EPISODE 1761

[INTRODUCTION]

[00:00:01] ANNOUNCER: CRV is a venture capital firm that invests in early-stage startups. The firm has invested in more than 600 startups, including Airtable, DoorDash, and Vercel. James Green is a general partner at CRV where he is known for investing in startups focused on security, infrastructure, and financial services. He joins the show to talk about his path into tech, CRV, life as a VC, and more. 

Gregor Vand is a security-focused technologist and is the founder and CTO of Mailpass. Previously, Gregor was a CTO across cyber security, cyber insurance, and general software engineering companies. He has been based in Asia-Pacific for almost a decade and can be found via his profile at vand.hk.

[INTERVIEW]

[00:00:58] GV: Hi, James. Welcome to Software Engineering Daily. 

[00:01:01] JG: Hi. Good to see you. How you doing? 

[00:01:03] GV: Yeah. I am good. It's great to have you here. You are at the VC outfit CRV, which is pretty exciting for us to have a VC speaking to us today. We don't often get to do that. We just kind of hear about it. 

[00:01:16] JG: No. I feel extremely privileged. Most people are extremely deep in the weeds and here's just me being a fake technical user.

[00:01:23] GV: No. No. I mean, we all know in technology the amount of technology that's only been possible because of VC. I think that's why it's exciting to get to talk to the people that are ultimately backing some of the moonshot projects as well as the ones that were sort of in theory obvious. I think maybe kind of the first question is what was actually your path to becoming a VC? Many people don't maybe think that VC sort of had a life before VC. What's kind of your path to CRV in total? 

[00:01:54] JG: Yeah. The path before VC is both like long-winding and not all that interesting. But, hopefully, you and your listeners will find interesting. But I'm originally from the UK. I grew up near Birmingham. Those folks who aren't from there, it's a place in the middle of the UK. Moved here to the States in 2012. Went to Harvard undergrad. 

Initially, I was a neuroscience, neurobiology major with the rest of my time spent doing computer science on the side, which really has nothing to do with the stuff I invest in now other than I invest in very technical projects and people who are doing stuff that is much more complicated than brain science as I would have called it back then. 

I went and spent a ton of time in labs. And I was modeling brain behavior and doing a number of stuff there. And this would have been in 2013. I think I quickly realized the US pharma system is deeply broken. And so, switched from wanting to create pharma drugs, which actually now given LLMs might have been a really interesting career. But switched from that into more kind of pure software startups tech. 

And I was like, "Okay, I have no skills other than basic computer science skills and engineering skills. And that I don't need to sleep very much." I was like, "Okay, how do I get into tech?" There was a website in the UK at the time called workinstartups.co.uk. Very, very sophisticated. And I scraped all the emails from it and sent cold emails to hundreds of startups. Hundreds of startups of which many, I'm assume, went to spam. Because this was early days of matching identifications and IDs to names. But ended up working at a company called Carwow, which was a consumer internet. You go online, buy a new car, which at the time seemed very innovative but is less innovative now.

[00:03:58] GV: And they're still around. Carwell. Right? 

[00:04:01] JG: They are still around. It's actually now. I was there when it was nine people. And then they subsequently raised hundreds of millions of dollars from well-known - like Excel, and Vitruvian, and Balderton, and big global VC firms. But I started really got into tech was - that were those days. And spending time with engineers, with salespeople, with marketing, all around it. 

And what I was really interested in from those folks, to the similar conversation we're having now, is like, "Okay, why did," in that case, Excel, "give nine people in a room who were selling 200 Corsas, or Vauxhalls, or not very nice cars? Why did they just give us millions of dollars?" Because it made no sense other than maybe they thought we were on a dream or on a path. 

And so, I did something very similar to my scraping and emailing with VC firms. And I interned shortly at a firm called Insight Partners just for a little bit to learn a little about what they do. And I was going - I got a return offer and I was going back and forth on should I go back. I kind of had like ultimate downside protection, to be honest. 

And so, I went and started my own company with a guy who I met in the Boston ecosystem, which was a manufacturing IoT business, which for various reasons was not successful. But did, one, gave me a lot of empathy for startup founders. But, two, just gave a lot of interesting stories of getting things off the ground. Doing all sorts, as we can go into, on roles of co-founders and early employees. Doing lots of different things and roles. 

And then I decided to go back to Insight in 2016, I went in my first foray into investing. And so, through an up and down roundabout journey, I went from startups to do my own thing and all in between. Ended up investing. It was 8 years ago now, I guess. Yeah. That was the journey.

[00:06:09] GV: Yeah. Wow. I like that you actually did start in startups. I mean, I also - although it kind of doesn't work today, I don't think, but the cold email and the scraping of the cold email. For those that - 

[00:06:20] JG: I think it totally works. I would argue being from the UK and not knowing anyone, moving to the states, I would argue my entire career has been made on the power of a cold email. 

[00:06:28] GV: Awesome. Okay. Well, I think likewise, ultimately, being from the UK as well and the US for me was always kind of realizing this opportunity out there. I very much agree with that sort of sentiment. And maybe I should do more of it for my startup. Anyway, I think what I'd also love to hear is I've read the book The Power Law, which for anyone that's listening and doesn't know that book, it's basically just a kind of history of VC. And I think that's been quite helpful for me to understand kind of the backgrounds of many VC outfits. But also realizing there's no kind of one type. And there's many myths even around kind of the VC world. I guess for our listeners, what's maybe a myth about VC that you'd like to kind of debunk from the outset? We've got a lot of technical listeners and what's probably something that they might be thinking. But, actually, you can say that's just not kind of how it works in VC. 

[00:07:24] JG: I think one of the really important things I've learned in this job is that there are really only a couple of very, very high leverage points in companies' lifetimes. I think it's the role of the VC to have seen high leverage points and have experienced high leverage points and help founders, entrepreneurs, engineers in those moments. 

[00:07:49] GV: Can you define what is a high leverage point in - 

[00:07:52] JG: I would say the classic one people talk in Silicon Valley is Mark Zuckerberg turning down the billion-acquisition offer, right? Or it could be as simple as, "Hey, this first employee you've out-scaled and you need to fire them." Or it could be identifying a new technical wave and making sure you're on the right side of history. 

I think to the current movement of AI and the historical ML-based architectures - and people have built hundred-million ARR businesses on historical technology. And if they don't move with the new architecture and the new models, they're going to become irrelevant real fast. And this is not dissimilar to the movement to the cloud in 2012, 2013, 2014 where companies historically were built with on-premise license maintenance business models. And the ones that evolved are still relevant today. 

You think that people no longer talk about PeopleSoft but instead talk about Workday. But people still talk about Adobe, right? Because Adobe moved from what was shipping boxes and shipping CDs to selling cloud software. And I continue to believe actually Adobe is a fantastic example of like large at-scale entrepreneurship in action. They're somewhat of a poster child for an adoption of generative AI in their products. And they have some of the advantages that it's Creative Suite than a GitHub-like product or something like that. But it's still an amazing example of that. Actually riding a wave and finding yourself inside of a high leverage point where, if you don't make the right decision, it could be fatal for the business. 

[00:09:44] GV: Yeah. I think Adobe is a really good example. I feel it's a company that's never - they just kind of - they're not cool, unfortunately. People talk - 

[00:09:52] JG: They are really not cool. But they are also really big and have a - 

[00:09:56] GV: Yeah. They're really big. They're really successful financially, as you've just called out. They have written these what I now understand to be high-leverage points pretty well. Yeah. They can sort of catch a break when it comes to, I would say, their image, unfortunately. But, yeah. Great company. 

I think we're going to talk a bit about sort of things that will kind of help our listeners a little bit. I think a lot of our listeners are technical and are often probably right now pitching to VCs or thinking about it. Just before we kind of dive into that, maybe just what is like a little bit of background on CRV itself. Whereas it's a pretty sort of very established. I believe it's sort of 1970s Palo Alto outfit, which is pretty cool. Yeah, what was kind of just the brief history I guess of CRV? And sort of what's the focus areas now for the firm? 

[00:10:45] JG: Yeah. The firm now and acronym of CRV historically started as Charles River Ventures. Actually, started as a $5-ish million fund in Boston in 1970 with basically the pitch to invest in MIT-based and MIT spin-off founders. We've since very much evolved from that missive. And we are now, to your point, Palo Alto and San Francisco-based rather than Boston-based. 

Similarly, in terms of the high leverage or crucible moments, I think it became pretty obvious to the partnership that the epicenter of technology was moving to the Bay Area and that we needed to remain relevant, we needed to move here and have a big presence here. We were bicoastal for a very long time but actually are now only on the West coast. 

It's been 54 years. We're in our 19th fund. We're one of the oldest venture capital films in the world. It's been just over 600 startups that we've invested in out of the 19 funds. You will have seen or used many of them. I think Twitter, DoorDash, Ring on the consumer side and a number of large enterprise companies. I think Airtable, and Vercel, and Crible in more recent days. 

The firm is pretty much all seed and A. We'll do some Bs as well as that core business. But very early stage. We'll focus on first, second-time, third-time founders with unique point of views in mostly areas that we understand. Think enterprise software, marketplaces, AI, cyber security, developer tools. Those are kind of the areas in which we're spending most of our time. Yeah, that's the firm. We're really excited to partner with the next generation of founders. 

[00:12:30] GV: Yeah. I mean, there's obviously some good names there I'm sure that will resonate to listeners. Vercel, etc. The location thing is interesting. I'm based in Singapore. At least I understand sort of the benefit of being in a place where there are actually VCs situated and kind of what that brings. But maybe, again, people live all over the place these days. What is it about actually being based in Palo Alto that brings benefit to CRV today? I can totally - might not been such a question, I don't know, 10 years ago or more. But, today, there clearly must still be benefits. But how do you articulate that? 

[00:13:10] JG: Yeah. I actually think we're very fortunate that we have a Palo Alto and San Francisco office. Palo Alto continues to have Stanford as a large driver of innovation and young engineers. As well as the behemoth of Facebook, and Google, and Nvidia now and all the other things in the Palo Alto South Bay Area. 

I'd also say I spent a lot of time with cyber security. There are generations of cyber security founders who have come out of South Bay. Think Palo Alto Networks, Zscaler, Netskope, and all those. And then with San Francisco and the rise of all these AI businesses. OpenAI being founded here. As well as Anthropic and many, many others. The network effects are actually amazing. 

And this is a very random story. But I was walking down the street probably half an hour ago before this and ran into the head of security from Harvey. And I know him. And we caught up. And it's like the serendipity of meeting people just on the street is very hard to recreate. You can create it on university campuses, obviously within PHD programs and within large companies. There are network effects within large businesses in different regions. 

I think, classically, Duo in Michigan of all places. Or Datadog in New York or things like those. But the San Francisco and Palo Alto ecosystem continues to be amazing. I moved here from New York in March of '21. And that was like the depths of Covid, if you will. And I was pretty skeptical, to be honest. I turned up. No one was going into offices. No one was here. I was having lots of Zoom meetings. I thought, "Well, other than the fact the weather's lovely, why am I here?" And the last two years have just been very obvious that network effects exist. And that being surrounded by like-minded people who are ambitious, extremely talented where serendipity can play a part in your life I think is really, really important to building successful businesses. Not necessary. Clearly, you can build successful businesses not in San Francisco and Palo Alto. But it's advantageous.

[00:15:23] GV: Yeah. I think that's a really, really good point. Yeah. I mean, I see even here in Singapore where I sit every day with our company. We've got a whole bunch of other startups around us. And just those interactions and who's then coming in and out of the office each day from other venture outfits or from wherever. Just having those moments where you actually get to meet someone that you never have met before and have a conversation that can actually change things is super true. 

Yeah. I mean, actually to Palo Alto specifically. Just coming out of Covid, I was working at a cyber security firm which was based basically between Singapore and Hong Kong. But I got sent out to Palo Alto, too, and I thought I may had the same skepticism slightly. I was like, "Are people really still out here? Why am I going out here to meet supposedly some people?" And I came back from that trip just going, "Oh, wow. I need to spend more time here." 

I went to Palo Alto. And I went to Boston as well in the same trip and realized big time this is why people are here. Because you can meet so many really smart, really interesting people, and the really great companies have a presence here. I mean, especially in cyber security. But in other industries of course as well. 

[00:16:27] JG: I would argue that some industries are more relevant than others for the network effects San Francisco and Palo Alto give you if you're building application-level AI software. [Habio inaudible 0:16:42] is an amazing business based in New York. Build a great business in-person in New York. Clearly, they don't need to be San Francisco. If you're building foundational models, next-generation cyber security, developer tools, you need to be where those people are. 

There is a limited number - and many, I assume, all your listeners. There are a limited number of people who actually built these. And so, you should surround yourself with people who are actually building these things rather than listening to them online or reading about them on a Reddit thread or pick a thread. But you know what I mean. And so, the sector-specific serendipity has become super obvious to me over the last couple of years. 

[00:17:26] GV: Yeah. And I think I totally appreciate there are listeners probably listening right now who maybe kind of in the starts of their sort of careers or startups. And these places are incredibly expensive. But at the same time, it's almost like an investment in your future. That sounds really corny. But it is kind of the way people I feel have to look at it, which is something that I did back in the day was effectively invest in living in New York. And then that turned into a few things. And has maybe been lost I think today slightly. It sounds like you kind of did almost the same thing [inaudible 0:17:59].

[00:18:00] JG: I am very cautious over calling San Francisco expensive versus overpriced. 

[00:18:07] GV: Yes. 

[00:18:08] JG: San Francisco is clearly expensive. Rent is expensive. Food is expensive. Everything is expensive. It feels underpriced relative to the impact it could have. Most of your listeners have probably had the experience where they've gone to a restaurant and they feel like they've got ripped off because it was overpriced. They've also probably gone to a restaurant where like it was objectively expensive but an amazing experience. That's kind of how I describe San Francisco to people, which is like pretty expensive but a really amazing experience. 

[00:18:35] GV: Yeah. That's a great way to I think explain it. New York may be still overpriced. I don't know. Nowadays. 

[00:18:42] JG: Depends on what you're anchoring to. And I don't want to get many angry emails from New Yorkers. 

[00:18:48] GV: Yeah. This may be a good place to kind of move along. As you know, we got a lot of technical listeners. They could be sort of in the thick of or kind of thinking about starting their own thing. You've just called out CRV does invest early. I think some of these questions hopefully are fairly - they're coming from someone who does assess early companies. What are some of the most common mistakes or pitfalls that you do see early-stage software? Let's mainly go with software because we are Software Engineering Daily. Pitfalls that you see early stage software startups make. 

[00:19:19] JG: I mean, there are so many. But one of the things that I often see is that someone works - call it a big tech company. Let's pick on Google because they're a giant. And they come up with a problem, an engineering problem which maybe they solve by building a tool that's proliferates within Google. Airbnb has done this. Netflix has done this. And the person who builds it is like, "This was amazing product. Everyone must need this." And they say, "Okay, I'm going to leave my job at Google. I'm going to go off with one, two, three of my friends and I'm going to go build this." 

And it turns out no one else has that problem. Because Google is a unique behemoth that does not reflect the real world. And there were and continue to be so many examples of this where people build something internally and they're like, "Let's go commercialize it," and no one ever wants to buy it. I think that's an extremely common example. 

I would say something I cue in on a lot is if there are named projects doing the same thing instead of tech organizations I respect. If you think early days of container orchestration before Kubernetes became a thing, that was a very known problem inside of every tech company respected in the world. And it's like, "Okay. Clearly, this is a problem not just Google has." They're going to throw hundreds of millions of dollars at Kubernetes. But everyone has this. And so, that was clearly a problem that resonates. But I'm not going to pick on people's companies. But there are lots of examples where that does not resonate with the rest of the world. I think that's a really big problem. 

The other thing that continues to amaze me is that people build companies where they do not have a unique insight. Where they are unique people. They've probably done amazing things. They have PhDs. They have lots of things. But it drives me nuts when they're like, "I am a world expert in protecting applications at the edge. And I am building a marketplace for pillows." And you're like, "What is going on here?" 

You're clearly amazing. And lots of people, they've become jaded. Right? If you spend 10 years doing one thing and become a world expert at it, you may just be tired of it, which is totally reasonable. But then don't go build something that you do not have a unique insight about. 

One of my favorite, favorite questions, which I definitely stole from somebody over the years who definitely stole from someone else, is what do you know that other people don't? Really, what do you know? I'll take Mercury, the startup bank, for example. What they know more than other people is that, actually, startups are not all as valuable as their revenue. But if they've raised money, they are super valuable customers. Because you can sell them lots of other things and make money on deposits. And you can sell them other products. And then some of them scale and you can make money with going there. And so, it actually is a better business than selling these people software, or cards, or anything. It's like a full stack. Really, that question, "What do you know that other people don't?" is really the cusp of really big businesses going back a really long time. And so, that unique insight is so important. 

[00:22:52] GV: Yeah. You can still be - I mean, I really like that question because you must see products that do look similar. Someone comes along and it does look maybe something you have seen already but because - right. However, if whoever you're talking to can answer that question in such a way that you suddenly realize, "Well, sure. We've got two on the face of it kind of similar products." However, the way this one's being approached, if that insight is what's going into that approach, that makes it more interesting is what I assume. Yeah.

[00:23:25] JG: Yeah. We just have so many examples of this. We see lots of people, especially at similar times, building something. And it's my business to try and meet as many of them as possible. And, occasionally, you'll meet one who's like we're building it. But the reason we're building it this way is because we know this. And this thing and how we're building an architect in the product, or how we're doing the go-to-market, or how we thinking about the fundraising, or how we think about the hiring and capital structure of the business. If you know that unique thing gives you a differentiated advantage over the competitors. 

The reality is most categories that are worth going after, you'll probably have lots of competitors. There are some examples. When Anduril got started, people weren't building Anduril competitors. And, clearly, an amazing business, which what they knew, which probably other people didn't know, which is stupid, because of course we knew it, is that the US is bipartisan and they're going to spend a lot of money on weapons. It turns out that's pretty obvious. 

But I think if you really know something unique, you should push on it and use that as a structural advantage for your business while you continue to know it and other people don't. As soon as they know it, it's commodity market and it's just execution. But I think that's pretty important. 

[00:24:34] GV: Yeah. That makes a lot of sense. I think, again, one thing that's maybe misunderstood or just not talked about so much in terms of actually the tech due diligence that a firm like CRV will do on a company. Walk us through that. Especially, is there differences between seed and series A in terms of how deep that goes and who does it? Or, actually, you guys don't particularly care that much about the stack and what's going on in the code? Yeah. Talk to us about that.

[00:25:04] JG: Yeah. You hit it right there, which it does depend on the stage. And, also, depends on the type of business. The application layer software businesses, sales and marketing, software stack, that tech stack matters less than if it's a cyber security business or foundational LLM. 

Here at CVR, we are super early. We can invest pre-idea. The tech stack is not determined at that point. And so, at that point, we're investing on the people and what they know. And in that case, we often put them in front of our network where they spend time with the other engineering leaders, or CEOs, or even CROs in our network. We do that. 

We also put people in front of buyers. If you're a cyber security business, we put them in front of our network of CISOs where they say, "Okay, here's what we're building. Here's how we view the pain point. Does that resonate." That's when it's super early. I would describe like pre-idea, post-idea, pre-product, post-product, pre-revenue, post-revenue, product-market fit. Kind of like many stages of what we do. 

In terms of the tech stack when we're investing post-product, I tend to determine is it in production or not in production as a very clear delineation on the product. Because, I'm going to generalize all of you all listeners, we've probably all built applications or played with applications that are really not production ready. You would never put it into any sort of real production environment. And it would break everything probably. 

And so, that is a big part of when we are evaluating post-product businesses. Okay, is it in production somewhere? Or is it production ready? And so, we do it that way. And, today, the architecture or model of choice is so interchangeable. If the rate of improvement from the models as well as the tooling around it is almost faster than anything I've ever seen. I don't really care like what cloud provider you're using. Or maybe I care roughly what database you're using. But not really. 

I care that is it extensible. Can you move with the times? Have you built it in such a way that allows you to, in the case of LLMs, swap in LLMs in and out? And so, you need to. Or you build things around it. We do really care. But we really care about the ethos and how you're building it. Less so what you build. Because it's probably going to change just given a lot of the current movements.

[00:27:43] GV: Yeah. The reasoning behind the stack as opposed to the stack I think is kind of a good way to summarize. Yeah. 

[00:27:50] JG: Yeah, for sure. I think unless you are Sam Altman, we have no idea how these models and LLMs are going to progress. And so, he knows something where the rest of us don't. Maybe he'd be a good person to talk about it. But, otherwise, we have to build it in an extensible manner where we assume they're going to get better and we assume that they're going to get cheaper. And the rate of which that change happens is what we're all betting on. Not is it going to happen. 

[00:28:13] GV: Yeah. And just sticking on kind of - I'll go to a kind of post-investment question in a second. But sticking on sort of before you have actually, let's say, issued a term sheet, any metrics, milestones? Again, you just sort of delineated all those steps, all those stages a minute ago. Pre-product, post-product, et cetera. But are there any kind of - I'm speaking a little from experience here where I'm in this part of the world, I would say, that most VCs are coming up with kind of exactly the same metric for, I would say, pre-seed. It's different. I mean, very different in the US, I believe.

[00:28:50] JG: I'd be very interested to hear what that is. Because the metrics for pre-seed don't really exist here.

[00:28:55] GV: Well, that's exactly, I would say, the problem. Again, even I read a great book Venture Deals. I'm sure a lot of people have probably read it as well. And it's written very much from a US perspective of we're going to talk about how you should actually talk to VCs, as well as lawyers, et cetera. And kind of one of the points that was made was if someone is throwing metrics at you pre-seed, they probably aren't actually a pr-seed investor was basically the - 

[00:29:22] JG: Yeah. I would resonate with that. I still have it where there are not tech angels. People who made their money in other categories who are starting to try and invest in pre-seed companies. And a lot of them just need educating that, at least in the US pre-seed market, you're investing in teams, ideas, visions, the ability to execute. Sometimes there are. But it's very rarely metrics. And if there are, it's proof of concept users. And how often they're using it? Or if it's a consumer, it's number of consumers who are using this basic application. 

I can't remember the last - I'm going to use air quotes for the people who can't see me, "hot precede investment" in a very technical category. I do cyber security. In cyber security that had tons of metrics. Just in the many years I've been doing this, I haven't seen that.

[00:30:19] GV: I think that's good to kind of clarify, I say. We've got listeners all over the world. And at least I can speak from a region that is very different I think to the US in that respect. And I would say it's just sort of misunderstood actually amongst a lot of people what pre-seed and even angel - quite frankly, angel investments sort of should be. You're now invested in a company. 

[00:30:41] JG: Before we go to that, I want to have a quick comment on the revenue metrics question. 

[00:30:46] GV: Sure. Yeah. Yeah. 

[00:30:47] JG: We internally talk a lot about this. Because we do pre-seed, seed, A, B. Mostly, the series A and B's have revenue. We've started asking this question, which I really like, which is if it didn't have revenue or didn't have metrics, would you still invest? We don't care if the business has a million of ARR or revenue. We'll try to invest in businesses that can get to a billion in revenue. They could get it there. Obviously, not all of them will. But could get there. 

Who really cares about the first million? It matters in that it's proof that you can sell and proof that the product works and all those things. But isn't overly indicative of do you believe it could get to a billion in revenue. And so, we've started asking this question around the metrics. It's like, "Okay, if it had no metrics, would you still be interested in investing?" And I think that that keeps you very honest of not leaning on five customer calls or five customer points where I say, "Okay, five people are paying them a 100 grand," which is objectively a lot of money for a seed company. But it's like not that much money if you're trying to build a billion-dollar revenue business. Sorry to cut you short. But I think that's a really important question that we think about a lot. 


[00:32:05] GV: No, absolutely. On the podcast, I spoke to one other VC, and something that came up there was the point of if an early-stage company has achieved revenue very early, does that mean it's too easy for someone else? I mean, why have they been able to achieve revenue that early when - does that mean other companies could just come in very fast as opposed to a company that focuses more on product and aiming for that billion-dollar revenue, which might not come as quickly? But I I'm curious if you've got the same or different view on that. 

[00:32:36] JG: I try not to generalize in that I've invested in companies that had revenue day one and were very successful. I've had invested in companies that didn't have revenue for multiple years and then end up being very successful. People often point to OpenAI and Figma as companies that took a long time to get to revenue and then were very successful. Airtable and Apple follow as a great example of that. 

There are also examples of ServiceNow, or Viva, or companies which are now very successful, that revenue pretty good pretty early. I try not to generalize. I agree that, especially in today's GPT Wrappers where people get revenue day one, it's easy to say, "Okay, this is really easy to go build." But the reason the revenue probably isn't sustainable is because it's GPT Wrapper, not because it had revenue day one. 

[00:33:28] GV: Yes. We'll get on to that area in a little bit. I think that's a good one to call out, so yes. Post-investment, again just through reading Power Law, I guess. That was maybe the most insight I got from that book which was just that post-investment, the approach of a VC firm can differ very wildly. It's not one's better than the other. Just I think this is where both the investee and the investor want to be, hopefully, on the same page about what their approach is. 

How would you describe CRV from a post-investment perspective, and how much involvement do you like to have? Do you bring in technical and product guidance? Or just how do you interface with your portfolio?

[00:34:16] JG: Yes. I think there's three questions there. How do we work as a firm? Do we get in the weeds of technical and product guidance? Then how do we interact with founders? But I would say how do we work as a firm is pretty aligned from partner to partner in that we really work on a poll model. My analogy is I'm on a bus, and the founder is driving the bus. I could definitely shout at the driver, but the driver is really not going to listen to me in that he may listen to me. But if he's driving in a certain route and once he's driving a certain route, it's what he's doing. Apologize to any of the founders I have back listen to this because they've probably heard that analogy 10 times. But if you need something, then we will do everything in our power to go achieve that. 

This morning, one of my portfolio companies is hiring a VP of customer success for your listeners. I'm very in the weeds; helping interview, identify score, negotiate, all of that. But that's because I've done it a bunch of times, and so I'm very, very involved. I am not involved at all in them hiring engineering leaders and not because I haven't done it before because I have, but just because they feel very well-equipped to go do that. They've done it multiple times from previous experiences. They don't need my help. Great, happy to get out of the way. 

I think that leads into the second point which is if you are opinionated, which you probably are, and if you know things that other people don't, which is one of the reasons we invested, we probably are going to be that involved in your product and engineering road map. We may share examples from the other portfolio companies and leaders of what's going on there, and probably introduce you to people who are at the forefront of innovation in different categories, simply as a way to provide you different perspectives. You can listen or not listen. It's totally up to you. 

We really, really try not to be heavy-handed at all, which I think resonates with the people I work with. I can't speak for every founder and CFO I've ever worked with, but I think that resonates. It's super, super important that we are there when people need us. It's kind of like a time to response or whatever that metric is. I'd actually be very interested to look at my time to response for the founders I work with, but I'd hope it's under an hour of my SLA. If you need something, we should be there. Barring me being asleep or, yes, getting married, which I did a couple years ago, I'm going to respond. My wife doesn't love the fact that I pretty much always respond to, yes, insert name found on my phone, but that's where I'm at. 

I'd say to summarize, we work on a poll model. We try to get you answers quickly from us or from our network. You're in charge of product and engineering, and that's one of the reasons we want to work with you. Not because we think we're the best product engineering people in the world but probably because either you know them, we know them. Or between us, we can get to them. 

[00:37:30] GV: Yes. It almost sounded like you were saying that you have four-nines SLA on response time to founders, which is nice. I like that. I'm sure you've heard this question many times. But at the same time, I do feel it's misunderstood or people, yes, have the wrong impression one way or the other, which is co-founders versus solo founders. I think YC gets looked at as almost the most publicly available information of how venture potentially works. How do you look at this? Then why would you back a solo founder over a team of two or three? How does that play into things?

[00:38:11] JG: Startups is a business of outliers, so there are going to be exceptions to every rule. I also use the saying or the phrase that you don't have to play your life on hard mode. I do believe that building a company on your own is playing life on hard mode, which people could be very successful playing life on hard mode. It's very true, and I do not have the exact stats. But if you pulled every tech IPO in the last - since 2010, after the Great Recession. How many of them were solo founders, and how many of them were multi founders, so two or more? 

My guess would be the vast majority have two or more founders. I think that's for a couple of reasons. One is because not only is being a founder hard. It's also lonely. It's a shared experience with somebody, and it's nonlinear. I believe pretty strongly in co-founders. For the reason that it's hard, it's nice to have people to do it with, and you're very, very unlikely to have all the skills to be super successful. 

If I think about - let's take Sierra, Bret Taylor's new company. Bret Taylor was the CTO of Facebook and the Co-CEO of Salesforce. He decided to have a co-founder. If he decided that he should have a co-founder, why do you not have a co-founder? Who knows that Sierra is going to be a public company? But just in terms of looking at your peer set of who is a very good founder, he's probably a pretty good phenotype. 

I would say if you're not going to have a co-founder, you should have a really good understanding of why and not for VCS but for yourself. It shouldn't be I haven't found someone good enough to partner with because there are a lot of really good people in the world. That's not a good answer. 

[00:40:11] GV: Yes. I think that's a great call out that if you haven't found that person, that's not the answer. It doesn't mean you shouldn't start the journey, but be very willing to take in that person or people as co-founders when you do find them because it does completely change the journey. I like that you said it's nonlinear. I think that piece is also misunderstood when you're trudging into the office every day, and you don't know what's going to happen. You don't know is this a progress day, is this a two steps forward, one step back day. 

But it helps big time when you've got someone that you're doing it with that you come out the office at the end of the day and go, "Well, at least we were doing it together, and we can figure out. There's this problem that we got to figure out, but it's not just me sitting in an office on my own trying to figure this one out." 

[00:40:56] JG: There is a reason that if your company is successful, you make a bunch of money. Companies are hard, and that's why they're valuable because there's not that many of them that are really successful. It's really hard. You shouldn't do it alone. It doesn't make a ton of sense that people put a co-found relationship akin to a romantic relationship. Similar way, life is nonlinear. Companies are nonlinear. It's just a little easier if you have someone to share it with. I think it's a really important fact. Again, you can do life on your own. You can do companies on your own. People have done it, but it's probably harder. 

[00:41:37] GV: Yes, for sure. Just coming on to really some companies maybe in the portfolio, as well as just looking at the general environment at the moment, what would you say are some of the most - maybe let's just take the last two years and what to you have been some of the most impressive or transformative, again, software today products that you've now invested in. 

[00:42:04] JG: Yes. I invested in a business called Astrix, which is really focusing on the non-human identities in the world. If you think about a public company, it probably has 5,000-plus humans, probably has 50,000-plus non-human identities. They keep getting breached, and thus they keep having access to your organization. I think about the phrase that identity is the new perimeter, and I think that's right. But I actually don't think it's humans. I think we're pretty well-protected as humans, unless you're getting fished, which happens. 

What I don't think is pretty well-connected is the non-humans. As well as if we fast-forward three years, I think there's a real chance that the vast majority of Internet traffic is agent-based that there are so many non-human identities on the Internet and interacting with your applications that there's just no way you can manage that security breach. It's been really amazing to watch this business execute in that it turns out people keep getting hacked with their non-human identities. 

If you're a CISO or a security leader, you're trying not to end up on the front of the Wall Street Journal, and so you don't want to be that person who ends up on the front as a function of you getting popped from one of your unsecured tokens that got breached from GitHub or Dropbox. Or pick one of the hacks, and next week there'll be another. That has been deeply amazing to watch happen. 

[00:43:42] GV: I'm curious with Astrix. Obviously, I'm asking for slight specifics here. I mean, there are a lot of people tackling this space. What would you say Astrix is doing to that question? What do they know that no one else does? Or how are they approaching it in such a way that was why you were wanting to invest in them specifically?

[00:44:00] JG: Yes. It's busy or crowded or category. I invested pre it being called non-human identity. I invested when it was - it wasn't two guys. There was probably 15 of them. I like the series A. But I invested when what they were saying to me was there are so many third-party connections and that the non-humans will be the new perimeter. That you don't know this, James, but agents are the future. 

We probably made the investment about a year ago, and I remember being like, "Okay, I'm sure agents at some point will be the future." But you don't really know that, and they were right. If you think fast-forward 12 months, the rate of which agents have improved is way faster than I thought it was going to be. I think it's going to continue to do the same trajectory. They've architected the product in a way that is extensible across infrastructures and agent-based tooling that is very hard to retroactively do. They have very extensible coverage, and they have lots of tools that allow you to do remediation post-breach or post-third-party breach. 

The fact that they've built this for the future in mind, knowing what they know, and with a product that can be reactive, not just protective, I think is very unique. There are lots of competitors. I mean, to our previous conversation, in categories that matter, there often are, in security especially. But when I spoke to the customers - when I made the investment, I spoke to customers actually more recently. These people have leaned on them many, many, many times. If you can't imagine a world without it, it's probably a pretty good sign that it's doing what they think it's going to do. 

[00:46:10] GV: Yes, for sure. I've always loved the Superhuman, the email platform. Their PMF playbook which was talk to the customers and then find the ones that say, "If I took this product away from you tomorrow, how would you feel," and then focus on the ones that say, "Very disappointed." I think, yes, it's a great way to look at it. 

We're coming to the end here, but I'd just love to get your take as I'm sure many people would like to understand LLMs currently, ChatGPT for X. As you called out earlier, GPT Wrappers, et cetera. How are you evaluating? There are so many across different sectors, of course. Maybe, I don't know, let's take security and maybe one other. But how are you evaluating these? To me, it's just there's a tidal wave of things. I'm really curious how you're able to cut through noise and evaluate these. 

[00:47:00] JG: Well, let me tell you. We're spending every day thinking about it. What is today's nugget of wisdom may be different next week's. But it's hard to ignore how valuable many of these platforms are. I take what the CEO of Klarna says with a pinch of salt. But he came out today saying they were going to replace Salesforce and Workday with homegrown AI-built applications, which I assume they pay $10 million-plus four, given how big Klarna is. 

There is a lot of worry in that GPT Wrappers or AI cloud applications or anything is not overly sticky. I think a couple of things that I'm queuing in on or queuing in on to give me excitement is one of the things we always do and did in the cloud area and the on-premise area, in the box area which is just the people. We're spending so much time with the people. That is one of the number one drivers of why we make an investment. 

As we think about the business, do they have unique distribution? Do they have unique data set? Do they have unique fine-tuning models that have come from building themselves? Have they built workflow that people are building careers around? I'll go back to Harvey. I was in the Harvey office today, and that is a "GPT Wrapper," which is totally unfair because they've built amazing product that people live in and that people deeply love. The same is true of many of our investments that are AI-powered applications. 

I think it's unfair to totally call things GPT Wrappers. There are clearly some, but we didn't call SAS S3 bucket wrappers and the Postgres database wrappers. It's not a little bit unfair. We have to assume that there is value in the data, in the workflow, in the distribution, or in the talent. If there is defensibility and/or a moat in those things, then that gets us really excited. If you are a world expert in sales tech and then, again, you build a pillar marketplace, we're probably going to get less excited. 

[00:49:19] GV: Yes. I think someone did ask me the other day just from where I sit. What were my thoughts on - call it AI, but I have to always delineate and say, "Well, LLMs is what we're talking about really here." But AI - it's actually just the quality of product. I don't think we just haven't had enough time to see the potential of quality of product is so huge. I think it's just that GPT, and OpenAI, APIs, and Anthropics APIs, et cetera. 

They are relatively new. I mean, people still forget this. We're talking one and a half years. For products to have actually fully realized the possibilities, it's so early. People are almost passing over it. Now, it's like, "Oh, well. It only does this little thing." Sure. But we've only had 1.5 years to even scratch the surface on what you can build with these things, so yes. 

[00:50:06] JG: It's been really interesting. These AI products, they demo really well in that when you see them on demos, it's - I was at a demo there a couple of days ago. I was just amazed at the demo. Then it doesn't always work in production. Go back to what we were saying earlier. It's like it's early. This stuff is still somewhat brittle at some areas, but it is going real fast. I continue to be absolutely amazed at the capabilities of some of this stuff. We're in for a really interesting ride over the next two to five years that many people our age or still "early in their careers," this is a big wave. We should make sure we're paying attention to it and writing as much as we can.

[00:51:00] GV: Yes, for sure. James, it's been such a pleasure to have you today. Really appreciate the time you could give, and I think you've given a ton of insights. I think our listeners will be able to take away today. Just getting to hear really from someone deep in VC and how you look at things I think is just super valuable. I just want to say thanks so much and hope we can chat again in the future. 

[00:51:25] JG: Thanks for having me. 

[END]