20VC: Cognition CEO Scott Wu on Acquiring Windsurf: The Process, The Deal, The Rationale | Did Google Overlook a Goldmine in the Core Asset and Did Founders Leave a Sinking Ship | How Cursor and Cognition Deal with Ever Increasing Reliance on Anthropic

I think there's an unspoken covenant that as a founder, you go down with the ship for better or for worse. It's, it's changed a bit over the last year, and I think it's a bit disappointing, to be honest. I think there's some real truth to your point that often there are actually a lot of really valuable pieces that, that get left behind in AI code. To be truly honest, if there was zero progress, the world would still be entirely different.

This is 20VC with me, Harry Stebbings. Now, you want to know how real-time this episode is? It was recorded about 12 hours before you're listening to it. It was recorded late late on Thursday night. And yes, 20VC team worked through the night to edit it and turn it around. That is commitment for you. The only thing that anyone wants to talk about this week is Cognition buying Windsurf. What happened? Did the team get screwed over?

What happens from here?

How does it fit into Cognition's grand master plan? What does it mean for the space in terms of who wins? Cursor versus Cognition versus Claude Code. We unpack it all today with Scott Wu, co-founder and CEO at Cognition.

But before we dive into the show today, Harvard Management Company is constantly seeking out the next generation of great investors and entrepreneurs. HMC has managed Harvard University's endowment for nearly 50 years and was one of the first institutional investors in venture capital. Their experience and long-term investment horizon makes them ideal partners to get world-changing ideas on a path to viability and success. They work as a true partner, providing insightful perspectives to help managers succeed. I personally have had the pleasure of working with the HMC team and can say that they are truly exceptional partners and savvy investors. So whether you're launching your first fund or your fifth, HMC welcomes the opportunity to partner with both developing and established managers. Have an idea you want to share with the team? Just send it over to venture@hmc.harvard.edu. And after securing top-tier partnerships like those with Harvard, the next step is building trust with customers, and Secureframe makes that simple and seamless. Secureframe empowers businesses to build trust with customers by simplifying information security and compliance through AI and automation. Thousands of fast-growing Businesses including NASDAQ, AngelList, Doodle, and Coda trust Secureframe to expedite their compliance journey for global security and privacy standards such as SOC 2 and ISO 27001, CMMC, NIST standards, and more. Backed by top-tier investors and corporations like Google and Kleiner Perkins, the company is among Forbes' list of the top 100 startup employers for 2024, G2's Best Software Awards for highest satisfaction products, and a recipient of the 2024 Cybersecurity Excellence Awards. Something I I definitely never got in school myself. Learn more today at secureframe.com.

You have now arrived at your destination.

Scott, I am so excited for this, dude. I was literally just telling you, it is a cool prep when you get to call up like Vinod Khosla, Joe Lonsdale, and you call them and you're like, hey, what should I ask? And they're just like, ah, I love this guy. So thank you so much for joining me.

Yeah, thanks for having me. It's been a crazy few days for us all.

It's been pretty insane.

I just want to start present day and work our way around.

How did you learn about the potential ability for you to acquire Windsurf and how did that opportunity arise?

Yeah, we found out honestly on Friday, the same time everyone else did, that all of this was happening and that this was the split and that it was Google and, and here's what happens next and here's what happens with the team. And we were kind of talking about it, which is, you know, on the Devon side, I think in a few ways it seems to be a pretty natural fit. One, because I think obviously there is an incredible you know, team left behind. And if anything, I think what we have at Cognition, I would say, is especially focused on the core engineering and product team, whereas obviously, like, you know, I think Windsurf has built an amazing go-to-market team, marketing team, finance, operations. And then similarly, in terms of the products, you know, we found that actually it was— there's a very naturally complementary lean. And so we reached out cold Friday evening and our first conversation was Friday night.

Just take me to it. Then you reach out to Jeff and you're like, hey, figure something's going down, should we jump on a call?

Yeah, yeah. And so, I mean, we basically said, look, it seems like there's enough here that it's worth having the conversation and talking about it. We got on the call together. We were talking about, you know, the ways to partner and what things would look like. I think what we came to pretty quickly— I always think about these things in terms of what is the correct way that things should happen here. And I think from Windsurf's perspective, the way that we kind of saw it was You know, you could talk about the product of the team and the business, right? And what is the right thing for each of those? And I think for the business, the thing that you have to call out is sooner the better. You could do this and you could have like, you know, months of diligence and everything. And that has apparently happened multiple times already, you know, in Windsurf's company history. And I think in this particular case, I think it's everyone is scrambling. You know, all the customers want to know what's going on. The whole team wants to know what's going on. Everyone in Silicon Valley is reaching out to them saying, hey, like, I heard the news, you know, you want to come interview here. And so, you know, we kind of just came to this conclusion, look, a month long is not the way to do this. I mean, even like a week long, you know, it's, it's, it is Friday night now. We want to have something ready to go by Monday. I'll pause because it sounds like you have that. Yeah.

I'm just like, when you look at like scale, it's like, you know, there's a husk left behind which kind of deteriorates and you're seeing that deteriorate in real time. Why did you believe that this was inherently different and that the husk was not a husk? It was actually a treasure chest of value. In a way that was different from Scale.

Yeah, so there was definitely this conversation online, right, about all the best researchers left and there's just nothing left behind. And I don't think so is basically how we felt about that. And we looked at it and I mean, it's an amazing product that a lot of people use. It's the entirety of the customer book as well. It's all of the code, a lot of the data and the proprietary IP. And then obviously it's a really incredible team. I think it's an interesting state in the sense that you could call it perhaps like a somewhat incomplete piece. I mean, as it turns out for us, I mean, we had the very nice complement of what were the pieces that were needed, right?

Those pieces are the rationale that lead to the decision.

How do you actually execute it?

How does that work in process?

Yeah, yeah, for sure. So we came to them and we said, look, I think there's a very natural partnership here. I think the team pieces are very complementary. I think there's a lot we can do on the product. The thing that just seems, you know, absolutely clear from both of our sides is if we're going to do something, we want to be ready to do this and announce it by Monday morning. You know, it kind of reminds me of like the bank goes into receivership Friday night. You got to have an answer, you know, by Sunday and by Monday morning. Right. And I think we thought of this as a similar situation, which is, look, there's a ton of value here, but everyone's wondering what's going on and we need to be able to give them a clear answer as soon as possible that we are going to do this. We're going to support, you know, the product. We're going to, we're going to make sure things, everything's going smoothly and running smoothly and so on. And so that's what we came to. And we gave them our honest take of like, here's what we think we can do together. We are not going to have time to get to go like as deep into the diligence and details as we would like to. But, but I think at a high level, we at least are in a position to understand the business very well, get to a verbal agreement on Saturday, hash out all the details and the terms and the legal on Sunday. Get everything signed and done on Monday morning, and then we're ready to announce. Like, that's— there's no other way to make this work out. And Jeff and Graham and Kevin and, and the whole crew, honestly, like, I mean, it's a huge shout out to them because they really, they really worked around the clock for, for their team.

Were they left in the lurch by this deal?

Yeah. So I think they had known about— I think they had known about the previous thing, you know, maybe a few days in advance. It was not a ton of, uh, advance notice. And I think they were honestly really thoughtful with how they went about it, which is, look, here are the options. We can operate as an independent company. We can go and raise a new round of venture capital now that there are actually no investors, or we can find someone who we think are going to be the right partners to work with. But either way, you know, it's going to be— we're going to have to do this quickly and we're going to have to figure out what we think is the right long-term fit for it.

I think most people see this as a really epic deal make from you, respectfully. And I mean, total hat tip and respect to you for it. The question that I have is, and actually I think it was Ruchi at South Park who asked this, but she said, was this a Google blunder by leaving such a valuable asset up for grabs and not seeing the true value also in the asset?

Yeah, I think there's some real truth to your point that often there are actually a lot of really valuable pieces that, that, that get left behind.

Given that you've been through this deal structure that's different and novel, do you think this becomes a new norm for how companies are acquired and how people join companies today?

I, I, I, I think there's an unspoken covenant that as a founder, you go down with the ship for better or for worse. It's, it's changed a bit over the last year. And I think it's a bit disappointing, to be honest.

Is the talent war getting a little bit out of hand? When you look at Meta's hiring spree and just how much money— is it getting out of hand?

Yeah, maybe crazy opinion, controversial opinion, but I actually think it's, it's, it's, it's quite reasonable, is my view. And the reason I say that is because AI, I think we are truly on the cusp of the greatest technology shift in our lives. That's already clear. Even if you froze all the capabilities today, you know, and then the question was just if you said that there were no new research breakthroughs, no new discoveries, you know, you don't even have to bet on that. And you just said, all right, we're going to, you know, grow products and figure out how to build the right experiences and get those out to users and meet users where they are in all these different verticals all over of the world, I think it would still be the biggest thing. I think the biggest thing, in my mind, the biggest technology shift in our lives has been the internet. I think there are some others which are obviously pretty massive as well, the mobile phone, the personal computer, and so on. But yeah, I think even if you froze all the capabilities today, I think AI already would be bigger than that. And the thing that's crazy about AI, and by the way, I don't think it's gonna freeze, I think it's gonna keep moving even faster. But the crazy thing about AI is there really are so few people that are really just determining the trajectory. Of AI. But I think in aggregate, I think the view is right, which is, you know, there's so few people that matter.

Can I interrupt you?

I'm so sorry. I totally agree in terms of what—

is it like 100?

Is it like 10,000? Just what quantum is it?

It's a great question. Somewhere in between those two is right. There are at least 100 folks that really matter in making a lot of differences. Certainly less than 10,000 and probably a good bit less than that, if I had to guess.

Okay, so we've got like 100 to 1,000 is kind of a range there. Do you need those people unless you're working on the most cutting-edge frontier models? If you're in the application layer, which is very hard, I'm not denying it, but do you need them if you're in the application layer?

I think the level of talent and the fierceness of competition in many of these is actually greater even in the application layer, which is crazy to say because I think the competition in the foundation labs is extremely strong.

When you think like, holy shit, anything can happen style, which it is today, one of the like, oh, shits that could happen that's very obvious to I think me and everyone is like the dependency now around Anthropic. Obviously, Windsurf, when I had Varun on the show, they were very much reliant on them then, and then they got cut off when the OpenAI deal was done. Is the reliance on Anthropic greater than it's ever been?

Yeah, look, we work very closely with Anthropic. I think Anthropic is a great company, and I think the Foundation Labs in general are great companies. People always ask this question. About, you know, where does the value accrue in AI? Does it— is it chips? Is it semiconductors? Is it foundation models? The application layer? Is it the infra layer? I think the boring but true answer is it accrues wherever you are able to establish real differentiation in your space. And I think there will be a lot of spaces for which that's the case. I think the foundation labs as businesses will do extremely well. I think the thing that we own is and that we spend all of our time on, I think, is actually a nicely complementary piece, which is in some ways really thinking about how to best optimize for very particular capabilities within software engineering and how to deliver and ship like a really great product experience around that. Right.

Do you not want though the developer products for these foundation models to commoditize? Because then you gain leverage. If Anthropic's Claude continues to be so much better than it is, their leverage is gone. Do you not want that commoditization?

Yeah, look, I mean, I think it's, uh, To be honest, I mean, I think whether we want it or not, it seems like folks are pushing ahead in the foundation model space and making a ton of progress. We've seen a lot of launches even in the last week or two, right? I mean, there's Kemi and Groq and a ton of great progress that's been happening. And I think that is the natural way of things, which is, you know, there will be competition in that layer. And I think some folks will, you know, especially in particular verticals and use cases, be able to establish differentiation. And there will be competition in our layer, and then some folks will be able to establish differentiation. As this equilibrium develops, I think it is just naturally the case that folks in these layers want to collaborate and figure out the right ways to work together, as long as that holds.

Why would Anthropic want to collaborate when they could just own it? Sorry, I'm a VC, so I'm naive.

No, no, of course, of course. I love it. Hitting me with all the hard questions. Look, I think the answer at the end of the day What we focus on is very, very different. The kinds of questions that we think about, for example, on both the Dev and the Windsurf side, by the way, is how should humans and AI work together to produce code? It's one thing to just kind of like solve for generally smarter and smarter base models. It's another thing to teach your specific model, all right, here's, here's how to go, you know, to Datadog and pull up the logs for this thing. And here's how you debug a frontend live. And here's what the— here's our representation of the codebase, which we're learning and iterating over time and so on. There are a lot of different verticals to work in and to solve for. I think we have parts of this equation, and I think the Foundation Labs have parts of this equation. But I think the truth is you really need both.

You prefaced that one with, you can ask the hard questions. So fuck it. It's the end of the day in the UK, so I'm using that as an excuse. What percent of your revenue goes to Anthropic, do you think? Ballpark?

I have to pass on that one.

It's like with children, you know, where they'll push and they'll see how far they can go. Yeah. When you speak about that progress and the incredible progress that we've seen, some think that it will take the same route as self-driving cars, where there is this kind of plateauing, and then we'll see again another inflection point. What gives you such confidence that we will see the continual progression in models that we've seen over the last 12 months over the next 36 months?

Yeah, so I'll give you two thoughts there. The first thought is there's very strong signs of this continued progression, largely because, by the way, there's, there's a lot of specific work, which is you can call it research, you You can call it engineering, you can call it infra or whatever you like, but there's a lot which is roughly that we know the techniques and we have a lot more to do to go and scale those. It's RL, for example, is I would say the biggest breakthrough of the last year and a half, call it. It's crazy to think that you roughly can solve any benchmark. It sounds insane to say, but that's really what RL is converging on, which is basically if you have a clean enough set of here are exactly the behaviors that I want, Here are the environments that you need to be able to operate in. Here's what it means to succeed or fail. You can just train a model that does that. Naturally, that is just such a powerful capability, which I think is going to be applied to more and more spaces. I think we will see that progression. The other thought I would give though, and I mean this truly sincerely, is there are a lot of spaces and a lot of different things going on in AI. In AI code, to be truly honest, if there was zero progress, the world would still be entirely different. You know, I think there are a lot of spaces that are early today that you see AI making progress. It's getting better and better and you're going to the next step and the next step. In code, you are just slower as a software engineer if you're not using AI. That is the truth. And it is a no-brainer already today. And look, I think it will be more of a no-brainer and I think we'll be able to make engineers even more efficient and, you know, be able to do even more with code. But I think that is already the case.

So many things to unpack there, dude. Do you think tools like Davin and others make 1x engineers 10x engineers or 10x engineers 100x engineers, if I were to put you in one camp?

Interesting. It actually really does depend tool by tool. I do think that the product experiences for the 1x to 10x and the 10x to 100x, or even the 0x to 1x or whatever you want to call it, are somewhat different project experiences.

I've interviewed Benioff at Salesforce. I interviewed Vlad at Robinhood, both very recently, and they both said that 50% of their net new code is created by AI. Would you agree with that in what you see with your customers and peers, one? And then where will that be in 3 years' time?

Yeah, absolutely. People often talk about the percent of code that's written. I think the obvious thing to call out is, well, it's a little odd because for one, you have to factor in how important each line of code is, right? If it's just like a ton of Protobufs, then that's one thing versus a lot of the core business logic is another thing. And then two, of course, is written with the help of AI is one thing, but how much— and so And for those reasons, we often like to think about in terms of like how much faster is an engineer using all these tools, you know, and an engineer using the best AI tools and, you know, who really understands how to get value out of them, how much are they doing in 1 hour versus how much they would do in 1 hour with no access to AI? And I would say that something in that range of 1.5 to 2x feels right to me today in aggregate. Look, I think in 3 years there's no reason that shouldn't be a 10x. The thing that's really fun, by the way, is I think we're going to have more than 10x more code. It's something I've been saying recently is I keep track of every time in my daily life where software fails me. We've become conditioned to be okay with it, but the truth is it happens all the time. You know, you think about products out there and, you know, the way I like to say it is there are, I say, this top tier of products which are in some sense like I'll call them the, the best made products in the world. And I'm thinking of like YouTube, TikTok, Instagram, and so on. And you can really feel every little detail that they've done with a ton of care. You know, it's, they're streaming in tons and tons of data and it's always super efficient. It never goes down and it's super reliable. The algorithm basically knows you better than you know yourself. And that is like, call it like hundreds of millions of hours of engineering time, you know, that went into building that, that piece of software, right? Then you kind of go down to the next layer of the next order of magnitude of software that has tens of millions of hours spent, and then single-digit millions and so on. You see the differences really quickly. I mean, when you're logging into your bank or when you're dealing with your healthcare, working with your insurance and trying to get things going, or when you're trying to go and navigate your customer list or things like that. The truth is all the software can be 10x better, and I think there actually is 10x more of it to write. I think that's one of the fun things in code, right, is that it really does have this Jevons paradox.

When you think about that developer efficiency, what are agents not able to do today that they really need to be able to do?

So I think there is a real point at which agents are truly able to take over ownership, I guess is how I would describe it, of the work that is done. And the way I kind of want to say it is like, look, I think in the future we will get to a point where you're not looking at your code, you know, you're looking at your product, right? And you're looking at your product. And at the end of the day, code Software engineering, this whole thing is just telling your computer what to do. Code is the language that your computer happens to speak, and that's why we all have to learn how to write code in order to do it. I think over time, yeah, you get to the point where you can just say, "All right, yeah, this website, let's add a new tab here and let's put this and that information, and maybe let's go collect this info from the user and we'll save it in the database this way. This button should be a little bit rounder." You're just able to make those calls and make all those decisions. I think we'll still call it programming, But I think it's going to transition into being more of like a technical architect, a technical product manager, you know, someone who's really owning these decisions. Agents have to get to the point where they are basically that high touch and that kind of understanding of context that you can give them that level of instruction and they'll just go and do that.

In a world like that, what skills become more valuable?

Yeah.

And what skills become less valuable?

Everything that you do is going to be about essentially this kind of core thing of deciding what is the solution you're going to build? You know, what is the problem that we're facing? What is the solution that we want to build? How exactly do we want to architect that solution? And I think that's going to be the most important skill.

There's so many things I want to ask. I do just want to get back before I forget it. You said about the 1.5x more efficient and more productive.

Yeah.

Do you think we are struggling now in a value chasm gap? And what I mean by that is how much do you charge for Devon today per seat average?

So it's all usage-based. It's essentially by the hour. We try to make things so that they're about 10x cheaper than basically the value of your time.

Because if you think about, say, a software engineer being $300,000 a year, give or take. $150,000 then would be that 0.5x that you are adding.

Yeah.

Are these tools going to be sufficiently paid for the value creation that they are enacting?

Yeah, value creation is a beautiful thing. And to this point, by the way, of where does the value accrue in AI that we were talking about earlier, there are 30 million software engineers in the world, you know, we're going to make them all 10x more efficient over these coming years. We're going to be writing 10x more code. We're going to be doing a lot here. And I think that We could talk about whether it's 5% or 10% or 20% or 30% of the value that actually gets collected by the companies doing this. But honestly, I think the highest order bid is actually less that and more about getting the technology and building the products to a point where everyone is going a lot faster with them.

When you look at something that no one sees that everyone should see, when you think about the future of AI code and the future of software engineering, What does no one talk about that you think more people should be talking about?

Yeah, a focus on deep context. It's already better than us honestly at these sandbox problems. A lot of the tough questions that we answer today are things like, well, there's this project that we're trying to do today and it's very similar to what somebody else actually asked Devin a month ago. How do we use that knowledge and kind of improve on that in order to make Devin smarter? Or there's these little things of like you want to go test the front end for your code base and make sure everything looks as expected. You know, you should be able to understand like what is that supposed to look like? Why is this different from what, you know, if you find a bug, you should be able to understand like how you found the bug. And that's a lot of the little detail of, honestly, I'll call it the difference between code and software engineering, right? Which is basically working in a large complex code base, building some intuition and some representation of all of the different pieces and how they interact with each other, learning how to use all the various tools at your disposal to actually understand what's going on and to debug and diagnose. And I think that is actually the big problem, honestly, in AI coding. And I think that's the big thing that folks will work on next. The thing that's fun, by the way, it's a very practical problem. You can call it a research problem. Obviously, in many ways it is a research problem in terms of how to push for these capabilities and make these things better. But it's not the kind of thing that you can solve in a sandbox. It's the kind of thing that you actually really just need to think about the practicality of software engineering to get into.

You said there about, hey, coding agents was what you saw before other people saw. What no one was discussing. It's a hard one, and I don't know how blunt I can be, but it felt like Devon fell out of the zeitgeist a bit, if we're honest. Elkurser and Windsurf owned consumer attention and owned the brand. Do you think that was a case that Devon just wasn't very good at marketing, or do you think that was like a product misstep?

Yeah, so it's funny because I think there's always an external perception and then there's an internal what's actually going on with the numbers. Without going too much into detail, what I I can say is that in the last 6 months, between January and now, even aside from this, obviously the latest deal this week with Windsor, the usage of Devin has actually grown something like 5 to 10x. And that's been the case in both self-serve and enterprise. I think the pattern, for what it's worth, is it's really like real engineering teams that bring on Devin and they tag Devin all the time in Slack, they tag Devin in Linear and so on, and they kind of use it and grow it and share it that way. And so in many ways it is not the same kind of like single non-engineer just going and, and signing up for an account can immediately just like build something like really cool with it. People do use it that way, but it's not the majority of our usage, right? I think the majority is really just kind of like real teams using it. But, but I think the point that you're saying, I think does get to, I think an important thing, which is that look, IDEs and the IDE experience are obviously, they, they came and, and they really started working, I think about a year before the agent experience really did. You know, I kind of think of agents as having taken off in the last 6 months or so. And I think of IDEs as last year when they really started to become kind of like no-brainer obvious, let's say, in terms of the value that they provided. I think what we're kind of seeing is an artifact of that, which is that, yeah, like folks are more familiar with IDEs because they've been around for a while. You fast forward 6 to 12 months from now, and I think people will be familiar with agents to the same level that they are familiar with IDEs today. And I think it's a—

I'm 10x is mega. It's fantastic. The hard thing is, if you compare it to like a Rappler or a Lovable, the growth is just fucking nuts. Like, I'm in Lovable, Scott, and I just look at the numbers and I'm like, what?

What?

How does that work?

Is that not a fair like-for-like comparison if you're thinking about growth rates?

The reason I would say that it's not— also, actually, I don't even know the number. What are their numbers over the last 6 months, or what are their They've gone from basically zero to $80 million an hour. For what it's worth. Well, yeah, look, I don't want to get into exact numbers, but even with the windsurf deal aside, like not including that at all, we've done roughly that as well in the last 6 to 8 months.

I stand corrected on my consumers like guys comment.

Which is, I think, an important point to consider. By the way, I think these are great companies and, you know, I obviously, I think they're great.

But then I think it's a marketing problem, dude. I'm saying this to you as now, I hope we can be friends. So I'm projecting forward this relationship as a friend. I would say you guys should be better at selling yourself. That's great.

Yeah, well, perhaps we should be. And the great news is we've just now inherited a great marketing team and we get to do that. But the only thing I would call out, I know Omdred from Replit. I've met a lot of these folks before. I think of it as actually quite different products and businesses and so on, which to the point that we were saying earlier is, you know, there are a lot of things, a lot of different product experiences that you can solve for in code, right? There's like bringing a 10x engineer to 100x, there's a 1x engineer to 10x, there's, you know, someone who doesn't know how to code and bringing them to 1x. For better or for worse, Replit and Lovable, to my understanding at least, are much more consumer-y lean, right? And perhaps that's why you hear about them more on Twitter or on YouTube or things like that. Whereas Devin is really, it can be anywhere from startups to some of the biggest companies in the world, but it is really focused on specifically engineers on engineering teams trying to do their work and to go faster that way. And so you can see this in how this is all set up. Most Devin sessions are started through Slack or through Linear, and Devin makes these pull requests in GitHub that you go and review and merge, and Devin works with your whole development system and learns how to get onboarded to your repo and so on. And I think that's kind of one difference there. But with that said, you know, point taken. I take the feedback. Yeah, I think it's a very fair point.

Dude, I'm always here to show you how to do a good tweet. After 10 years, I now know how to do a good tweet. That's all I know. You're the champion.

I mean, you're the master at it. I was going to say, so after this, we'll talk and I want to get your tips.

Dude, honestly, I have so many for you with those numbers. I'm like, wow, dude, you need— Yeah, anyway, can I ask you, my job as an investor also is to think about like market makeup at an end state. Because that's where value kind of also kind of accrues in my mind. When you think about that and you think about like developer, where the market shakes out, is it that Cursor win bottoms up developer minds and Winsurf win and Cognition now win top-down large enterprise, super solid blue chip clients? Is that how you think that developer market shakes out?

I think the honest answer I would say is I think it's far too early to call on any of these. I'll give you a hot take, which is none of us are that close to the future of software engineering. And I think the future of software engineering is, look, this is going to take place over the next couple of years or so, but it really is this version where we call it an IDE or a coding agent or this or that. These are the terms that we use. I think if we're being real about what we're building here and what this is all going towards, I think of this as the next generation of human-computer interface is the problem that's being solved here. Which is basically, as we said, code, software, the whole point of that is just telling your computer what to do. At some point, telling your computer what to do is not gonna take place with code. It is going to take place with you just expressing your intent. And there are a lot of things, don't get me wrong, there are a lot of capabilities problems to solve to get there. There are a lot of, you know, interface problems to get there. But, but I think that is essentially what we get to is, you know, a simple thing I'd say is Tony Stark does not pull up his laptop, you know, Tony Stark goes and talks to Jarvis, right? Like, like I, I, I think there's, there is a point at which you just have a very clean connection, the ability to express your intent and do these things. People talk about generative UI, people talk about single-use software and so on. At the end of the day, what it all really boils down to is you to have a very clean connection and able to just being able to tell your computer what to do and it'll do that for you. And I think if we think about where everyone is today, like I think there are 10 or 20 levels of what the product experience looks like until we get there. And the fun thing is every level is like 2 or 3 months. And so if you kind of just multiply that out, it means we're gonna be there in like a few years.

You know, if this pace keeps going, I believe a common enemy is a very important thing within a company. If I were to push you, Scott, and say, hey, who is the competitor that you most look to think about, respect even, who would that be?

Jensen said this once and it always stuck with me. When you have figured out a way for your company to win, that means that no one else has to lose, then you will know that you have found your path.

Do you think he's done that? I think that's a huge amount.

I certainly think he's done that. Yeah. I mean, you know, Nvidia is an incredible business. Jensen obviously has done quite well.

It's a freaking monopoly.

Like, well, and I think his point, right, is, is that I think at the end of the day, there are so many different verticals to serve. There are so many different niches to own. And there are a lot of businesses which, of course, will work in adjacent spaces and will see things. And, you know, you'll, you'll, you'll run into folks on a deal or whatever it is. But at the end of the day, people have the things that they specialize in, and I firmly believe that. I think the thing that's been fun, I think, throughout the history of Cognition is, for better or for worse, maybe it's because we're insane, I don't know, but we've always had a pretty unique approach and a pretty unique view about, like, here's what we think the future is going to be, here's what we want to build for. And that's different from Cursor, that's different from OpenAI or Anthropic themselves and how they think about these things and so on. Right. The fun thing about that is there is so much to build on code, and I actually truly do think there will be more than one winner.

What have you not built that you would most like to build?

Yeah, the answer I would give today, which is, I mean, the biggest thing that we've been thinking about in the last few days since this deal, is really in figuring out what is that combined experience of IDE and agent, because I think there really is something here. To what we said earlier, I think years from now, all these systems will look so different, and we may not even have the same terms for them. But I think in the immediate future, I think both both IDEs and agents will be an important part of a developer's workflow. And you can imagine all sorts of things of, hey, I want to go and plan out a task in the IDE. I want to be able to use the intelligence and the retrieval and, you know, Dev Search and, you know, read on the wiki to understand exactly what decisions I need to make or what parts of the code this is going to touch and what things I need to plan out. And then I want to be able to hand that off to an agent, have the agent go do the bulk of the work, and then I'm going to go and review the code. And naturally, it'd be great to review the code locally in my IDE if there's any touch-up that I need to do, I can use that, you know, I can use my in-IDE tooling to go and do that. But, but basically, you know, figuring out what is that combined experience where you can go from synchronous to asynchronous to synchronous for, you know, and, and be there for whatever parts need you and, and be able to go and parallelize and do more for, for the parts that don't need you. I think that's gonna be a really fun question to figure out in terms of, you know, what happens with the, the combination of Windsurf and Devin. I think in the immediate short term, obviously like I think I think that there's a lot of Devon to run and a lot of Windsurf to run, and we plan on kind of like, we certainly plan on maintaining the philosophy of both of those products, but I think kind of finding that intersection in between of how do you make it a really smooth experience to go between for the folks who use both is gonna be the fun one.

Dude, I've so enjoyed this. It was, yeah, it's just been a really fun conversation. I do wanna do a quickfire round if that's okay. So I'm gonna say a short stuff. Okay.

I have to say I'm a bit underslept, and so it's, please pardon me if I need like an extra, you know, 1 or 2 seconds for the fire ones.

How have you been sleeping?

Next question. No, look, I mean, it's, we had some crazy, we had some, we've had some crazy days. And I mean, Saturday, Friday, Saturday, Sunday obviously was figuring out, you know, if there was a deal to be done and, and making that happen. And Monday, Tuesday, Wednesday has been figuring out how we bring the teams together and how we build the, the really great thing. And obviously like the, the immediate thing is going out to the customers and letting them know, look, it's like we're, we're here, you know, we've got the the, the engineering firepower and the work to, to be able to support things. We're gonna make sure you guys have a super solid experience. And, and if anything, we're gonna be able to make it better even faster and, and figure out the better together.

You, you even upleveled the engineering team.

Hey. Um, and, uh, and, and like, so, so long story short, not, not a lot of sleep this week. I love it.

Uh, okay, dude, what's one widely held belief about AI that you think is completely wrong.

Yeah, I'll give you a take. Sam Altman had this post 10 years ago called Bubble Theory. He was basically saying— this was in his YC days— and he said, you know, everyone here says that all these companies are overvalued and we're in a bubble, and it's obviously fun to talk about bubbles and all, but I don't believe that, and I'm going to give you a bet. And he said, it looks hilarious, by the way, when you read it in retrospect, which is like, here are the top, you know, YC companies today, and it's literally like Uber and Airbnb and like, you know, whatever. And it's like, today these are worth this much. I predict that 5 years from now they're going to be worth, you know, 3x more, you know. And then it's like, here are these kind of like emerging, you know, mid, mid-level companies. And then that list was like, you know, Stripe and some of these others. And it's like, today this is worth this much, and I predict these will be up at least 3x, you know. And he basically said, everyone says we're in a bubble. If you'd like to go and bet on that and put your own money down, I am happy to put, you know, money on this. I have to say, in AI, I really feel this. I feel this, I think, now more strongly than ever, which is over the last 2 years, People kind of think of it as this GenAI wave. Again, like we were saying, lots of incremental jumps. But the truth is RL is the master, like the big story, I'd say, over the last year or two in capabilities. And I think people have really underappreciated how much is possible with RL.

What do people not see with RL that they should see? Because I kind of think I don't.

Yeah, we've had a few years of, I would call, imitation learning before RL, which is, that's what went into GPT-3 and 3.5 and so on, which is basically you take the entire internet and you read the whole internet and you get a model that sounds like somebody on Reddit. And it was crazy. I mean, it was obviously ChatGPT like shocked the world and it was a huge moment for everybody. But yeah, you could kind of see why, well, you know, talking like the average is not necessarily the thing that you need to do to go build a great, you know, medicine specialist or a great, you know, software engineer or any of these things. And it actually got a lot of the way there, to be fair, more than people might've guessed. But I think that the next big thing with RL is, as we said, you can take any benchmark and solve it. It. It's crazy to say, but I think the next step of that is honestly an open question for everyone in every application, which is, hey, what is the benchmark? If you're an accountant, you know, the benchmark— I mean, in real-world terms, the benchmark is roughly like, all right, you submit returns on behalf of the client, and, you know, if you get audited by the IRS and something comes up, then that was a fail for the RL. And if you don't, then you did a good job. But obviously, you know, you want to be able to constrain it to shorter feedback cycles than that and think about what are the ways that you can determine if an agent's work was a success or a failure. But once you have that, it turns out that you can just train agents to do all sorts of things. And that's what we've seen, by the way. I mean, we've seen folks getting gold medals in these international math competitions and things like that with AI. And the truth is that this is really hard stuff, right? And obviously you have to go and apply it to every vertical, but it's going to happen. So I guess on that point, I guess I was just going to say it's like you think about the companies today, there's the foundation layer companies, right? There's the, the application layer, there's a lot of the tooling. I would happily bet, take all the foundation labs today, right? It's, it's the private ones, right? There's, there's OpenAI, there's Anthropic, there's X, you know, let's put in SSI, we can include Thinking Machines in that list as well. If you add all their valuations today, what is that, like $500B, something like that? I, I think that's gonna go way up in the next 5 years. And then if you take all the application layer companies, you know, The top ones that come to mind are, you know, there's Perplexity, there's Sierra, there's Decagon, there's Harvey, there's us, there's Cursor. There's, you know, and you take that list of companies and today that list of companies is probably worth, I don't know, $50 to $100 billion. And I think that's going to go way up in aggregate. And obviously it doesn't mean every single company is going to go up.

But I think that the value that we're going to produce is so massive here that you can invest in OpenAI at $300 billion or Anthropic at $60 billion. Which one do you choose to invest in?

Honestly, and I mean this sincerely, I think both are great investments.

Amazing. You can put money in one.

Oh man. Anthropic has quadrupled their revenue since they were valued at $60 billion. So it seems like a fair one to pick.

Do you think we will have— you mentioned that like 6 in the foundation model there.

Yeah.

Do you think we will have 6? Do you not think the consolidation already begun?

So I think the consolidation happens and obviously, you know, there's Google and there's Meta I think the consolidation happens and I think in total we probably end up with something like 2 to 6 players. 3 to 5 seems pretty reasonable, but 2, you think just 2?

You got ChatGPT and OpenAI on consumer and then you've got Anthropic on enterprise.

Really? You think it's just those 2?

You might have like really tertiary ones. It's just like search or like minor ones like your DuckDuckGo of the world, but like, the DuckDuckGo of AI.

I'm sure someone's looking forward to winning that elusive title. I think the thing is funny on consumer, I totally agree, which is ChatGPT. I mean, everyone knows just ChatGPT as the term. I think there's an interesting effect, which is obviously, you know, companies want choice. And I think this is gonna happen in a few spaces as well, which is, you know, it doesn't necessarily, it is always a power law distribution where number 1 is like 75% and number 2 is 20% and number 3 is like 4% and everyone else combined is 1%. Or things like that. But I think in terms of folks, in terms of market share, let's say, but in terms of people getting there on capabilities and being able to offer something which is at least competitive enough that 4% or 20% of people would consider it, I think there is reason to believe that there will be at least a few folks that get there. Yeah.

For the acquisition, what was the percentage between stock versus cash?

It was a mixture of both. I probably can't comment on the exact split.

I'm so enjoying this constant game of like push, push.

Well, I'm very happy to— you can ask all your— I'll push back when I need to push back.

Can I be honest? I think you missed a trick with your acquisition video. I'm just gonna be—

I just totally—

I didn't know you from shit, okay? Obviously.

And I watched it and I was like, no offense, it's a bit personalityless.

And meeting you, you're like, I'd love to hang out with you. And you're like amazing personality, charisma. You're awesome. I think that an acquisition video could have been more like personable.

That's fair. I agree. Dude, we were in between, like the lawyers all pulled an all-nighter as well going and getting this because it was like, yeah, I mean, we need to get this ready to go. But you know, and there's just all the various little things of this term and this term. But yeah, it was in between that and in between sorting out all of the questions with the team. And then, you know, we basically had like 20, 30 minutes to go just film something real quick.

Oh yeah, dude, you're on ChatGPT, write a script for an acquisition.

Yeah, I was going to say, I was pretty happy with the output relative to the amount of effort that was put in, but I fully agree that there's more to it.

I'm 100% with you. My favorite bit was, funnily enough, we did have one competitor who we thought really highly of.

Cognition. Okay, hit me with your hard question. Again.

So what did you believe in the world of AI that you have subsequently changed your mind on in the last 12 to 24 months?

I think one thing which has just really started to change in the space— 24 months ago, actually, the story was all about data and just more and more and more data, right? And quantity of data and just figuring out how you get even more. I think that has changed a lot actually towards figuring out a small set of highly curated data for exactly the use case that you care Yeah, and that's been a fun one. You know, I'll just give an example. We, a lot of our research work, we obviously, we obviously can't talk about publicly, but, but we did, we did share one project, which was, um, a model that we released called Kevin. And so it basically, KernelBench is, you know, a benchmark that involves your ability to write CUDA kernels.

Uh, I'm gonna be, I'm gonna be honest, whoever names your, your products really needs to be fired. Devin and Kevin, like, no offense, dude, like, that's not very creative. Damn.

Well, we'll work on that. We'll work on on that. But yeah, so we rolled out Kevin32B, and the whole idea is basically with RL on agent trajectories, on very specifically CUDA kernel agent trajectories, you can make it way better than the SOTA-based models on this stuff. And this is kind of the story that we're seeing in a lot of these, which is rather than just this mass volume of data, if you have a very particular use case and a very particular behavior that you're looking for, honestly, it turns out that a small amount of data in exactly that vertical and you set up the environments correctly and you set up the feedback loop correctly is more what you need. So it's a lot of compute, not a lot of data, I think is what we're seeing. And it's like quality of data over quantity.

Would you advise a new young student to study CS today?

Yes, absolutely. If anything, I would say, I'll give you my hot take, which is the complaint that we've had all along about schools is they they're supposed to teach you CS and software engineering. You go to school and you learn about garbage collection and algorithms and architecture and everything, and then you go on the job and then a lot of what you have to do is debugging JavaScript traces, right? I think the thing that's kind of funny is, as we said, this fundamental skill set of how you solve problems and how you think from first principles and understanding the model of a computer and understanding a lot of these kind of architectures of, yeah, how does a database work or things like that, Like those are actually gonna be the important things I think going forward. And so, you know, maybe one way to say it is at the end of the day, like the degree in CS is more a degree in how to think. And I think that will always be valuable.

Scott, I wanna finish on one, which is I think really important. I've so enjoyed this and I watched this actually TikTok and it said, if you wanna build a relationship with someone, ask them about the trait within themselves that they are most proud of or they like the most. Because in future interactions, you're able to refer to it and it makes them feel good. I'm intrigued. What trait are you most proud of? What's your favorite trait of yourself and why?

Yeah, that's a fun one. It's, uh, I actually am very curious to hear your answer to this question as well, by the way. I think my answer, funnily enough, is I actually would say it's like a certain emotional calmness, I guess. I'm a very competitive person. I get very frustrated and things like that. But, but I think for better or for worse, we've been in a lot of stressful situations. As you can imagine, we've been in a lot of them in the last 6 days in particular. But honestly, that's been the entire story of the company. It's probably a bit more than average, you know, than an average week, but like not by that much. And I think the— yeah, I think, I think for better or for— I'm proud that I'm just able to stay composed. In the situation.

Dude, I love you. You're great. I think you are like the most underdiscussed personality in this business. Like, you need to do more. Like, the Scott Wood brand needs to be a bigger personal brand.

What do you suggest? I'm looking to you as the interim chief marketing officer at Cognition.

The thing I think that's nice with your product is actually that you fall into the Nike bucket. And what do I mean by the Nike bucket? The Nike bucket is their success was they make you feel like a superhero. They tell you that everyone is an athlete, that even if you don't have the skills, you are able to do things you never could before. And I think that you and Devin fall into that similar bucket of human enhancement. And so I think you should tell the stories of that much better. And then I think you should also bluntly, very clearly tell the story of your own growth much more deliberately. People want to be part of a rocket ship, and gossip is very vicious, and it happens when you don't shape the narrative yourself. If you shape the narrative, gossip doesn't really happen, because I can't say you're at $20 million in revenue when you're not, you're much higher. And so don't let other people shape your narrative for you.

Yeah, no, fair enough. I think it's— there's been a real shift, which is, I think a few months ago, it's like we were like, you know, it's, it's almost better if people don't hear that agents are working You know, and we'd rather it that way. I think in the last few months everyone and their mother is now trying to do agents anyway, and so we might as well, we might as well be more public.

But I also think just for talent, I think for funding, I think, yeah, people also just want the brand.

Yeah, yeah, it makes sense.

I mean, what a fantastic guest. He put up with some pretty tough questions there which he was not prepared for. It was an emergency podcast. He was just a fantastic personality. If you want to watch the episode, you can find it on YouTube by searching for 20VC.

That's 20VC. But before we leave you today, Harvard Management Company is constantly seeking out the next generation of great investors and entrepreneurs. HMC has managed Harvard University's endowment for nearly 50 years and was one of the first institutional investors in venture capital. Their experience and long-term investment horizon makes them ideal partners to get world-changing ideas on a path to viability and success. They work as a true partner, providing insightful perspectives to help managers succeed. I personally have had the pleasure of working with the HMC team and can say that they are truly exceptional partners and savvy investors. So whether you're launching your first fund or your fifth, HMC welcomes the opportunity to partner with both developing and established managers. Have an idea you want to share with the team? Just send it over to venture@hmc.harvard.edu. And after securing top-tier partnerships like those with Harvard, the next step is building trust with customers, and Secureframe makes that Simple and seamless. Secureframe empowers businesses to build trust with customers by simplifying information security and compliance through AI and automation. Thousands of fast-growing businesses including NASDAQ, AngelList, Doodle, and Coda trust Secureframe to expedite their compliance journey for global security and privacy standards such as SOC 2 and ISO 27001, CMMC, NIST standards, and more. Backed by top-tier investors in corporations like Google and Kleiner Perkins, the company is among Forbes' list of the top 100 startup employers for 2024, G2's Best Software Awards for highest satisfaction products, and a recipient of the 2024 Cybersecurity Excellence Awards— something I definitely never got in school myself. Learn more today at secureframe.com.

As always, I so appreciate all your support, and stay tuned for an incredible episode coming on Monday.