Investing Billions - E418: AI, Venture Capital & the Future of Investing

Episode Date: August 19, 2026

Most venture firms think AI is another productivity tool. David sits down with John Melas-Kyriazi co-founder and CEO of Standard Metrics—the AI-native portfolio management platform used by leading venture capital and private equity firms—to discuss how AI is transforming every stage of investing, from sourcing and diligence to portfolio management, follow-on decisions, and firm operations. John explains why the best venture firms are becoming AI-native organizations, how investors are using large language models today, why every investment memo should be "red teamed" by AI, the rise of MCPs, when firms should build software versus buy it, what Standard Metrics is seeing across more than 12,000 portfolio companies, and why human judgment will become even more valuable as AI automates everything else.

Everyone says AI is transforming companies, but you think AI is going to transform venture capital itself? Why? We're seeing people experiment with new workflows across the entire life cycle of investing from the research, sourcing, kind of the earliest stages of identifying companies to invest in, diligence, and then all the way through to portfolio management and back office. There have been some companies that have been built to assist with various parts of that, but a lot of it right now is grassroots. And so it's just really interesting. Like every time I meet an investor, I'll ask them, hey, what are you using AI for? And oftentimes you get a different answer every time. I think it's actually really exciting time to be an investor

because people are kind of redefining what it means to be an investor in the age of AI and trying lots of new things. What's the most effective way that you know that venture firms are using AI today. A firm is going to meet with an entrepreneur. It's a perfect kind of deep research type use case of, hey, we're going to sit down for this entrepreneur, put together a full dossier on everything that this person's ever worked on, built, professional history, go do research on all of the things that have ever been published about his or her company. Another one that we see a lot is that many firms are generalist firms and or people. at the firms cover many different categories. It's pretty rare these days that you'll have

someone who's just focusing on, you know, semiconductors or just focusing on fintech. Of course, it does occur. But a lot of people are dabbling in a number of different areas and forced to go, in many cases, like mile wide and an inch deep. I think LLMs and, you know, combined with internet research, LLMs, are really good at helping people get up to speed in new technical areas. So, for example, if you're meeting a new AI chip company and you don't come from a semiconductor background, the ability of one of these models to go conduct a lot of research and tailor explanations for you for what you might be hearing from a company that meet your level of technical abilities is pretty incredible. And so we're seeing a lot of investors rely more on LLMs to help them get smart fast

on new technical areas. Probably not enough for them to get all the way over a line, over the line with an investment, but enough where they can actually follow along the narrative arc of a company that might have been just way too technical for them even a few years ago. You mentioned deep research, getting ready for a meeting. What are some of the specific tools that VCs are using in order to be the most ready for those meetings? Maybe three sources of data that we see people rely on. The first is internal, let's call it proprietary data. That might be notes that have been taken,

that might be stored in Notion, previous meetings that have occurred, CRM-related notes, the company being mentioned in various different memos internally. Actually getting all of that kind of all in one place and organized well as a non-trivial task, although I think the advent of MCP has made that easier and easier for people. for people, people will connect to Affinities MCP and they'll connect to Notions MCP and they'll connect to Salesforce MCP or whatever tools they use. So that's one source is just all this internal

information. Another source is information that's just on the internet, news articles, social media, Wikipedia pages, whatever it might be, published academic papers. Information that's always been accessible to people, but it would have been very challenging. to go and crawl through all of that information very, very quickly. SEC filings, right? There's just tons of different sources of data on the internet. And then the last one is industry-specific data products that have been built. So take like a harmonic or a specter, maybe like two examples that are more focused on sourcing data,

where they've gone and they've taken oftentimes public or public-ish data, and they've built like derivative products on top of it that might show you things like, like here's a list of founders that are in stealth mode according to LinkedIn that came from very interesting other startups so they've kind of put a layer of almost like judgment on top of that data and so the combination of you know the internal data the external raw data and then some of this more you know kind of industry-specific curated data I think makes for really interesting research opportunities that let people walk into meetings with a, just like a completely different level of knowledge of what they're

walking into than they did previously. It could be true for an LP meeting too, by the way. Like, if you're sitting down with an LP, you know, if that LP happened to have been on a podcast or, you know, written a paper 15 years ago or whatever, you'd be much more likely to understand that and have the, be able to speak to that or have an interesting conversation about that going in. So I don't think it's just impacting the, you know, VC to founder relationship. I think it's also impacting the way that VC's interact with other investors and LPs and other constituents as well. And are venture firms using clot agents? Are they just putting in a prompt into chat GPT or anthropic?

We see a lot of firms that use, that are, quote-unquote, clod shops. We see firms that use chat GPT, kind of enterprise level across the whole firm. We see firms that do both. We also see firms that use other tools like perplexity and Gemini and the kind of Google family of products. Oftentimes it comes down to individual investor preference. And because a lot of the people that are putting together these workflows are doing so in more of an experimental fashion, it oftentimes is very path dependent on how that person got into AI. If you happen to have a friend who worked at Anthropark or something like that and got into using cloths,

then maybe you would build that workflow in Claude, but someone else in the firm might build it in a completely different tool. So we're still at that place where relatively few of these workflows are enforced in a top-down manner. Some of them are shared and discussed, of course, but oftentimes it's very driven by the individual and what he or she's kind of comfortable using. And the capabilities of these different tools for things like I just discussed, maybe some of these research-oriented tasks are, they're all quite good at doing stuff like this. So it's a little bit maybe more just the preference of the person and then what tools the firm has approved from like a compliance perspective.

Full disclosure, I'm investor in Anthropic. Congratulations. That being said, when I look at these revenue numbers, $60 billion, revenue run rate, and I talk to people in the industry across different verticals, and they're all saying a similar thing, which is most of the AI usage within their organization, outside of developers, they're kind of more mature in their use, is very grassroots, individuals using their own tools, and most organizations have not actually institutionalized or operationalized this as a firm capacity.

Do you see any venture firms operationalizing this? And if so, how common is that? That's a really good question. I have seen some, I think that the workflows, especially for these general purpose tools, like a Claude, where you can literally ask Claude to kind of do anything you want. That's part of the beauty of it, but it's also part of the challenge. There's a sort of blank screen problem that you have where you,

infinite possibilities. Infinite possibilities. So for some of these horizontal tools, I see a lot of the workflows that are being developed start bottoms up, and then sometimes people realize how useful they are, and then they start becoming more broadly established and enforced. For example, there's one person I talked to who had a very specific process that he ran through around effectively read teaming investment memos.

So once an investment memo was published, by someone in the firm, they would effectively go red-team that investment memo and they would use the LLM to go and poke as many holes as possible in the logic and the strategy behind the investment. And then the firm would kind of come together around that and that would help provide some of the structure for their team conversation. It wasn't the only thing that they talked about, I think, but it was, hey, look, here are these like four major themes that came up as we read team this with the LLM and also here these other topics. And I think that was an example of one that was really valuable and then it just started becoming kind of like a policy across

the firm to work on to work that way it became a step in the process of let's make sure that every investment memo is kind of running through a similar process some of the more vertical specific tools that are built for venture capital and standard metrics might be one of them oftentimes those are ones that are easier to go and build kind of like full institutional buy-in from day one because they probably go and fulfill a certain very specific workflow that the firm needs to do for example, like portfolio reporting. So we're seeing a little bit of different behavior where there's some tools that are coming top down

and then a lot of grassroots workflows that are being built bottoms up. And I think the interesting place is actually the intersection of those two where because a lot of companies are now building very agent-friendly, you know, connectors like MCP, for example, a lot of firms are starting to go in,

quote-unquote, like vertical AI tools for their firm, but then they're also starting to go and build a lot of their own very bespoke workflows and even kind of software on top of those. So we're sort of seeing a layering now of those two. MCP model context protocol. Tell me about these MCPs and what exactly do they accomplish? The way I think about MCP is as a layer that sits on top of your application programming interface or API that makes it really easy for external large language models, AI tools,

agents to go and interface with your product, but allows the user to stay within the realm of prompting. So for example, we have an API for our product, and there's a way of querying the API as a customer where you could say, hey, I want you to return this data point at this time for this company. It's a function, a programming function, and there's a bunch of inputs, and there's an output that's received. It's very structured, and it needs to be done in a very specific way. What an MCP allows you to do is a layer of abstraction above that where you could go to Claude and you could say, hey, Claude, build me a report on this company. And it'll go in and it'll do all the appropriate underlying API calls. It'll perform actions on top of all of that data and it'll

present it to you. So it makes it much easier for a human being to use an API. It kind of answers the questions to how software firms could continue to stay relevant with this advertiser. of these large LLM models and how they could partner with the LLM models. Sometimes the phrase that people are using these days is, you know, headless. I think Mark Benioff used that. The idea that, you know, you may not want to interface directly with the product. You may want to stay at the level of prompting or working in an AI tool. The underlying work on kind of organizing the data, the ontology of the data,

the surfacing of the data might be the heavy, lifting might be done by another product, but it's like a little bit of an iceberg where you're experiencing the tip of the iceberg and there's all this stuff happening below the surface. And for a lot of users, it's very empowering. For example, our products, it's a data-rich product. And so sometimes there might be a user who needs to get something from standard metrics, but they're on the go, they're super busy, you know, they're looking for an answer to a question. It might be a much better experience for that person to prompt an agent to grab that data for them than to log in and have a very kind of data, heavy visualization type experience. There might be other users that really want to have that much more controlled experience, and so it's actually useful to be able to support both.

Prior standard metrics, you were a VC for six years. You were at Spark Capital. If you were starting a VC firm today and you wanted to make it AI native from the get-go, what would you do? A lot of it comes down to the culture of the people because the tools are improving so quickly. And every day, week, month, year, there's so many new things that are getting launched and built that help people to be more effective at their jobs. And it's true in venture capital and private equity where we focus. But it's also true, obviously, in every other industry. And I think it's actually becoming easier and easier to adopt these tools, too. There's more guidance, there's more services that are available.

If you look at Open AI and Anthropic, they're building out these forward deployed models, where they actually go and help people to determine how to do this. There's third party service providers that'll do that. Also just people are getting better at bringing their learnings from their personal lives and other aspects of their work life and helping to kind of iterate and experiment. So I think the most important thing is the culture of the team.

If you wanted to build an AI native firm, I think you kind of need to start with people who are AI native. And or people who want to become AI native. I've met so many people who, you know, didn't use a lot of AI tools for a long time. And then one day they decided, I want to go learn about this. And then you fast forward a couple of months, and they've learned a tremendous amount. So I think the most important thing is probably the people that really desire to build that kind of firm. And then making smart decisions around process and tooling, I think, flows from the culture of the culture of the business. the people that are getting built.

Everyone I talked to on the show is chasing the same thing, an edge. And more and more, the edge comes down to your information, not just having it, but being able to trust it when the stakes are highest. AI is doing more of the information gathering for you every day, and most tools are very good at sounding right. The summary reads clean, but can you trace it back to the filing, the transcript, the specific passage that drove the answer? Or are you just trusting the confidence of the output?

For investors, that's not a minor concern. A missed filing, a misweight of source, a context that got lost somewhere in the retrieval chain. Those aren't edge cases. They're how decisions go wrong. AlphaSense is the AI market intelligence platform built specifically for this. They own the content. Over 500 million curated documents from broker research and expert transcripts to filings and earning calls, and they own the retrieval layer on top of it.

So every answer links back to an exact verifiable source because the answer is only as good as what's underneath it. And with Alpha Sense, you know exactly what that is. The edge goes to whoever could trust their information and prove it. See it for yourself. Start your free trial at Alpha-Sense.com slash how I invest. That's Alpha-Sense.com, how I invest. Have you seen full-time AI engineers internally at venture capital? Yes.

Is that something you see as a best practice? It's becoming extremely common. And I think it's a really, interesting idea. I think one of the big questions that we ask ourselves a lot is how much software our firm is going to build themselves, which is relevant to our business, because we sell software to firms. So it's sort of in our best interest for firms to be very AI native and want to build certain types of software. Of course we have the perspective that firms shouldn't build their entire software stack from scratch. I can talk

through that specifically if you're interested in why, especially multiplayer software. I'm curious why not build your own software. So I was actually with a prospective customer right before this, talking about this very question. When we started the company six years ago, we would sometimes meet firms that decided to go and build their own portfolio management software or other types of software in-house. And most of them ran into trouble at some point. in the kind of, let's call it, pre-coding agents era, someone would build something, and then that person would inevitably leave,

and the team would inherit this, like, unmaintainable, difficult thing. Not every single firm, but that was what we saw most commonly. Now in the coding agents era, that still is a concern, but it's so much faster and easier to build software than it was before, that I think that's actually, that barrier has gotten lower.

So I think it's actually a good question. I think you really need to justify why would I buy versus building. I think one reason to buy software versus build software is, you know, there's obviously considerations around continuity, maintenance, you know, dealing with sensitive data, that kind of stuff. Another big one is who needs to use this software? Is it just us? or is it other constituents outside of the firm? And that's where this distinction between single player and multiplayer really makes a big difference.

Imagine you worked at a firm and you wanted to build a new research tool that will allow you to go conduct really, really deep research on a specific founder or a company and, you know, run it through some sort of interesting internal process. That's actually probably a really good example saying you should build yourself because every firm kind of has a different taste, and it's just the internal team using it. It doesn't need to be used by thousands of people. The stakes are not low, but if it goes down for some reason,

okay, well, that's too bad, we'll get it back up and running, you know. On the flip side, stuff like what we work on, which is, you know, this is a tool that's getting used by a firm. It's also getting used by a firm's portfolio companies that are not part of the firm. They are their own independent entities. Also, my, by the way, be getting used by the firm's auditors. And the auditors probably care about the verifiability of the audit.

And where is the data come from? And a trusted third party, you know, if an auditor was looking through a firm's books, for example, to try to justify evaluation. And a firm said, hey, we did this analysis based on like revenue and EBITDA and here the revenue and EBITDA numbers. if those revenue in eBay. Numbers come with an audit trail directly from the portfolio company on a trusted third party platform

it probably feels pretty different to the auditors than if it's on like a vibe-coded system that the firm built themselves. One of them is a little bit more verifiable to an outsider, whereas one is like, okay, well, it's cool that you're showing us this, but this doesn't actually help the audit. We need to see the underlying source documents

and be able to trace that back and we might even go back to that company and ask them to share that data with us again. So I think the single-player versus multiplayer is going to be a really, really important distinction for where buy versus build makes more sense. And what we've seen so far is that firms are much less eager to build their own software in a place where they need to be responsible for the user experience, data privacy, security,

compliance needs of other people outside of their firm versus just their own internal needs, which also might be important for them too, things like GDPR and things like them. Also the whole concept of, is it a network effect business? In other words, as more users use it, they input more data, and is there an inherent data mode in the product?

Totally, yeah. Those are two other elements that I think are extremely important, which really tie back to the quality of the user experience. If you're working with your portfolio companies with a specific product, and most of those companies are already using it before you sign up,

Is it faster to onboard? Is it faster to get value? Is it better on the other side for portfolio companies to work with an investor? If they're already using that product with their other investors, we think the answer to that obviously is yes. And it's definitely borne out in our data. So there's, yeah, network effects are important. And then the other side of it is where can you build interesting proprietary data assets, you know, in 2026?

Private markets are actually an interesting case where you're dealing with a lot of, sensitive, private, non-public data that, you know, individually can never be shared, but there is interesting opportunities to build derivative, aggregated, and anonymized data products that set on top of that with enough scale, which is something that we've, you know, we've spent time working on and building around. You have a really interesting vantage point in that you're not only building a software product for VCs, but you're constantly. talking to them about their infrastructure and their strategy.

Yep. In the next five years, what do you think becomes commoditized within venture, and what do you think the alpha comes from? That's interesting. I was talking about this the other day with somebody around the idea of e-vals, building kind of evaluations for different workflows that AI agents are executing, and the sort of frontier of where human beings create the most value. continues to push more and more toward workflows or activities where it's very difficult to build a good eval.

Coding is a great example of a place where so much more of the coding work is being done by agents than it was in the past at almost every company. But the number of software engineers that are getting hired is actually increasing over time, which is counterintuitive for a lot of people. And then the questions, well, how is that the case? And there's just interesting non-zero-sum aspect of software engineering, which is that there's literally a million problems to go and solve. So as it becomes more and more efficient to solve problems, it pushes human beings to go and spend more time on tasks that require lots of taste and judgment. And it turns out that there's so many of those things to do that the more and more coding agents there are, the more and more human beings we're hiring. We'll see if that trend persists forever.

it's kind of the same thing. Parting data out of documents, that is not something that firms are going to be doing, you know, by hand. Most firms still do that by hand today, and products like ours and others are helping to alleviate that strain because you can build really good e-vals around that. You can say, hey, let's parse like a thousand cash flow statements, and then let's figure out how accurate we were at doing that. And then let's go and update our agent harness to do a better and better job of parsing

cash flow statements. And by the way, the models are also getting better. And you can just sort of relentlessly focus on improving that flow. It's a closed loop. You're getting immediate feedback. Totally. And some of them are harder problems than others. For example, we do board decks as well as financial statements. And board decks are much more challenging than financial statements because they're much less structured and they're much longer. And they tend to be in PDF format, which can have challenges compared to the rows and columns of a spreadsheet. So an activity like parsing data out of documents, I think is a good example of one. that, you know, ought to become fully automated.

Doing that well is actually quite hard, but those types of activities are probably not where a firm goes and pitches in LP and says, oh, we're going to be the best in the world at X. I think where we start to see human beings, you know, lean in more and more is the human aspects of the job. The valuation and support of founders, for example, is something that... Value at. value add and also the personal evaluation.

Can you build a good eval for determining if a founder is a top quality founder based on hours spent together discussing the business, you know, friends calls, you know, all of the sort of the product planning and thinking and strategy? Maybe someday one can kind of get closer and closer to doing that. But I don't know, my perspective on that is there is going to be a lot of very important kind of human judgment-centric work that ties in with the way that investors work with and support founders, post-investment as well. When I think about investors that have been the most helpful to us along the way, I think about specific people who went out of their way to proactively go do things for us that move the needle for our business, whether it was introducing us to a customer, that bought our product or introducing us to an employee who came and joined our team or introducing us to another investor that invested into our company or something along those lines. That's actually a way that investors have to meaningfully move the needle on the near-term performance and progress of the investments that they make.

But investors oftentimes are bogged down with so many other things that they don't have as much time as they'd like to actually work with their portfolio companies. So I think that's another example of like the investors end up feeling like a closer and closer member of the team over time because there's less busy work to do. they're able to focus more of their time on, you know, adding value. There's a cognitive aspect to this, and this is why people are so concerned about AI disrupting jobs because they're not able to see what happens, a second and third order effect of AI making things more efficient because they're so consumed with what they have to do

on the daily basis. They don't think, what if that two, three, four times more time and more energy to spend on other things? They only focus on, well, today, I have to do all these things. So what happens if you take all these things away from me? They're not able to think two, three steps ahead. But in reality, when you see businesses as they grow, they bring in new talent. That talent starts to solve different problems. Now the CEO has more time to reflect, to think. Now they start new businesses.

And now there's new problems. So businesses want to grow, all things being equal. But it's very difficult when you're in the trenches and you're just overworked and have no time to really think about these net new opportunities. When I talk to founder friends, especially about kind of AI and employment, one thing that I think a lot of founders feel is this deep sense of being overwhelmed by the long backlog and laundry list of all the things that you wish you could spend time on, but you can't. And this sense of, gosh, like, if I only had more time, more resources, but I don't. So I'm going to focus my time on the most, ideally, mission critical, strategic activities that are going to have the highest probability

of moving my business forward. That's effectively what really good planning is, right? Really good planning is, okay, we are going to, we have a limited amount of resources, limited amount of time, what are the things we can focus our time on that are going to be the biggest needle move for our business? But it turns out that there's probably a lot of other things

you could also be doing. that would also be adding value in kind of compounding in various different ways. And as you get better and better at using AI to help those things go faster, it affords you the ability to tackle a much broader set of challenges and work your way down that backlog. And that backlog effectively is an infinite backlog. I mean, maybe there's some end to it. And David Deutsch talks about this in the beginning of infinity,

that as you innovate, there's an infinite amount of innovation. Why? Because the innovations start to innovate on each other. Oftentimes, the innovation is the result of multiple new innovations. People always see this retroactively. When the television was invented, I think, like three people invented within a year. Why? Because all these technologies were now available that made it all possible. So they see it in retroactively. They don't think that there's something that hasn't been invented yet,

that if there's a couple of new innovations that could come about, now you can have a whole new technology. It's really interesting to consider the negative side of this, though, too, which is because suddenly everyone gets more productive and certainly feels more productive, and you start to open up that backlog. It'll be interesting to see what happens from a product experience and design perspective. Because I think if companies aren't careful, they can also just do so many different things that they actually go past that optimal point, perhaps, and then their products start to feel

overly complex and overly convoluted, and they start to accrue all sorts of technical debt that they didn't probably need to. There's probably a limit in some sense for how many things people should do, and maybe some people kind of move past that optimal point, but I think the common feeling of being a startup founder is sort of feeling like you're drowning, because there are so many different things that you ought to be doing, and I do think the AI is this incredibly empowering force to help people to kind of pull their heads above the water a little bit. And it's true for investors too. I mean, imagine just the thousands of companies that you wish you had time to go research and meet. And it's easier to work with a much larger volume now. There's a great thought

experiment that two different people, two very different people, I brought up, Peter Thiel likes to ask this question. If you had to scale 10x in the next six months, what would you do if I forced you to scale 10 times? Alex Hermosie talks about what is the one thing that you could accomplish that would make the next 10 things irrelevant. So, for example, for us, sometimes I find myself stuck on a micro issue, oh, I need a new sponsor for the podcast, and then I'm like, what am I thinking about?

I could just do a new fund or new product, and that pays for 1,000 years of sponsors. So it's easy to get stuck in these problems that don't actually have to be solved if you solve other problems. Growing up, I thought managing money meant paying bills and balancing a checkbook. But as you know,

that is only a small piece of the financial puzzle. Managing your money takes more than just checking your bank account every once in a while, and great financial decisions come from having a complete picture and proactive management of your income, expenses, and investments. Take control of your finances with Monarch. It brings together all of your accounts, investments, saving goals, and spending into one place, making it much easier to understand where your money is going and whether you're actually on track to achieve your financial goals.

Why, like most about Monarch is that doesn't just tell me what already happened. It helps me plan ahead. The AI system lets me ask questions about my finances in plain English, and the AI weekly recap, highlight spending changes or upcoming expenses before they become surprises. It's like having a financial advisor in your pocket. Write your own money story with Monarch. Use code invest at Monarch.com to get your first year of Monarch core half off at just $50. That's 50% off your first year at Monarch.com with code invest.

Support for today's episode comes from Square, the all in one way for business owners to take payments, book appointments, manned staff, and keep everything running in one place. Whether you're selling lattes, cutting hair, running a boutique, or managing a service business, Square helps you run your business without running yourself into the ground. I was actually thinking about this other day when I stopped by a local cafe here. They use Square and everything just works. Check out as fast, receipts are instant, sometimes I even get loyalty rewards automatically. There's something about businesses that use Square.

They just feel more put together. The experience is smoother for them and it's smoother for me as a customer. Square makes it easy to sell wherever your customers are. in store online, on your phone, or even at pop-ups, and everything stay synced in real time. You could track sales, manage inventory, book appointments, and see reports instantly, whether you're in your shop or on the go. And when you make a sale, you don't have to wait days to get paid. Square gives you fast access to your earnings through Square checking.

They also have built-in tools like loyalty and marketing, so your best customers keep coming back. And right now, you can get up to $200 off Square hardware when you sign up at square.com slash go slash how I invest. That's SQU-U-A-R-E dot com slash go-slash how I invest. With Square, you get all the tools to run your business with none of the contracts or complexity. Run your business smarter or Square. Get started today.

Support for today's episode comes from Square, the all-in-one way for business owners to take payments, book appointments, manned staff, and keep everything running in one place. Whether you're selling lattes, cutting hair, running a boutique, or managing a service business, Square helps you run your business without running yourself into the ground. I was actually thinking about this other day when I stopped by a local cafe here. They use Square and everything just works. Check out is fast, receipts are instant. and sometimes I even get loyalty rewards automatically. There's something about businesses that use Square.

They just feel more put together. The experience is smoother for them, and it's smoother for me as a customer. Square makes it easy to sell wherever your customers are. In store, online, on your phone, or even at pop-ups, and everything stay synced in real-time. You could track sales, manage inventory, book appointments, and see reports instantly whether you're in your shop or on the go. And when you make a sale, you don't have to wait days to get paid. Square gives you fast access to your earnings through Square checking.

They also have built-in tools like loyalty and marketing, so your best customers keep coming back. And right now, you can get up to $200 off Square hardware when you sign up at square.com slash go slash how I invest. That's SQU-A-R-E.com slash go slash how I invest. With Square, you get all the tools to run your business with none of the contracts or complexity. Run your business smarter or Square. Get started today. Framework a lot.

I think it's true for investors too. I think, you know, there's this expression in startups where it's kind of like growth solves all problems, which I don't think is always true, but it certainly is true that it's a lot easier to let fires burn or fires seem smaller when you're growing a lot. And I think as an investor, it's also true that it's solving smaller problems might seem less important if you get to invest in in Anthropic. That might make some of those other problems. As you were mentioning, investing, I was thinking about SpaceX also was investing in SpaceX. But when I look at a couple different parties around it. Antonio Gracius. He invested in it 30 times. It was publicly made

of, I think the number was he returned 127 billion. I think it was somewhere around that. And Lucas Nosek, it was a former Founders Fund partner. I remember meeting him at Founders Fund. He told me that he was starting a fund based on SpaceX called Gigafund. I thought that was the most absurd idea. And of course, the 137 ventures. I think they owned roughly 1% of SpaceX. So there are some investors that take this thought experiment literally and realize, well, SpaceX is once-in-a-generation company, I should just keep on backing up the truck. There's obviously portfolio construction limitations to that, and it's not for the faint of heart, but it also applies to investing as well, not just productivity.

It's impressive. Some of these folks that backed up the truck for, in some cases, I don't know when Founders Fund made its first SpaceX investment, but it must have been well over 15 years ago, right? It's like a very, very long period of time. One thing we think about a lot at our company is how do investors handle everything that happens after that initial investment? Because people spend so much time and energy thinking about making a new investment, the rigor, the customer calls, the references, the research, the memo, the debate, the investment committee. But then for a lot of investors, after that investment occurs, the next investment in that company might be a quick, conversation at the partner meeting. Oh yeah, the company's doing great. They're racing their series A.

We own 15% let's do our probrata. Everyone sounds good. Okay, great, let's do it. But in reality, some of those conversations, you know, some of those decisions probably ought to have a lot more scrutiny applied to them. Particularly when you look at the percentage of these funds that are used for reserves, you know, it depends on the construction, but a lot of these funds will be 30, 40, 50%, percent reserves. So if you think about how does the firm generate, you know, alpha, good reserves planning and really prudent and oftentimes maybe more aggressive follow-on investment decisions can have a huge impact on fund performance, but probably get a lot less scrutiny than they deserve. And that's something that we're interested in kind of as a theme and an opportunity to help the industry. Speaking of Founders Fund, they've famous for concentrating in their winners and making so many

fund returners on their winners. Hunter and Satya from Homebrew says that every single round for every investment you need to be a net buyer and net seller. There's no hold. There's no pro rata that's just lazy thinking. That's an interesting framework, net buyer or net seller. You have over 12,000 companies on your platform, some reporting, some not reporting, but you have this rich data set. You do research on this data set. What does surprise you the most about what you found? It's interesting. We find new things in the data every time we run these analyses and we're also learning and we've hired recently brought on a new exec on our team named Ani Kotiash who was running portfolio analytics at battery

ventures and was previously a white combinator who's really helped us to kind of like level up some of the analysis that we're doing around this we see surprising things in the data every time and some of them are very clear reflections of trends that are happening externally and some of them are almost like counterintuitive so I'll give you a couple of examples. So like when the ZERP crash happened, you know, late 2022, early 2023, we saw huge swings in revenue growth rates and profitability, as you might expect, right? Growth rates went down, profitability went up, companies tighten their belts, you know, ARR per FTE or revenue per FTE, probably a better way of thinking about it.

you know, it was kind of climbing across all sorts of different segments, sectors, stages, et cetera. People got more efficient at generating revenue because funding dried up. Then over the last couple of years, you see this boom in very well-funded AI companies. And there's also this narrative that you might see it on LinkedIn or something like that, which shows like, oh, we're like a three-person company and we have like 30 million in ARR. I'm sure you've seen these. And seats trapping. Yeah.

And those companies do exist, of course. But interestingly, what we found in our data was that if you look at the AI companies versus the non-AI companies, for a long time, the non-AI companies were actually more efficient on a revenue per FTE basis than the AI companies. Because we think access to capital was so much more readily available for the AI companies that they would go raise a big round. They'd go hire tons of people and SDRs and like tons of engineers. And that gap has closed.

And we just published, I think, with Emergence last week, a short kind of add-on to a report that we did with them that showed that for the first time, revenue per FTE for the $100 million-plus revenue companies had actually crossed. And now the AI companies were actually producing more on a per-employee basis than the non-A-I companies. So there's these interesting things that you see in the data,

and you also start to see really large changes over time. The data from 2021, is not relevant, really, to what's going on in 2023. The growth rates of the top decile companies in 2023 are not relevant to what's happening in 26. There are these major swings that occur that are driven by some of these underlying dynamics with interest rates and also new technology, of course.

Tell me about the metrics behind top decile companies today in 26. One slice that we find useful to take are AI companies versus non-AI companies. And it's a little bit difficult to distinguish between the two, to be honest with you. We do our best. Companies sort of self-identify on our platform, and they can adjust that and change that over time. And we do our best to verify that. But there's a pretty fine line between a software company that builds a bunch of really useful AI tooling into their product

and that their customers use and, like, using. And a company that, you know, sort of is like truly AI-native, day one kind of in the token flow. Those two things may also converge more and more over time. But regardless, there's a very large difference between the AI companies and non-AI companies when it comes to growth at the top decile. If you look at median growth rates across different revenue bands, AI companies are typically growing faster than their counterparts.

I don't have exact data points for you, unfortunately, like during this podcast, I wish I did. But if you look at the top decile, that's where you see like a huge, huge swing. The fastest growing AI companies are growing faster than any companies ever in human existence. That's why they're such a focus on them. Yeah. So it's interesting. You end up learning a lot more from the top desal, I think, in certain cases than you would just looking at media. And that's one reason why having a huge data set is very useful. How would you explain that intuition? Why are the top desal AI companies just growing so much faster than everybody else? That's a really good question from first principles, it would be that either the level of product market fit is so much stronger

and or the ability to distribute effectively is so much stronger, and probably some combination of the two, of course, because product market fit can help to drive distribution. And then distribution also is driven by fundraising, too. So companies with incredible access to capital, the best teams, the most backable teams that are able to raise capital very aggressively, very early, with amazing products, those companies have significant advantages in terms of distribution. And if they can translate those advantages into just a radically superior product experience, it can lead to pretty unprecedented growth.

The other thing I'd say is that people talk a lot about gross margins. I think that one thing that's kind of interesting about the AI era is that, you investors' attitudes toward gross margins have changed a little bit, meaning people still care about gross margins, but I think a lot of people are taking a pretty forward-looking view on gross margins. Not necessarily what are gross margins today, but hey, this is a venture bet. If this company is successful in doing what they say they're going to do, what does the long-term gross margin structure of this company look like? And hey, this company might need to raise a billion dollars in capital to get there, but if the end result is a company that's worth, tens of billions of dollars they're going to be able to go and do that and so we saw you know

this this host of companies with you know relatively mediocre gross margins and in certain cases maybe perhaps negative gross margins if the full accounting was done be able to go raise enough capital to go and really scale incredibly rapidly to get to that place or suddenly the gross margins end up looking pretty good they didn't get stuck in that valley of death the market was forward-thinking enough and efficient enough that it allowed those companies to kind of hop over that valley and get to the place where I guess are generating significant amounts of gross profit. It's just pretty interesting that it played out that way. And I think that the net result is now a lot of companies that have kind of crossed that chasm. And they're now very powerful.

It makes absolute sense if you think about it from the power law lens, which is all the returns come from the biggest high flyer. and if a company's growing really fast, even if it has lower margins, you're really taking one bet, which is that they could improve their margins. And if they can, then it could be a 10, 50, 100,000 X return. Totally. Power law versus in private equity. You can't have 50, 70% of your companies go out of business, and it would be nonsensical to make that bet.

That's totally right. And then I think the interesting middle case between private equity and venture capital are going to be, you know, things like AI Native Services companies, right? that sort of can have venture like growth, but maybe map a little bit more to the type of business model that one would have seen on the private equity side. I'm really excited about that category.

And we, in some sense, consider ourselves, we provide a lot of heavily AI-assisted services, and we also, of course, sell software. So we're kind of straddling the lines between those two business models. There's this thesis that as software becomes easier to produce, the next layers where the value is going to come. The services, the etch cases,

helping customers deal with their problems that are not directly solved by the product. What do you think about that? In many cases, that's absolutely true. It kind of gets back a little bit to the question around, you know, network effects and single player versus multiplayer. There's some question, I think, around responsibility.

Like, who's responsible for something? AI can't be punished for making a mistake. There's sort of a level of responsibility that a human being can take, that a model can never take. And so you want to engage with third parties sometimes to take on very important tasks where somebody needs to. I think it has to do to some extent with just liability. There's a benefit to others taking responsibilities and liabilities off of your plate, especially if they're not your core competency. and there's a benefit for human beings being held accountable for the quality of that work. It's almost another form of cognitive load.

There's only so much responsibility you could take on. Yeah, it's a form of cognitive load. It helps you go to bed at night in the future if there's sort of an issue, there's somebody on the other end, even if it's not like a legal issue, there's somebody who is responsible for making things right. Whereas if you insource every activity, if you're your own law firm and you're your own software, you know, developer and there's just so many things you could take on your own plate. If those things aren't your core competency, there's, you know, distraction-related challenges.

And then there's also this challenge that comes up, which is like, if something goes wrong, like I have to be the one to go and fix it versus somebody else is very incentivized to go and fix it. So, yeah, maybe cognitive load is a good way of thinking about it. Pep peeve with my wife sometimes. I'll ask her to find a place for us to go grab dinner and she'll come back to me with three choices and I just wanted to make a decision yeah I just wanted to completely be just decided by somebody else there's only so many problems I could deal with that analogy is great and the other analogy that I like is like when you talk to people I remember having this experience

who like bought their first apartment or their first house about what it's like one of the first things people always say is when something breaks now it now I have to go and figure it out you know it's like if there's a problem with the toilet or there's a problem with the sink I can't call the super and someone comes and fixed it. Like, I got to be the one to go and find the right subcontractor and go do it. So look, in certain cases, people are going to internalize more things. If people are more productive with AI, it gives people the opportunity to go and internalize more things. However, the opportunity cost is that instead of, you could, instead of internalizing those things,

you could instead go and take on more areas in your core competency. You know, if you're a venture capital firm, you could take off. on a wider scope where you could take on a broader strategy or you could go talk to more entrepreneurs or spend more time helping your portfolio companies. So I think every company will evaluate to some extent on a case-by-case basis where it makes sense for them to use AI to internalize more activities outside of their core competency and where it makes sense for them to continue to rely on others to be responsible for that work, probably in a more AI-assisted fashion, and go deeper into their core competency.

That's kind of interesting, broader economic question, I think. How do you see AI changing the relationship between you and your customers? That's a really interesting question. One thing I will say is that nothing beats getting on a plane or Ubering over or whatever to go see a customer in person. The trust that's built in person, I think, is something that is impossible to replicate and may never be possible to replicate through technology. particularly per our previous conversation, the idea of, you know,

externalizing responsibility for certain core areas or sharing responsibility for certain core areas. I do think that in-person conversations are incredibly important. So that hasn't changed. Although interestingly, one thing that we've been doing, and I know that other software companies do too, is actually sitting down with customers in person to work on AI-related projects together. Because oftentimes there can be a gap in understanding. understanding sometimes where the customer isn't as familiar with using different AI tools.

For example, a lot of our customers use our MCP to go and build a bunch of really cool automations, but some firms have never used an MCP before. So there might be knowledge gap. There's also knowledge gap on the other side, which is that sometimes our customers are using our product in ways that we didn't necessarily design. And we're learning from them. That's a good sign. It's a great sign. Yeah.

I mean, I think it's so exciting. Like, we'll get these reports back. like, oh, this customers, you know, found this really creative mechanism for doing X, Y, and Z. And we're like, oh, that's so cool. So I think that the in-person stuff for us is critically important. I think where it helps us in terms of the customer relationship, we've gotten a little bit better, and we're still working on this, processing a lot more qualitative and quantitative information alongside our customers to show them all the different ways that they're using and getting value from our product, or maybe gaps in our product.

And so the idea of like a partnership review, right? Okay, let's sit down. Let's talk about how things are going. That's one area where it does feel like there's an opportunity to have a much, much richer conversation than there was in the past. Because going all of these different gong calls and dashboards and data, it was doable, but there's a lot of work and you can kind of automate more of. that now and make that make that easier. We work with a lot of our customers in Slack also. And, you know, I think that there's going to be some cool ways for, over time, for more, let's call it like external agents to become kind of part of that workflow. My simple response

to your question would be like the direct conversations and whether it's on Zoom, but especially, you know, at least occasionally in person, that is as valuable or more valuable than ever. And for products like ours, where it's a considered purchase, where there's at least an important shared level of responsibility around the outcomes, I think that that will continue to be really, really important for companies like ours. If you go back six years ago and you could give yourself one timeless piece of advice, what would that be? There's certain things that you learn building a company that are difficult to fully

internalize until you go through them yourself. Like words of wisdom about company building that you could. read in a book you could understand theoretically but it's only until you've actually gone through it and or made mistakes that you fully internalize one example that comes to mind for me the idea of hiring people for a startup that really care about that startup and really care about the mission of that startup and are aligned with the values that startup I think I had this idea in the early days of standard metrics sometimes that the goal was to hire really smart people and then

you could figure out how to mold them into your values. You know, as long as they were like a reasonably good values fit, you can kind of mold them into your values and as long as they kind of cared about the mission, you can get them really excited about your mission. I think for some people that's true. But for a lot of people, building a startup is really hard and you need people to be incredibly well aligned on the way in, which forces you to be much tougher and much more stringent about how you hire, not just on the, hey, this person did great on the, you know, coding interviews,

or has great references, or crush their quota at their last company or whatever, the evaluation process is for their function, but more figuring out, like, why do they want to be here? So if I could go back in time and give myself advice before starting standard metrics, I think the biggest piece of advice that I would give myself is hire slow and screen relentlessly for mission and values alignment and be more comfortable saying no to really talented smart people because they're not a good mission and values match. Is that because startups, as you mentioned, are hard and when times get tough, they're going to leave to another firm or mentally check out? Yeah, that's part of it. I think that part of it is, you know, what's the reason why this person gets up for the fifth time after they've gotten, you know, bad news or punched in the face or their colleague is left or the customer issue happened or they lost the deal or whatever it might be. And then the other part of it is that I think that especially in a small company, you know, we're 68 people now, so we're not tiny, but we're not a big company by any stretch of the imagination.

Every person has such a deep impact on the culture of the company. and it only takes one or two people who are not aligned with the values of a company or the certain way of doing things for, you know, distractions to emerge. And we're talking a little bit before about like the backlog and the core competencies and, you know, where should you focus your time. I think that when you look around and everyone is on the same page for why we're here and why what we're doing matters and how we do business and how we work together, it makes it so much easier to reduce that mental load and just focus on the work.

You're working on the mission, not in essentially babysitting and aligning people to that mission. One of our company, so one of our four core values is work in the open. The idea is that we lean into transparency at our company. There are certain things that we don't work in the open on. For example, we don't do like performance reviews in front of the whole company or whatever. Some companies do, like I think a Bridgewater or something like that probably does. But the default at our company is if you have a meeting, write it down, it goes in notion. Everyone can see it.

If you're creating a doc, write it down. It goes a notion. Everyone can see it. Slack default to open channels. If you're telling someone you're running late, it's fine. If you're talking about something that's really sensitive, totally fine. But in general, make things discoverable to other people. And it turns out that that's now extraordinarily valuable because of AI tools.

Suddenly we have this exhaust that, especially via... to MCPs, think like Slack, Notion, Salesforce, all these tools, you can start to get really, really valuable insights for what's going on across the whole company. But even before MCPs was really useful, there's certain people that we've hired over the years that did not buy into that. And I kind of thought to myself, like, I'm going to work on that. I'm going to figure out how to get them aligned with that. And that's a risky endeavor.

And it turns out that if you're kind of misaligned with one of those core values, can create so many downstream challenges. And then other people say, oh, well, you know, David's not, you know, taking all their meeting notes and notion. Like, I don't need to either. And it's kind of like spreading on the team. That's probably one of the biggest things that I've learned is knowing that people can change and having a growth mindset, I think, is important.

But relentlessly screening for values in alignment and mission alignment on the way in is kind of a critical tool. And that's something that I continue to interview every single. person that we hire and I plan to do that as long as I possibly can for mostly for that reason. My second master's is in psychology and gives me an unique lens into to these problems. And one of the new ideas in psychology is this concept of internal family systems. If you want to look at there's a kid depressed or a kid has anxiety, oftentimes it's not actually, you can't just look at it in a vacuum.

You have to look at it within the family system. And oftentimes it's these family systems that have these dynamics that actually cause this anxiety or these problems in the kids. And same goes with organizations. You can't just take the person outside of the organization, try to psychoanalyze them or understand what's going on. Oftentimes it's this effect of the organization having. And sometimes you take one part out and the problem goes away. Sometimes you put it one person in. The problem comes back.

So it's much more interconnected the way organizations work than most people realize. I think it's a really good point and it actually, there's actually a couple of interesting downstream implications of that. Like one of them is, you know, how do you assess somebody well if they haven't actually worked inside of your company? Everything that you're doing during an interview process is kind of a simulation, right?

It's like, we're gonna meet the head of sales, you're gonna present this thing, you're gonna have this interview with John. It's all trying to get at predicting how well this is going to perform within your company and what kind of team member you're going to be. Some companies have taken the approach of these effectively like firing up these working relationships before someone becomes a full-time member of the team. How do you do that?

There's one way that we do that. We don't do that with most employees. We have done a lot of internships that have converted into full-time roles, especially for folks that are coming out of undergrad or in certain cases out of business school. That has been really powerful for us. Because if someone spends a summer at your company and is just an unbelievable team member and you've gotten to know them really well and you've seen them under pressure and you've seen them perform and you've seen how they align with your values and how deeply they're kind of embodying the mission of your company, you have a really good sense for what they're going to be like as a full-time employee. So I think the hit rate there for us, you know, knock on wood has been very, very high. And then the other thing that I think is interesting is like people oftentimes in the earliest days form a very cohesive startup kind of culture by hiring people that they've worked with and that they already know.

That's your risks. Which has pros and cons. It's not perfect, but if somebody's worked with someone on the team professionally, and that person is a great, is like a great, member of our team and they strongly vouched for them and they bring them into a recruiting process. That doesn't mean that we don't go through our full process and kind of try to evaluate them very strictly, but I think there's a much higher probability that that person is going to sort of work very well within our company because that person who's referring them has a strong incentive for them to work out. If they don't work out, it's kind of a painful experience for

everyone. So there's this additional layer of filtering from the team that oftentimes extends into that area. There's a couple of ways we sort of get at it. Have you seen any interesting examples of companies that do these? I don't even remember what people are calling it now, but these kind of like work practice or kind of consulting engagements before they start working together and that kind of stuff. I've seen on a customer basis the fully deployed engineers, Palantir would send over the engineer on Friday to work on your problem

before everyone else, their competitors pitched on Monday. So they would literally give you a work sample. Yeah. We use internships as well. Yeah. It's a big problem. And a lot of the top CEOs that run organizations, with 10, 50,000 people, they've oftentimes, much of them, many of them have told me that

they know whether someone's going to work out the second week. Yeah. And then the question is, how do you, how do you compact that? I think internships are a great thing. And whenever possible, doing projects, which I think with any form of information, the goal is not to reduce 100% of the errors, but can you, is there an 80-20? Is there a one-week project that you could give or a case study that you could give a one-day case study that maybe takes your failure rate from 50% to 20%. That's a massive difference just

right there. So I think it's an unsolved problem. Yeah. The other thing that we've tried to think about a little bit is how do we make sure that we do everything within our power to make sure that people leave their onboarding process aligned, excited, and well-ramped. One of the challenges that we face, which is also strength in other ways, is that we are a remote first company. We have an office in San Francisco. We have an office here in New York, but we also have team members all over the world. One change that we made over the last year, which I'm actually really excited about, is that whenever a new person joins our company, we fly them either to San Francisco or New York for the better part of a week for onboarding with their manager. That act helps to really

reinforce and strengthen some of those more kind of qualitative aspects that we discussed before. But yeah, it does feel like an unsolved problem. There's so many people that I would love to say, hey, come work at a company for a month. We'll pay you handsomely for that time, you know, if it's a great fit, amazing, if it's not, no hard feelings. I think the challenge that we face there is just that people are so busy, you know. People have so much going on, you know, oftentimes it's an existing job or it's hard to do that. But I know other companies, I'm trying to remember which ones, but there's a few other companies that I've seen that have leaned in hard there. and sometimes only hired people after they've done,

at least maybe like a week of plus of work with them. And maybe we should give that a go. The last thing that I forgot to mention before is developing more managers from within. It is very hard to know what someone is going to be like as a manager at your company. And if someone's been an IC on your team for a year, two years, three years, four years,

and you've seen them every day, you probably have a lot better sense for what they're going to be like as a people manager. Right now, for example, We have four engineering managers at our company, and all four of them started out as ICs on our team, which I believe is the way that RAMP has approached it. I believe they may have a rule around that, where it's like, Eng managers all need to start out as ICs first.

Then I think helps to reinforce some of these kind of values and cultural norms. John, this is an absolute masterclass. Thanks so much for jumping on. Thanks for having me.