-- you know, you're, you're able to do a lot of that in different ways. I'd love to hear you talk about that with, you know, the advent of AI and how that's impacting the ecosystem as well.

    2. DS

      Yeah, uh, and I've been a kind of consumer of platforms and ecosystems for, you know, most of my professional career, so I sort of know what it's like on both sides. Um, and this is, uh, no disparagement to any of the existing platforms then or now, but, uh, what tends to happen with, um, ecosystems and platforms is that, um, the platform providers will start, um, you know, looking to attract the developer ecosystem and, and get people to build apps or build, uh, you know, companies that are kind of adding value to the core platform. Uh, but over time, what ends up happening is that, um, either intentionally or, or, uh, unintentionally, they sort of kind of stomp over, uh, those developers. They'll close things up. They'll shut down. And we saw this, I can, we can talk about this, it's, uh, story's 15-plus years old now. Like Twitter had this, uh, in the early days, right? They had-- the original Twitter was this kind of very open ecosystem. We're a platform and w- and, and literally a thousand flowers bloomed back then. Lots of Twitter clients built, lots of people, um, you know, built, uh, products and, and, uh, and companies, uh, on that original set of APIs. And it's not exclusive to Twitter. That's just what jumps to mind right now. But, um, but then they sort of closed it down and became k- you know, everything from like developer unfriendly to almost developer hostile, right? Um, and so one of the things that I think this has kinda helped us build the ecosystem is that we've sort of take the, taken those lessons to heart that says, okay, well, let's be the platform that we wished would have existed for us back when we were doing, you know, varying and sundry things. It's like, let's, you know, be the, sort of carry that HubSpot brand of being kind of friendly, likable, um, and that's like we're, we're a, we're red-blooded capitalists. Uh, we like to think of ourselves as like warm-hearted, red-blooded capitalists. Like, okay, everyone's here to make money. Here's what we're doing. We try to be, uh, even with our, uh, with our partners, uh, try to be a, a as open a book as we can be in terms of like, okay, here, here are things that we're-- we can't promise you we're not gonna do this, but here's what we're doing. If we were looking for white space, if we were you, here's the things, the kinds of things we would do that we think there's an opportunity that, you know, we're highly, highly unlikely to kind of traverse into that territory. Uh, and that's worked. So we've built trust, um, you know, across a, a wide pool of people now. And then some of it's just the, um, you know, building an ecosystem is a, a lot like building a product, but you have to think of it that way, is to say, okay, well, you know, um, as a developer at heart, it's like, okay, well, in the same way we think about time to value for the core customers, we have to think about time to joy for a developer. How fast can someone get to, "Oh, I got something working on, with the HubSpot APIs," to the, "Oh, I made my first dollar," you know, working with the HubSpot, uh, within the HubSpot ecosystem. Whatever that is, can we compress that? Can we reduce the time to joy, um, you know, for, for that, that class of customer? Um-

    3. JS

      Yeah.

    4. DS

      Yeah, so.

  7. 23:33 – 27:23

    Staying Off the Yellow Brick Road

    1. JS

      No, total-totally, and, uh, you know, congrats on, on building such a thriving ecosystem. Maybe just transitioning a little bit now into, you know, the, the essay and the conversation we were having around it.

    2. DS

      Mm-hmm.

    3. JS

      I'd love to hear how you think about the things that you're maybe concerned about, uh, with, with regard to the workflows that the, the labs are kinda going after. And of course, there's many things inside of the HubSpot experience, I think that, that those, those horizontal products are trying to pull in. Um-

    4. DS

      Yep

    5. JS

      ... and how you think about like the, the area that by which you wanna defend and, and where you can kind of build the most durable value. So would love your commentary on that.

    6. DS

      Yeah. A, a few thoughts. Uh, one is I'm gonna open this with a, you have a very, very biased, uh, point of view, uh, given I'm, uh, co-founder, CTO of HubSpot.

    7. JS

      I would assume so.

    8. DS

      That doesn't necessarily make me wrong. Um, so the... A, a few thoughts. One is around, um, and we'll just talk more broadly, and then we can kinda zero in on, uh, some of the specific things, is that- You know, for those on the outside looking in, um, it's like it feels like, and this is your kinda Yellow Brick, uh, Road, which I, you know, I love the essay. The, the, the idea and the analogy was, you know, very apt, which is like there's this, you know, well-paved, shiny bright road that's kinda pointed to the Emerald City or whatever. How could you not, like, walk down it? It's like, and it feels empty at the time, right? But you don't know that there's like a, like a fleet of tanks coming down that road.

    9. JS

      [laughs] Yeah.

    10. DS

      Um, and if, and they're getting faster. It's not that they're even slow because they've invented this new thing that makes them even go fast.

    11. JS

      Yeah.

    12. DS

      Um, and so, and, and this was kind of the, the premise of your essay, which I g- wholeheartedly agree with, is that, uh, yes, they're going to, um, they're going to barrel down that road. Um, and it's, it is a shiny, nice road, and they're... and hence the trillion dollar, um, you know, valuations, um, and, and soon to be market caps. But, uh, but there's so much opportunity outside of that. And so here's, like, a, I'm a framework kinda guy. Like I... Here's a mistake I think entrepreneurs make on both sides of it. Uh, one, when they dismiss it, they say, "Oh, well, ChatGPT doesn't do this or Cloud Cowork doesn't do that. Well, we're gonna do this. We're gonna fill this gap." Um, a- and those are the ideas that, uh, make me nervous, right?

    13. JS

      Yeah.

    14. DS

      Because there's, and there's two things sort of, I think, misguided about that thinking. One is if you can a- just append, uh, the word "yet" to it, uh, and it still holds water, that's a problem. It's like, "Oh, ChatGPT can't do that yet." Does that ring true? If it-

    15. JS

      Yeah

    16. DS

      ... sorta does, it's like, okay, well, that, that's a problem. That's thing number one. Thing number two is what I think of as, uh, measuring the product versus the company. It's like, "Oh, ChatGPT can't do that." It's like, okay, but can OpenAI the company do that?

    17. JS

      Yeah.

    18. DS

      Like, maybe, you know, it's a new product, a new product line or whatever. It's like, so you have to sorta think through how wide is that road. Um-

    19. JS

      Right

    20. DS

      ... you know, it's like, okay, maybe the road itself is getting wider, fine. So maybe stay off the road, but maybe also stay o- you know, a little bit outside. It's like, okay, if there's all things being equal, if you can just be a little bit further away, so even if as they expand, when you can... And it's somewhat predictable. Um, so that's kinda thing number one, is to make sure, um, that... And, and either, you know, pick things that the frontier model companies, um, either have a structural constraint, uh, that they just can't do regulatory-wise, whatever it is, um, or have a, what I think of as, like, an incentive constraint.

    21. JS

      Yeah.

    22. DS

      It's like, yes, they could do that, but the odds of them doing that are really, really low because, uh, I'll just... I'll use HubSpot as an example. So let's say, um, any of the frontier model companies, um, I'll, I'll, I'll pick on Sam Altman since, uh, I'm an investor in OpenAI and, and know Sam. It's like he could wake up tomorrow and say, "You know, that CRM industry, that looks awfully tempting. And that, uh, you know, three-plus billion dollars of revenue that, that HubSpot has, we're gonna make it our mission or somebody within the company's mission to go, go take that revenue." Are they smart enough? Yes. Do they have ambition? Yes. But it's that very ambition that makes it highly unlikely that the team at OpenAI or Sam or anyone, or same thing with Anthropic or any of the frontier model companies, they're unlikely to do that. Um, and but that's what it would

  8. 27:23 – 31:49

    Why Models Are Becoming Infrastructure

    1. DS

      take to be able to... It's like, okay, so my thesis on this kinda winds back to where do you find the pockets of opportunity that are off the Yellow Brick Road, which is one thing we know for sure is the frontier model companies have invested hundreds of billions of dollars kinda building, doing the pre-training as, as you-

    2. JS

      Yeah

    3. DS

      ... as you well know. And so their whole business effectively now is, um, taking the CapEx that they've spent and the money they're continuing to spend and say, "How do we sell inference? How do we sell tokens? How do we kinda monetize the thing that we've built?"

    4. JS

      Yep.

    5. DS

      Awesome. Uh, it's a tale as old as time. And they have two choices. Uh, like one is they can say, "Oh, we're just gonna go directly to everybody on the planet and sell them tokens." And they'll do that by virtue of the consumer product, and they'll do that. Uh, and even on the high end, they'll say, "Okay, well, we're gonna go directly to Fortune 500 JP Morgans, and we're gonna sell them tokens because we're gonna sell them. There doesn't need to be an intermediary. We'll have an army of FDEs that can help us do that." Uh, and that's certainly a path, and they're kinda walking down that path. Another possible path is it's like, okay, well, there are, uh, you know, SaaS companies and products and companies and startups or whatever that know their particular domain, know their market, have some distribution, or have something that allows them to sell t- tokens and be a reseller of inference-

    6. JS

      Yeah

    7. DS

      ... for lack of a better, lack of a better term. It's like, okay, well... And they don't actually, this is not a binary choice. They can do both. They can say, "Oh, for these kinds of things, for consumer, we wanna own that. We don't want other things, you know, people doing, like, a ChatGPT equivalent. Uh, we're gonna own that." But there's going to be lots of opportunity to say, "Okay, well, think of, um, AI as infrastructure in the same way we thought of AWS and, and cloud-

    8. JS

      Yeah, of course

    9. DS

      ... computer as infrastructure 15 years ago." Uh, it's the exact same thing. It's like, and, um, and as entrepreneurs, and once again, I'm biased but not wrong, I think it's less scary now to build, um, in and around, uh, kind of the AI arena than it was in 2023. And the reason I say that is back in 2023, um, uh, and I have, you know, very few, uh, sim- uh, rules for business. One is, uh, never own a restaurant. Number two is-

    10. JS

      [laughs]

    11. DS

      ... don't ever be a landlord and charge rent, and number three is never compete with Sam Altman.

    12. JS

      [laughs]

    13. DS

      Um, which I've absolutely done a couple times. Um-

    14. JS

      Oops [laughs]

    15. DS

      ... but the, the, the reality though is that at the time in 2023, if we take ourselves back, um, if I were a betting person, uh, I would've said they are just so far ahead. It, it was like they had figured out this kind of-

    16. JS

      Yeah

    17. DS

      ... magic elixir or something like that. It's like I have no idea if anyone ever is going to be able to really solve it in the, and, and be able to pull all the things, including the CapEx spend that was gonna be necessary. Uh, I was thankfully wrong about that. Like, you know, thankfully in terms of from entrepreneurs and from an ecosystem perspective because n- where things stand now is that, um, models, I'm not gonna say they're a commodity, but part of the virtue of them being as common as they are is they are infrastructure, and they're going to trend toward both from a pricing perspective-

    18. JS

      Yeah

    19. DS

      ... uh, and even a fungibility perspective. Right now, you know, things like OpenRouter exist because it's easy to sorta swap out models back, and even within HubSpot and most SaaS companies are doing some form of, uh, use case decomposition and model routing to figure out what's the right model for the, for this task. The byproduct of that is that- Like, you can sort of swap out what you need or whatever, so there's less reason to be scared because you don't have this kinda, uh, supply chain power where, like, this one company controls the one crucial ingredient for you to build a product and solve the problem that you're solving. We don't have that right now. We have open source. We have lots of great commercial products. Um, so, hey, as long as you stay off the Yellow Brick Road, it is completely possible now that you're not gonna get margined out or whatever because you have options. There's not just-

    20. JS

      Yeah

    21. DS

      ... one company providing the one key ingredient to building a, an AI, an AI native company.

    22. JS

      No, totally, and I think that the other piece of this which you kinda touch on, it's like, okay, of course with enough focus, but it's, it's also just that, uh, the advances or the advantages you're gonna build as, uh, as a, as a startup or as a Hubspot or whoever it is-

    23. DS

      Yeah

    24. JS

      ... if you're not building a, a foundation model, it's you're actually b- you're building still high-quality software. I think that is the layer that you're going after to-

    25. DS

      Yeah

    26. JS

      ... provide value to your customer. And so of course there's all these other great pieces of infrastructure that you can now package up and then give to those, those, uh-

    27. DS

      Totally

    28. JS

      ... you know, those customers.

    29. DS

      Yeah.

    30. JS

      And, and the way that at least I, I think about it, and I, of course, I argue in, in this essay, it's like-

  9. 31:49 – 37:04

    ChatGPT Is the New OS

    1. JS

      build in.

    2. DS

      Yeah. It, it's compounding loops all the way down, Joe. Um-

    3. JS

      Of course

    4. DS

      ... that's, that's the, the entire name of the game before AI, after AI. But, um, a couple of things I'll say. So I'm gonna rewind, um, uh, push one idea on the, on the stack. I'm gonna pop it off and, and talk through it, is that so the... You know, one of the questions we get, you know, back on the, um, if you were an entrepreneur kinda evaluating the ideas, um, it's like, okay, well, so how was... Uh, today was, uh, OpenAI's, uh, dev day, by the way, so September 30th. So I-

    5. JS

      Yeah

    6. DS

      ... you know, right before this call was, uh-

    7. JS

      Date our recording. Our recording

    8. DS

      ... was watching it. [laughs] What's that?

    9. JS

      I said we're dating the recording for the, for the-

    10. DS

      Yeah

    11. JS

      ... for the audience. So you now know timestamped. Uh.

    12. DS

      Exactly. You know exactly when this is being recorded. Um, so yeah, and Hu- HubSpot launched our, um... We were the first, you know, CRM company to do so, uh, our connector to ChatGPT over a year ago, and we have a similar one for Anthropic. And one of the questions, you know, we get sometimes is like, "Oh, well, doesn't that sorta diminish the long-term value of HubSpot?" It's like, okay, if everyone's coming in through ChatGPT-

    13. JS

      Right

    14. DS

      ... or through Cloud Cowork or something like that, isn't that, like, just saying eroding the kinda long-term value? And, um, I under- and I st- understand the reasoning. My sorta counter to that is that, um, w- like the interface matters, uh, it, but it matters to varying degrees. There's a spectrum. So let's say one other spectrum, I'm making this up as I go. Um, let's say you look at something like an Uber Eats or a DoorDash, right? It's like, okay, they have very compelling apps, but it's like, okay, well, the app is trivial. What's actually valuable is the network that they have, the partnerships with the restaurants, and things like that, so a, a driver can show up, pick up food, and deliver it to the logistics and all of that. That's where the actual value is, and that's the part that's sorta hard to displace. Um, there are other things where the int- the interface is the entirety of the product, right? That's all there is. If it weren't for the interface, there would be no value, so. And, and, and I think most business applications actually sit more towards the Uber Eats kinda thing, right? Not in the s-sense that they have, uh, network effects like, uh, Uber Eats and DoorDash do, but in the sense that the actual thing that's valuable is the work that's happening. The interface is a means to an end.

    15. JS

      Mm-hmm.

    16. DS

      The only reason the interface is important to HubSpot is because, uh, marketers and salespeople are trying to get something done or deliver on certain outcomes, and the interface is the way b- in, in order for them to communicate with a computer. And before AI happened, it's like HubSpot's claim to fame was like, oh, we understand the problem you're trying to do, whatever, and we've actually built the interface that's super simple to help you kinda get there. But that was not the entirety of the value. Now, one can argue where on the spectrum, you know, you know, HubSpot falls, but I can guarantee you it does not fall on the, the interface has, you know, 99% of the value. And so my argument is that in the same way, you know, uh, you know, everything kinda comes back around again, is that, uh, when, when we saw the original kinda SaaS development just moving to the kinda web-based application world, we had the same kinda thing. It's like, okay, well, we had client server apps, you know, running on our desktops, you know, back in the day. Uh, and then we had, you know, this browser thing come along, the Internet Explorer and, you know, later, uh, Mozilla as, uh, a few folks, uh, uh, know well. Um, and it's like, okay, well, now my software is running inside of a browser. But as it turns out, like, Google did not own the entire application stack, right? There was an entire lots and lots of, lots of software companies that got built, and it's like, okay, the interface became the browser because that was a convenient way to get to the actual kinda product that you needed to get to. A browser was a better way than installing something on, you know, n-teen, uh, million, uh, computers. We have a similar thing now. Like, and it's different, right? Because, uh, the AI is-

    17. JS

      Yeah

    18. DS

      ... doing cognition work. It's doing actual smart things.

    19. JS

      Right.

    20. DS

      It can do more things. But if you just look at the kinda ChatGPT and Cloud Cowork thing or whatever, they're effectively the new OS.

    21. JS

      Yeah.

    22. DS

      They are the new browsers. Like, this is the way people are consuming value, and there will be companies like HubSpot that sorta plug in, and this is the thing we saw from dev day, and we saw it, you know, last year, and we continue to see. It's like, oh, they're gonna increase the richness of what's available within ChatGPT. They launched a, you know, this kinda go three at a, a marketplace for, you know, people to plug in, in a, um, you know, in plug-in architecture and all those things, which makes complete sense to me. Um-

    23. JS

      Yeah

    24. DS

      ... so, so my fundamental philosophy, which I think holds true for most people looking to, at least in the, in the B2B world, is that, like, don't overthink the interface. Figure out what's the best possible way to achieve the outcomes your customers are looking for. Um, and then the other th- you know, value that you can bring, uh, is around, like, what's the economic, um, you know, best way? What's the lowest latency way? What's the lowest friction way to kinda accomplish those outcomes? Um, and often, like an LLM is very, very good at lots and lots of things.

    25. JS

      Yeah.

    26. DS

      It's the most, you know, brilliant technology we've had in the history of humankind. But, uh, one of the things that I've kinda gone on record as saying is, like, anything that can be solved with a structured query deterministically, solve with a structured query deterministically.

    27. JS

      Yeah.

    28. DS

      Don't use inference to solve something that a query can solve.

    29. JS

      Yeah, exactly.

    30. DS

      It doesn't make any sense. So even if the data were there, it's like, oh- If your question is how many customers or how much revenue are we getting from, uh, from Ireland-

  10. 37:04 – 40:44

    Rethinking HubSpot's Primitives for AI

    1. JS

      uh-

    2. DS

      Yeah

    3. JS

      ... ideas here, which could be its own dang podcast, I think, um [chuckles] -

    4. DS

      Yeah

    5. JS

      ... for that matter. May- maybe last question on this, 'cause I think it's really interesting you talk through this. The architectural decisions that you made when you originally built HubSpot are probably-

    6. DS

      Yeah

    7. JS

      ... uh, you know, when we with the benefit of hindsight, I don't know how you would revisit those. I, I'm, I'm so curious to hear how you think about that because, of course, now the way that, you know, that you d-designed it originally was for a human to, to be navigating all of this-

    8. DS

      Yep

    9. JS

      ... this information, and you're storing that in a way that's intuitive to a human.

    10. DS

      Yeah.

    11. JS

      And so now we have these, you know, unbelievable, unbelievably powerful models that can go look at things in very different ways, and so I don't know how you think about the way you may have wanted to start differently or maybe what you're doing to address that now a-as, as a, as an executive at HubSpot.

    12. DS

      Yeah. Um, a couple things. One is, you know, so I'm a classically trained computer science software development nerd, um, forever, and one of the things I'm a big believer in, in terms of, uh, you know, large scale software applications, is around this notion of like frameworks and, and starting with a set of primitives that you sort of evolve over time, because that's the thing that gives you, that's coming back to kinda, uh, compounding and, and leverage. So one of the things I've learned, you know, across, um, every company I've, I've, I've done and worked in is that, you know, software, if you can kinda get those core primitives right, um, and then what you end up doing is sort of, um, remixing and recomposing those, uh, primitives in kinda novel ways based on how the market evolves and how the product evolves. And so that part has been true at HubSpot since the beginning, right? So the thing we built out was we had this kinda core, uh, kinda operating system or operating framework that we had on which we built the marketing automation and things like that. So when we went to product, you know, two, three, four, whatever, like most of the value that we needed to bring to market already existed in the form of primitives. Like for instance, we started with marketing automation. We just, "Oh, here's this deterministic workflow to do marketing auto-- like send an email out, wait three days, wait for them to click." It's like, oh, well then when we got into sales and CRM, it's like, well, our marketing automation primitive is actually a primitive. We can just reuse all of that technology, and we will have more on day one of our sales automation app or product than anyone likely has that's been at it for a few years, right? Like, and s- and so that part of it, I think, still holds true. Um, and so what we're doing now is to say, okay, what is the framework we need for the next 20 years of HubSpot now knowing that we have this AI world? And so you touched on one of the things is, oh, well, the data is no longer just this kinda structured data in a relational data store or whatever, which made sense for the kinds of queries, uh, deterministic ques-queries you're doing. Now it's both structured and unstructured data, and it's not like this data didn't exist within the organization, even within our customers' organizations. Like you just couldn't do anything with it, uh, that was meaningful, uh, at any sort of, uh, well, at, at any price actually back 10 years ago. But now you can, right? So, so now we revisit the primitives and say, okay, well we still need a structured data store. Um, and I will, uh, you know, happily debate this with whoever wants to debate it.

    13. JS

      [chuckles]

    14. DS

      Like there is still room, whether it's relational, some other... There needs to be some persistent amount of mechanism for storing kinda known facts, uh, and data, some sort of system of record, if you will, uh, alongside, uh, this kinda system we think of as a system of context, which is this unstructured data. And that's the thing that, um, you know, we're sort of building out. And the, the way I think this is going to evolve is, um, this notion of, um, and I use this term, uh, others use it too, like a, a second brain, uh-

    15. JS

      Yeah

    16. DS

      ... or like an open, like a... Can we put all the things that are happening, um, and, and the, the kind of the structure that likely makes sense, um, is some sort of graph structure. Can we store this in some sort of, uh, context graph? Um, another term that's, um, has been used. And, and that's the idea. It's like, okay, well, can we take this and then use that in combination with the structured data store and then draw inference and drive agents and drive outcomes and, and, and get work done?

  11. 40:44 – 43:47

    Replaying 20 Years of Code With AI

    1. DS

      So, uh, one thing that's interesting, uh, in terms of AI, you didn't ask this question, but, uh, it occurred to me as, as we were chatting, is that one of the things that's really beautiful about, uh, agentic coding now, um, is, and this is particularly true for, actually even more true for incumbents, it's like, yes, it allows us to produce software faster. We can test ideas faster, which we've talked about before, build products faster, which is all awesome. Uh, we benefit from that. Uh, but the other thing you can do is you can actually go, um, deeper and do larger projects that you might not have otherwise considered. So what's possible now, um, is that, you know, so we have a system of record for code, which most people, which is GitHub. So imagine you said, "Okay, well I have something like a, you know, a Fable 5.1 or like, um, 6 Astra, um, that's out now, or 6.1 that's out now." Um, but so imagine you say, "Okay, well, you know, I wish like younger me had made these choices when it came to... If I just could go back in time, here's how I would've done those primitives, or I would've introduced this as a primitive that's a reusable service versus because we didn't know that then, but we sorta know that now." In theory, it's actually possible, um, to now say, "I wanna, I wanna play that back. I wanna take this body of code that we used this data structure for, this language for, this framework for, this architecture for." And what would happen if we said, "We're gonna spend $200,000 on tokens to go just redo that," and basically just play that. It's like a, you know, infinite possible universes thing or let- let's jump to that branch of the universe where we have made a different choice and then see what happens, where do you wind up. And one of the advantages, um, you know, companies like HubSpot have is that we have, uh, what people used to call unit tests, uh, and now show up as evals. It says, can we actually measure the viability of the thing that we just did? Uh, does it still pass the test? Does it still produce the outcomes? Does it still do all the things that it needs to do? Like, we have that now, so we do, we run this test now when new models come out. It's like, okay, well, we can run our battery of evals across, uh, new models and say, "Okay, well here's a trade-off. Here's the economic value. Like we lose about 3% of quality points, uh, but we gain this much in terms of latency or, or economic." We can do the same thing with testing these kind of theories. Like, okay, if we did rewrite this whole thing into Rust, like I'm making that up.

    2. JS

      Yeah, whatever, whatever.

    3. DS

      What, what would it look like? What, what, what's the actual... And you can, you know, test it, uh, now with small bodies of work. So anyway, it's, it's, it's exciting, uh-

    4. JS

      No, I-

    5. DS

      Yeah. The kinds of things you can do.

    6. JS

      I think this is such a fascinating topic. I, I wrote a different essay, um, I think right around the same time I wrote "The Yellow Brick Road," maybe a little bit before, and I actually used Workday as an example, uh-

    7. DS

      Yep

    8. JS

      ... on this front of just like, hey, you have to kind of start over to think about being AI native, and here's all the, here's all the advantages you have, but here's all the disadvantages. And so-

    9. DS

      Sure

    10. JS

      ... you know, it's, it's almost, you know, the, the benefit I think that HubSpot has is it has, uh, an amazing, you know, founder who's still very, very involved in the company, of course, running the business and, and driving these changes. And so th- that... I, I love that example of, like, trying to think about these, like, alternate universes that you can build.

    11. DS

      Yep.

    12. JS

      Are there other processes that you're trying to create inside of HubSpot, or teams you're trying to create in HubSpot, or, or-

    13. DS

      Yeah, so-

    14. JS

      ... or destroying bureaucracy? I'm just curious to hear what you're doing as a, as a founder right now.

  12. 43:47 – 46:38

    HubSpot Next and the Birth of uSpot

    1. DS

      Yeah. Uh, so we have a, a group within HubSpot, um, called HubSpot Next, and the idea is it's, it's, uh, um, effectively an innovation lab, so it's a, it's a startup within a startup. Um, and so we'll have any number of these kind of going on at any particular point in time. And the motivation, and we really truly do treat it like a startup, and we, and we did a variation of this, uh, when we built the CRM. It's like, okay, we're gonna build a CRM thing. It's like, wait, what's that? We're in the marketing software business. And it's like, and we sort of have cordoned off the team. Uh, they had none of the rules. You didn't have to build it on our stack. You didn't have to use the brand. You didn't have to do anything, just go solve the problem. Um, and we sort of carried that idea through now, and so, um, and the, and, uh, and I've championed, uh, and led, uh, several of these projects over, um, over the years. And the one I'm working on right now, literally every, every single night, um, is a project called uSpot, um, and it's a one-person CRM. It's a CRM for solopreneurs. That's the market, and it's going back to the, uh, kind of narrow your, um, aperture of, uh, focus. But the HubSpot Next projects have, um... they're pursuing either a new product category or new product idea, pursuing a new market, or testing a new technology. And going back to the can you learn your lesson cheaply, it's much easier for me. It's like si- I have this idea for a second brain, right? Like, oh, maybe a graph th- you know, graph model works better. Maybe you plug in all the social media and everything that's coming in or whatever, and it's like, okay, well, let's test this in, in uSpot because that's a much easier f- kind of faster paced way to test it. Um, and so the product, uh, now is exactly 30 days old, uh, so-

    2. JS

      Congratulations.

    3. DS

      Yeah. And, and, and, and it's, and it's been fascinating, um, kind of learning. So the answer to the, the original question is, you know, can something like that work? The answer is yes, um, it, it does work, and I think, um, other entrepreneurs should kind of take away from, um, you know, this is that. So that, that notion of having a second brain, um, that holds, um, and it... The thing to, going back to the kind of Yellow Brick Road conversation and kind of avoiding b- uh, being run over, the thing you shouldn't do is to say, "Oh, well, ChatGPT and Cloud Core work are really bad at memory or whatever," and this, like-

    4. JS

      Yeah, it's like it's gonna get better.

    5. DS

      Don't solve that problem. Don't solve the horizontal use case or whatever. But say, "Hey, within this class of customer, here's a class of problems that they have that would be benefited by thinking about this-"

    6. JS

      Yeah.

    7. DS

      "... notion of having a second brain." What would the ingestion points be? How do you pull it together? What's the actual agents that you would produce or the work that you would kind of, sort of accomplish? But I think that underlying primitive, um, is actually a powerful one. I think it, it, it applies, um, so.

    8. JS

      Yeah. And then now that you... I mean, I, I think uSpot's so fascinating. I think it's, uh, it's very valuable for that, for surely for this type of user. I mean, it's, it's what they, they want, of course.

    9. DS

      Yeah.

    10. JS

      Um, how do you then take those learnings and, like, productize them at the mothership, if at, if at all, and, and what processes do you need to put in place to, to make that happen?

  13. 46:38 – 49:43

    A Startup Within a Startup

    1. DS

      Yeah. Um, the way we've done it, which w- you know, we've kind of stumbled over the years and tried various mechanisms, um, is to kind of pull on that original thread of, um, this is a startup within a startup. So we do a couple of things. Um, one is if the thing gets to scale, uh, and we've had this happen, you know, m- multiple times before. So there was a product I launched a couple years ago called, uh, Agent.ai.

    2. JS

      Mm.

    3. DS

      Uh, and Agent.ai was a marketplace for AI agents.

    4. JS

      Yes.

    5. DS

      Um, you know, back, back in the day. This was one of my accidentally competed with, uh, Sam Altman, uh, stories.

    6. JS

      Right. [laughs]

    7. DS

      Although it was focused on go-to-market, it was, uh, too horizontal. Uh, but there were two parts to it. There was these kind of agent AI marketplace, and there was an agent builder that says, "Oh, can we let mere mortals actually build these, uh, kind of-

    8. JS

      Right

    9. DS

      ... mix of deterministic, non-deterministic, um, agents?" And what we discovered is that, uh, half of the idea was right, uh, but actually belonged and needed, uh, the full HubSpot team behind it. So we did, uh, effectively an acquisition, right? So it's like we're gonna take that, and now we're gonna make agent builder a product within HubSpot or whatever, and that's what it lives, lives at right now. So in the uSpot case, right, like, uh, one possible outcome is, oh, like we're building the CRM for, uh, one-person companies. We got these things right, those things wrong, whatever. Maybe we just absorb it into, uh, HubSpot, becomes a product of HubSpot. Um, maybe we say, oh, well, the product itself still probably needs to be standalone, or maybe the market doesn't work. Uh, but the idea of a second brain maybe works and, like, and here are the things and the ingestion points that we think work for SMB, uh, and, and multiplayer kind of team-based mode or whatever. We just sort of let the core team decide what they want to pick and choose. It's-

    10. JS

      Yeah

    11. DS

      ... it's a little bit like they're looking, like, at, at an acqui-hire, and it's-

    12. JS

      Yeah

    13. DS

      ... literally like that. It's like, oh, here's what we can buy, um, at, at the low, low price of it's included, right?

    14. JS

      It's included. Yeah.

    15. DS

      [laughs] It's not, it's not free again.

    16. JS

      Already here.

    17. DS

      It's free. [laughs] Ex- exactly. It's free.

    18. JS

      Yeah, yeah.

    19. DS

      Um, so yeah, that's... And, and we sort of try to have an open mind about it, but the nice thing about doing it the way we do it is that... So you think it's like why can't you just do it within the, you know, the, the core, core business is that those degrees of freedom are super important to be able to run the experiments at the rate that you want to r- run them. So you have to... Like, one of the things I've, you know, long said, you know, being a startup guy is that, um, the, the issue that incumbents sort of... [sighs] Hard to use, uh, incumbent when I refer to HubSpot now. I still think of it, uh, in some ways as a startup, but-

    20. JS

      It's still a startup. Don't worry

    21. DS

      ... yeah, um, is that-

    22. JS

      You're, you're still building

    23. DS

      ... it's not the absence of something. It's not like startups, you know, have something that, um, that the larger companies don't. The issue is that- Is the kinda, I'll call it baggage, um-

    24. JS

      Yeah

    25. DS

      ... and the weight that the organization has, and if they can find a way to kinda shed that, then you get all the advantages of being larger. You have distribution, you have customer relationships, you have evals, you have-

    26. JS

      Yeah

    27. DS

      ... you know, 20 years of a code base that you can use as data to kinda inform the next product that you build. All this stuff is, is actually an asset, so if you can sorta shed the liabilities, um, and, you know, HubSpot Next is a way to say, "Okay, well, we're just gonna pretend that we have none of those liabilities," which we don't. As a startup with a startup, we can just go off and do our thing, which we've been doing. Um, but then when it comes time to lean into the asset, it's like, oh, it would be helpful to have the HubSpot brand here, or can we get some advice on distribution over there? It's like,

  14. 49:43 – 53:07

    The Rise of the One-Person Company

    1. DS

      we can do that, so.

    2. JS

      Yeah. It's so easy right now e- especially with some of the, the, the, the national discourse to, to think negatively about AI, about AI and, like, the impacts, but the reality is all the data that I've seen are so-- it's so exciting. Like, growth is unbelievable.

    3. DS

      Yeah.

    4. JS

      Um, and I actually think uSpot is a really good example of, of this potential growth, right? It's like when in the, in the history of mankind would it have been, like, a good idea to go build, like, this, like, super individualized CRM for one person? Like, it would've been almost-

    5. DS

      Totally

    6. JS

      ... impossible to do that-

    7. DS

      Yes

    8. JS

      ... historically, right?

    9. DS

      Yeah.

    10. JS

      But now we're able to do that, and-

    11. DS

      Yes

    12. JS

      ... what is that then gonna let those one-people companies do? Like, that is-- We're gonna see growth unlike, I think, anything we've ever seen before, and it's just so cool to dream. So maybe-

    13. DS

      I agree

    14. JS

      ... I'd love to just hear you talk a little bit about what you're seeing, like these initial use cases maybe from uSpot or, or what-

    15. DS

      Yeah

    16. JS

      ... you're seeing from, from growth. It's just, it's, it's, it is a really amazing time to be alive.

    17. DS

      Yeah. Um, so it's, it's, what's interesting, like the, uh, I knew this to be the case, but I wasn't sure to what degree. So a-as I mentioned, the one thing that I did differently with, uh, uSpot, um, versus prior kinda HubSpot Next projects that I've done is I, uh, I monetized really early. Um, and part of it's not to kinda cover costs or something like that. It's, like, that's one of the lessons I wanted to learn is I wanna learn, like, the pricing lessons and some of the things that we wanna learn more about kinda monetization pricing and what, you know, value do people ascribe to this. Um, that was part of it. So the thing that I've been really surprised by is, uh, willingness of, for people to put a credit card down on a consumption-based model. Um, like, uh, it's like nothing... I've, I've launched lots of projects, um, you know, over the years at HubSpot. This has been, in that regard, the most successful, so there's, like, this signal, um, out there. Um, and part of it, I think, is just driven by, you know, the sheer numbers, you know. We'll bring the, the US Census data into this right now is that, so in 2025, we had the most number of new company formation that the United States has ever seen. I think it was, like, 5.6 million, closing on 6 million, uh, which is great. Uh, this year, we-we're already at four-- like, by August, um, we were already over 4 million, so it's gonna be another record year. So it's not like, oh, we had this one-time spike, and now it's settling back in. So that's kinda fact number one that's super interesting. Uh, fact number two that's even more interesting is the percentage of those new company formations that are expected to remain one-person companies is the highest it's ever been, and it's increasing. And so the US Census Bureau actually tracks this data. They say they put every new company into two buckets, and the, one is that this company is predicted to h-hire an employee within, at least one, an employee within the next two years, and they bucket them in those two ways. And the bucket of they are unlikely to actually hire an employee in the next two years, uh, that percentage is now, it's, yeah, over 70%. Um, so, so it's exciting. So there's this new market, um, and so for the entrepreneurs that are out there, there's a entirely-- And just like SMB was back when, you know, and everybody sorta, uh, pushed back on HubSpot back in the day. It's like, "Oh, you can't make money in SMB. It's so hard." When no software company has ever, other than Intuit, has ever made a multibillion-dollar business in SMB. Um, but you can. It's, like, it's big enough. Yeah, there's gonna be painful parts to, uh, you know, those kinds of markets. But anyway, I'm, I'm excited for that. I'm excited for, excited for the economy. I'm excited for all the entrepreneurs that are gonna be starting these millions of businesses or whatever. Um, yeah, and, and then excited for the entrepreneurs that build things for the, for that market. Just, like, it's exciting all the way

  15. 53:07 – 56:06

    AI Is Still Just Software

    1. DS

      around. So I'm, uh, yeah.

    2. JS

      No, I totally agree with this, and, and we, we talk a bunch about, uh, you know, aspects of this a- all the time, and I think one of the, one of the great areas that we talked about it, it has been like, oh, well, vertical software in the past. Well, how did you make that work? Well, you made vertical software work with a low, you know, a-ACV or, uh-

    3. DS

      Yeah

    4. JS

      ... annual or monthly, monthly cost. But then you made it up because you had payments, uh, you know-

    5. DS

      Yeah

    6. JS

      ... inter-interchange revenue.

    7. DS

      Yeah.

    8. JS

      And, and I just don't know if people are fully appreciating the increase in, in opportunity size, as there's just so much more you can do, and-

    9. DS

      Yeah

    10. JS

      ... and so many oth-other ways you can, you can offer value and, and create value for your, your, your small customers, so.

    11. DS

      Yeah. And this is-- Yeah, and this is my, uh, kinda counter to the, sort of the S- the, the SaaSpocalypse, um, kinda idea is that the reality is that, you know s-- And AI, by the way, f-for the record, I know you know this, it, it is software, just so you know.

    12. JS

      Yeah. Yeah, yeah. No.

    13. DS

      It's not like it's not software, right? It's not-

    14. JS

      Yeah. It's, like, not, it's not an energy form.

    15. DS

      It's not a new energy source. It's, it's software.

    16. JS

      Yeah, yeah.

    17. DS

      Um, not really smart software, but software.

    18. JS

      Yeah.

    19. DS

      Uh, and what, what has really happened now is, like, if you look at the class of problems, um, that software, uh, was designed... Okay, I'll, I'm gonna, uh, I'm gonna take one minute because this is actually a good story. Uh-

    20. JS

      Yeah

    21. DS

      ... I only have a handful of good stories. So when we were first starting, uh, in the first year of HubSpot, and we were out raising venture capital, we said, "Hey, we're building this marketing software company. Um, here's what we're gonna do." Um, and, and the, almost universally across all the VCs that we talked to came back and said, "Well, there's not been a marketing software company in the billions or whatever." Um, like, that, that's, uh, y- so we had two strikes, like, the SMB strike and the fact that we were doing marketing software.

    22. JS

      Yeah.

    23. DS

      And our argument at the time was the reason... Okay, so if you look down the P&L of any, uh, any sized business, every line item on the, on the P&L has a hundred-plus billion dollar software company name attached to that line item-

    24. JS

      Yeah

    25. DS

      ... except for marketing. It's not like people don't spend. And the reason is, is that marketing historically has been an arts, craftsy kind of thing or whatever. It's like, yep, you have this amount of spend. You spend a little bit on, on creating the creative, but then you just spend it on ads and billboards-

    26. JS

      Yeah

    27. DS

      ... and all those kinds of things. Fine. And our thesis was that as a result of the internet, marketing the industry is going to look a lot like a problem that software was designed to solve.

    28. JS

      Hmm.

    29. DS

      That's why a software company needs to ex- didn't need to exist five years prior to HubSpot starting, but now someone will build a massive, uh, you know, marketing, uh, marketing software business. I'm gonna make the same argument now that before there was a class of problems that software wasn't really equipped to solve.

    30. JS

      Yes.

  16. 56:06 – 58:51

    Lightning Round

    1. JS

      agree with you. This has been, uh, unbelievably fun. I'm, I'm, I'm so grateful for your time.

    2. DS

      Yeah, thank you.

    3. JS

      I'm gonna end with a quick lightning round for you.

    4. DS

      Sure.

    5. JS

      Then we can go on, on our merry ways.

    6. DS

      I'm not very fast on my feet, so, uh-

    7. JS

      Yeah. Well, there's only-

    8. DS

      ... I'll hold you to your best, but-

    9. JS

      There's only a few. These are softballs.

    10. DS

      Okay.

    11. JS

      Um, I'll start with a very easy one, I, I, I would assume. Dunkin' or Starbucks?

    12. DS

      Uh, Dunkin'.

    13. JS

      Okay.

    14. DS

      But I don't drink coffee, by the way.

    15. JS

      Oh, wow. Okay. Well then, uh, what, what do you drink instead? Just only water? Are you a cyborg?

    16. DS

      Uh, just, just wa- Yeah, usually.

    17. JS

      Wow. Wow.

    18. DS

      I gave up caffeine, like, years ago, so anyway, that's, uh-

    19. JS

      Even more impressive-

    20. DS

      [laughs]

    21. JS

      ... Dharmesh. Okay. Um, uh, Instinct or Muse?

    22. DS

      I'm sorry, what?

    23. JS

      Instinct or Muse, you know, on the new, uh, the new agentic shopping experiences?

    24. DS

      I, I'm gonna have to say Instinct.

    25. JS

      Okay. Interesting. It, um, if you had to be an animal, which type would you be?

    26. DS

      I would still be a mammal. Maybe an elephant.

    27. JS

      An elephant. I like that. Okay, good. What's the most Boston thing about HubSpot?

    28. DS

      Um, almost everything. [laughs] It's, uh-

    29. JS

      Good, good, good answer. [laughs]

    30. DS

      Yeah. Um, we're very pragmatic, solve for the customer. Uh, it was... I, I say that. Um, we're sort of like a hybrid between, like, a, some of the best of the kinda West Coast, uh, kinda valley thinking in terms of, uh, ambition, aggression, but some of the pragmatism of the, the East Coast, uh-

Episode duration: 59:06

In the uListen app, the transcript lights up word by word as you listen. Get uListen

Transcript of episode s4lXP9IuHjk

Get more out of YouTube videos.

Read along on your phone, search it beside the video, or ask your favorite AI app what was said.

iOSAndroidClaudeChatGPTChrome