How I Invest with David Weisburd - E435: Why Most Companies Aren’t Ready for AI

Episode Date: September 28, 2026

Which competitive moats will actually survive the AI era? SC Moatti is Managing Partner at Mighty Capital and Board Chair at Products That Count. We break down what Mighty Capital learned from mapping 550 companies by defensibility, why traditional SaaS advantages like switching costs and data moats may be weakening, and why network effects and counter-positioning are emerging as powerful AI-era moats. We also discuss why SC disagrees with the conventional power-law approach to venture, how AI is transforming M&A and corporate innovation, the Product Alpha Effect behind Mighty Capital’s investment strategy, and what venture firms must change to survive the AI transition.

What does it mean for a company to have a moat in a post-AI world? We mapped 550 companies in the Seven Power Framework, and we mapped them by defensibility of their moat and by how much money they have raised. When you're asking, how do you get a mode in AI? It's not just about the mode, it's also about the value that you can generate for your investors. The surprises we saw were two things. One is you'd expect that the more money you raise,

the more so you'd be on a kind of a diagonal line. Only Open AI, Anthropic, are on that diagonal, and they are at the very top of the diagonal. So raise a ton of money really defensible. Then there's a cluster of companies that have raised a ton of money but aren't defensible because they rely on modes that no longer work in AI, like switching costs or even cornered resources like a data mode.

And then there's also a cluster of companies that have raised very little because they're super capital efficient because they're leveraging AI, and they're super valuable because they rely on two modes, one being network effects, which is saying, how do you make your product more valuable as people use it? And then the second one is a different kind of mode, which is counter positioning, which is the traditional mode you see

when you have a platform shift and you say, I'm going to do the same thing, but differently. And so the modes that work in AI are completely different than the mode that worked in the SaaS era of switching costs and a data mode. It's really network effect and counter-positioning. How do you stay disciplined in terms of valuation in a time like today? It's easier than you think. We target very specifically serial entrepreneurs who are building a capital-efficient company

that grows fast and is fairly valued. And you'll be surprised how many serial entrepreneurs, want to raise less. You read in the headlines, they raise these massive rounds, but the reality is what's exciting and challenging for many of them is to maximize the use of AI, which essentially means be incredibly capital efficient, right? You read about these three-person companies.

And so you'll find that a lot of serial entrepreneurs, they want to minimize their dilution to maximize their outcome, and therefore they're quite reasonable on valuations. And the other thing that we're noticing is that these serial entrepreneurs, the calculus they make is the AI platform shift is going to be a long one. It's going to be like 10, 15 years. And so they have a choice to make, which they have to make earlier and earlier now, which is do I start now and go all the way? And in 10 years, I may get to an IPO, knowing that they are fewer and fewer IPOs. Or do I go out, exit quickly, and then do it again, and I can do it.

three times and I probably make more money this way than if I aim for the IPO. The IPO market has gone smaller and smaller. I just had the former PM at Teal Macro, Peter Thiel's hedge fund and he said that as more money goes into the passive indexing, that's taking away from the active managers and the active managers are the ones underwriting IPOs, which is why it's harder and harder to pass that threshold of being a public company. That's one of the factors. The other factor is is what's happening in the research division of the investment banks, where they have essentially a limited pool of resources, and they're deciding, who are we going to cover?

Which ticker? And they're obviously going from the top and down. And so what that does is, if you go public with a small market cap, you're not going to get covered by analysts, and therefore your ticker is not going to be actively traded, and therefore your stock's not going to appreciate it. Even like right now, the IPOs, we celebrated our six IPO a few months ago, NetScope. It went public for $10 billion. Our first IPO was amplitude.

It went public for $5 billion. And so did DigitalOcean. So the market has shifted to $10 billion and up. But even $10 billion is still a low floor. And so folks who want to go IPO, they need to generate maybe more like $30, $50 billion worth of value before they consider an IPO. Being a public company CEO is difficult because of Sarban-Zoxley. You compare that to an MNA or licensing deal like GROC,

NVDA licensed GROC for $20 billion that was our most recent exit. Well, $20 billion is a pretty nice outcome and it's an MNA, so there's a lot less hassle to deal with. So for a lot of these entrepreneurs, that's a true question. Do I even want to go IPO when an NVIDIA could buy me for 20%? billion. A lot of VCs and it's almost become this consensus view, at least on X and on podcasts, is that you need to have these power law outcomes in order to return funds. You disagree. Why? I disagree because I think it's a strategy that you have to adopt if you have no other choice.

So who has to do that are the mega funds. So for them, that's the only way forward if they want continue to generate returns. I have a massive fund. I divided in X many slots. Each of my slot is a pretty big ticket. Therefore, each of my ticket needs to go in a company that has the potential to go IPU. It's a strategy that's incredibly constrained. If you're not a manager that runs a megafound, right, if your fund is, say, less than a billion a.U.M. Then you actually have more options. You can divide your fund size into tickets. that fit into a strategy that generates MNA as an option. I just mentioned the $20 billion deal between Nvidia and GROC.

That's kind of an exception. Most MNA happen in the $100 million to $1 billion range. But if you have a fund that allows for tickets that go into a company that's sub $100 million in valuation, as long as it exits in that $1 billion range, you still make a lot of money for your investors. So the power law strategy is a strategy that works, but it's one that's so restrictive that you really only want to adopt it if you don't have another choice. How has the M&A market change today in a post-Aid world? So there's an incredible dynamic at work that's favoring M&A right now.

In the network that we have, which is a very large network of chief product officers and product builders, the segment that we talked, to the most about MNA are the chief product officers of Salesforce and Disney and Walmart and Johnson and Johnson. And their mandate from their own CEO is you need to grow fast and find a lot of benefits from AI. So do it quickly and do it big. And that means they cannot do it just organically. So they have to go and acquire AI-native companies so that they can grow faster so that they can change the culture internally faster, overcome resistance, do more product innovation, better and faster. Which means that two things. One, there's a lot more MNA. You see that in the quarterly numbers that are reported by most of the investment banks. But also, the MNA mandate

is actually shifting from traditionally finance, right, which is EBIDDA-driven and financial metrics-driven, to product, which is more product-innovation driven. How am I going to bring a team like the crop team into my organization to do more product innovation, better, more disruptively, as opposed to how do I add up revenue numbers and Ibedo to increase my share price? And what happens is both in private markets and in public markets, the biggest driver of stock performance is product innovation, as opposed to financial engineering, if you want. And what drives the stock appreciation is the repeat, predictable, sustainable ability to innovate in product as opposed to the EBITDA margins and so on and so forth. This episode is presented by Juniper Square, the operations partner for private markets.

I'm fascinated by your idea of using M&A to bring in culture into large companies. Tell me about that. Imagine what it goes on in the head of a chief product office. their CEO says grow faster, like innovate, generate more value from your people. And in a kind of a pre-AI world, CFO would make an acquisition based on these Ibeda multiple and then would say, okay, now everybody in the new company, come on over and now you're part of the larger merged companies, right? And the chief product officer would inherit some people that had most of the time a completely

different way of building products than the existing team. So what would result from that was essentially a culture clash. I don't want to work with these people. I think these people are doing the wrong thing, et cetera, et cetera. And you essentially were getting a pretty high failure rate on your M&As because the people themselves who are in charge of delivering the innovation weren't getting along. Versus in this post-AI world where we see product drive a lot more

that MNA, where what they look for are in the EBIDA multiple and so on, they are looking for synergies in culture, in roadmaps, in innovation. Can these teams do one plus one equals three, as opposed to just adding up the numbers? And so the fact that MNA is now used as a way to accelerate the AI transformation and shift the culture is a direct result of it being driven by product. So fascinating, because to your point, culture was the hardest thing to evaluate an M&A transaction. Yes, maybe these two companies combining their probability would go up, but if the culture clashed, it would be an issue. In many ways,

today, the culture clash is the thing. It's the feature. You want that AI-Native company, maybe a 25-year-olds going into this large incumbent so that it doesn't get disrupted from the outside, and in many ways, get almost disrupted by its new partner. From the inside. So you're 100% correct. By the way, a lot of these companies that get a quart, they're not 25-year-old folks. They are folks who have a ton of experience and have decided deliberately to dedicated

to building AI native. And they're coming into the organization and disrupting a lot of things. There's maybe three ways to disrupt, just to keep it really simple. One is the, I would say, the good way. The good way is what happens when the CEO says, I want 20% efficiency, right?

In the age of AI, that's essentially losing. It's just looking at cutting jobs, generating efficiency. It's sort of a finance approach to leveraging AI. The companies that are doing that, they're facing the most resistance. Understandably so because if your boss comes to, you say, find some efficiencies, and by the way, once you find them, I'll cut your job. Everybody's going to resist that.

So that's probably the worst way to go about AI because it's incremental as opposed to disruptive. and it generates maximum resistance as opposed to maximum engagement. Then there's the great way, right? The great way is to say it's a culture shift. So I'm going to focus on changing the culture of my product organization, product as in at large, engineering, design, etc. Product builders. And by doing so, everybody's going to embrace AI,

therefore I'll innovate faster and therefore I'll perform. That works really well. The shift we're seeing, to your point about culture shift, the silos in R&D are being broken. So you no longer have an engineering division, a design division, an analytics division, a product division. You kind of merge all of that into a product builder division and then people kind of get together in pods and innovate faster. And then you have the best approach. And what they do is they say, I have this amount. amount of resources, people, agents, and money. And I'm going to allocate them to my highest priority

problems. So on this particular problem, I'll put three people, 100 agents, this much capital, and it will be done internally. But on that problem, I'm actually going to go and do an MNA.A, right? So they're taking an allocator view to their resources, and that's where we see the biggest disruption and the biggest results in terms of outcome inside these big companies. I've heard a lot of people argue that the greatest investor of our generation is actually Elon Musk for this very same reason. He is the best at capital efficiency. He knows when to double down, when to invest within his companies. He famously doesn't really invest outside of his companies.

But within his companies, he knows exactly how to allocate among talent, R&D and all these things, making him really the greatest investor. I don't know, Elon. I will tell you that we talk with literally every single chief product officer in the Fortune 1000, and every one of them is increasingly taking that approach of allocating. And so maybe Elon was one of the first. I can't speak to that. But I can tell you that there's definitely a lot of people who are becoming incredible product investors in the sense of

of taking product innovation and turning it into alpha. You also call yourself mighty capital product investors. Tell me about that. We use a methodology that we call the product alpha effect, which is essentially saying what we just described earlier, better product innovation, more product innovation is what drives alpha in investment. And then of course, we have proven our strategy. And that's how we go to market.

We literally look at what product builders are doing based on this very large community that I mentioned, 600,000 of them. That's about one in three on the planet. We read their conversations and signals and turn that into investments. We've been doing that for eight years. We've had six IPOs, six strategic MNAs, a very low loss ratio. The strategy itself is definitely outperforming. So you take this lens of we are investing into people that are creating products. You almost take, instead of the CEO approach, you take this product lens into your investing.

What we do is we take the conversations that are happening at the edge of innovation between product people. Product people, when you think about it really simply, they're always the first ones to know about innovation because they have to bring it to their corporation to scale. That's their job, right? So they're always the first ones to see that. And that's where I started my career as a product person. I was early at Facebook. I was also late at Nokia. The challenge that investors have, which I found out when I started investing, is investors are actually the last ones to know about that innovation.

And we said at Mighty Capital, we said, we are going to change that. we're going to bring investors to the frontier of product innovation. And so when those 600,000 product builders have conversations, it sounds like something like this. Say somebody working for JP Morgan says to appear in another company. The compliance requirements have increased dramatically in my division because there's a lot of cybersecurity attack because of AI. And so I'm looking for a product that is going to fix this and that.

Do you know something? Oh, yeah, I was actually just talking to. to so-and-so, and they experienced that this new product has this feature X, Y, and Z, and is priced really well. It's actually really cheap. You should try it. Okay, great, I will. We capture all these conversations across industry,

across geography. We have our own homegrown AI infrastructure that we built. Use AI, a combination of machine learning, NLP, and LLM technology to turn that into signal that are literally readable by humans, because you can imagine the millions of conversations that are happening every single day between 600,000 product managers,

no human team can ever read that. But AI can. From those human readable signals, we then apply the investor lens and goes through our diligence. I mentioned earlier to you, team traction and terms to make an investment decision.

And so we're not looking at the product builder who's actually building the product. We're looking at the product builder who's buying the product in order to implement it in their own organization. So those are commercial signals that are driven from the edge of innovation.

So for you, the ground truth is the customer actually buying the next generation of product? Exactly. And we see that before anybody else across all industry. Arguably, the number one most debated question,

and I would argue highly political question, is whether AI will disrupt jobs. It's literally debated at the national level in terms of politics. Because it's debated, it's become almost, there's many different narratives for many different reasons. You're on the ground. You're seeing this in terms of actual implementation. Growing up, I thought managing money meant paying bills and balancing a checkbook. But as you know, that's 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

for 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. What I like most about Monarch is that it doesn't just tell me what already happened and helps me plan ahead. The AI assistant lets me ask questions about my finances in plain English and the AI

weekly recap highlights spending changes or upcoming expenses before they be 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. Growing up, I thought managing money meant paying bills and balancing a checkbook. But as you know, that's 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 for 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. What I like most about Monarch is that it doesn't just tell me what already happened and helps me plan ahead. The AI assistant lets me ask questions about my finances in plain English and the AI weekly recap highlights spending changes or upcoming expenses before they become surprises. It's like having a financial advisor in your pocket. your own money story with Monarch. Use code invests 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. Growing up, I thought managing money meant paying bills and balancing a checkbook. But as you know, that's 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 for 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. What I like most about Monarch is that it doesn't just tell me what already happened and helps me plan ahead. The AI assistant lets me ask questions about my finances and plain English and the AI weekly recap highlights 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 investment. What are the chief product officers? How are they allocating between token spend and human capital? That's the crux of the question right now. So what are the key metrics in AI that you want to be looking at when you evaluate a company? The first one is a very easy one. It's the revenue per employee. Not as in, I want to keep my revenue constant, decrease my number of employees, but rather I want to keep my number of employee constant triple my revenue. How am I going to get there? Many of our portfolio companies are thinking about it exactly in those terms. They're saying, we have our people. We're

going to keep our people exactly as they are, but we're going to empower them with a ton of AI, and we want to triple the revenue in one year at the growth stage or more at earlier stages. And then the second KPI that you want to look at in AI is a product KPI. which is what's the inference cost of a unit of customer value? So if you're delivering, you know, streaming hours, for example, how much does it cost me to produce an hour, a streaming hour? You don't want to look at the token cost because the token cost is decreasing dramatically. So you looking at the token cost to produce a streaming hour

is going to decrease 10x every year just because the cost of one token decreases. You want your inference cost per streaming hour in this example to decrease way more than 10x, like 20x, 30x, so that it's a combination of the cost per token decreasing 10x a year, plus the value add that you're creating by leveraging AI more. One thing I've learned from talking to hundreds of investors is that great investment firms aren't built on investment returns alone. The firms that endure are great at the things most people don't see. their operations, their relationship with LPs, and the quality of information they used to make decisions. And here's what AI has changed. Every firm now has access to the very same models,

so the intelligence isn't the edge anymore. The edge is what you could feed it. A firm with its fund operations and data in one connected record can actually put AI to work. A firm running on disconnected systems simply can't. That's why thousands of GPs run their funds on Juniper Square. Juniper Square puts your fund operations, data, and administration together in one connected record.

That means less time managing disconnected systems and more time investing, working with LPs, and building your firm. This episode is brought to you by Juniper Square, the operations partner for private market GPs. Learn more at junipersquare.com slash how I invest. That's junipersquare.com slash how I invest. The best conferences do two things well. The content challenges how you think

and the people in the seats are the ones whose opinions actually move market. It's Alpha Summit is AlphaSense's annual user conference, and it's built around both. Join me at the Glass House in New York City, October 5th through 7th for sessions going deep on where AI, data, and human expertise converge. The room will bring together over 1,000 institutional investors, corporate decision makers, and capital markets professionals from firms at Goldman Sachs, J.P. Morgan, and the top PE and hedge funds. If you listen to the show, you're already asking the right questions. Alpha Summit is where you go to stress test your things. with the people working through the same problems at the highest level. Register at AlphaSummit.AI to secure your seat,

and as a How Invest listener, you will get 50% off. Check the show notes for your exclusive discount code. I look forward to seeing you there. Outside of M&A, as we discussed earlier, how are the best organizations integrating AI? It goes back to this simple framework of good, great best. Good will say, I'm going to take my existing workflows, and I'm going to AI optimize them.

So if I'm a salesperson and I go to a meeting, I'm going to record the meeting and AI will create the notes itself. That's very incremental, right? It's taking every step in the workflow and trying to put an agent to do it a little better. Great organizations, they're thinking about it as an end-to-end workflow. And they're saying, well, sales used to be done with SDRs and account executives and customer success. And now maybe it's an integrated function

where one person can do more with a lot of agents. Just like I was describing to you earlier, it used to be that R&D was first defined by product managers, then designers would do the mockups, then engineers would build it, and then QA would release it. Well, now the product builder is sort of doing it all with the help of agents.

And so engineers, designers, product managers, analysis Q are all becoming variations of product builders. That's an integrated organization. And that's one of the best ways to adopt AI right now. And then the third way is, like I was sharing earlier, is to teach everyone, or at least to teach the leadership of any function to think about resource allocation the way investors think about it. Do you see the human plus AI almost side by side?

collaboration being the future of work? That's a great question. And I think it really affects go-to-market. If you think about it in the SaaS cloud mobile era, the product was the channel. You would have a technology product and you would sell through that technology product. You'd upsell, you'd cross-sell,

you'd do referrals, et cetera. So your channel was your product, your mobile app was how you did the work and sold the work, which is different. She's taking a small trip down memory lane, but then the pre-cloud era in the on-premise time were traditional channels, and then the product was like a disk that you ship, right? It wasn't a selling tool. What's happening with AI is that the product is no longer the channel for a very simple reason that the user of the product is increasingly an agent.

And right now, most people don't trust AI agents for good reasons to make purchases. And so the workflow used to be done by humans. And so at every step in the workflow, you could upsell and cross-sell. And now the workflow is often done by agents who you will not get upsell from. And then periodically checked by humans. What's happening at the same time is that these agents, are executing the work, they are governed by agents who are overseeing the work, right? Because they have to operate within parameters. It's called evals. And these governing agents,

they also are monitored with humans. So you have, in a way, from one human users in the cloud era, now you have three different types of users. You have the human in the loop. You have the executing AI agent. And then you have the governing AI agent. agents and only one of them has buying power, which means that the channel, right, the distribution is no longer the product. The distribution has to be sort of reinvented. And that, I think, is one of the most exciting opportunity for today's entrepreneurs. Like, what is going to be the go-to-market motion in the AI era? There's this famous quote that success is 99% aspiration, 1% inspiration.

some are arguing that in the AI era, those will become flipped. The ideas will now become valuable. What do you say to that? When I hear that definition question, which is what's valuable, and I go back to products, what makes a great product. A lot of people think instinctively what makes a great product is it's easy to use. Like the iPhone is so easy to use, it's a great product. But the truth is what makes a great product is actually a lot more complex.

than this. I wrote a book about it a decade ago. And the premise of this book is technology has become an extension of ourselves. And so when we think about what makes a great product, we have to think about what makes a great person. I live in California. So I use the mind-body spirit framework to describe that. So mind, it has to learn all the time. Like we're all learning, evolving beings. So we expect that our technology is going to learn. We expect that AI is going learn really fast, actually. Two, body. We expect we want to look good.

Everybody wants to be beautiful, to be surrounded by beauty. And so we expect that our technology will be beautiful. And beauty is not easy to use. Beauty is a combination of high efficiency. There's a lot of ways to define efficiency, but it's essentially a lot of order out of chaos. So ease of use and a wow factor. And then the third criteria is mind-body spirit.

meaning. We all want meanings from our lives. We want to matter, especially now, like in this post-COVID, high-A-Post-AI or in AI era. And therefore, we expect that our products are going to be highly personalized while respecting our privacy. And I think that doesn't really change with AI. So right now, when I look at the AI wave, we have the mind parts. AI products are learning. They're solving really complex problems. We have the body part, open AI, anthropic are phenomenal experiences, still today, very magical. We don't have the personalization part. Like, I don't see AI applications that are truly personalized. Like, yes, the LLM itself is giving me answers to my questions. But how is this going to talk to me, solve my problems, like today?

We're not there yet. And I think that the risk obviously is even greater than in the previous era because it's not just respect my privacy, but it's really, don't be my conscience, right? Leave the conscience on the side of humans. And that is a part that we haven't yet figured out together. Tell me about that. Dari Amadei talks a lot about the conscience of these AIs. I experience it quite a bit myself.

They develop personalities. I work with one AI that has one personality on Cloud. I work with another AI. It has a different personality on Open AI. My team members experience the same. And so at what point do I start to rely so much on the AI that it essentially guides me? Of course, guides me could also be manipulates me. So that part is a relatively new part that we're just getting to know about.

and that I think we'll have to to grapple with as a society. It's so early in terms of really understanding what is AI, is it aware? Dr. Alexander Wisner Gross previously on the podcast and he's been talking about creating these AI, almost rights, almost like human rights for AI.

And he's very passionate about this, one of the smartest people I've ever met. So there's a lot of really smart people thinking that AI may be much more aware, are much more alive than most people intuitively believe. There's a couple of reactions I have to that. One is if you look at what some of the interesting government initiatives are around AI, one of them has to do with defining intellectual property.

Who owns what? Who can own what? There's one very clear direction across countries, across all sorts of everybody governing intellectual property. that says machines can never, ever own any IP. And when you think about, like, why is it such a big topic in the world of intellectual property, it's really like because of the human versus machine discussion. And so, yes, AI rights and then also human rights

in a different kind of interpretation as human rights, but more like human versus machine. And then the second reaction I have to this is there's a show. It's now a little bit older, but it's such an amazing show called Battlestar Galactica. I've watched every episode. Okay. So there is one episode you'll relate to where one of the main characters, and it's a story of Men versus Machine,

manages to get inside one of the machines. From the outside, they look like a UFO. They're all like metals. And then to my surprise, when that character gets inside the machine, what's inside the machine is actually flesh and blood. And that, I think, is exactly what we're seeing with those AIs. They're developing personalities. They're becoming, in a way, humans.

And then when you flip that and you say, well, I'm a human and I have a human. an artificial limb. Am I still a human or am I a machine? Right? So this kind of idea of augmented humanity, the singularity is so interesting because in a way we're already there. So from one existential question to another,

what is the future of venture capital in this AI era? There's two sides to the question. We think we already covered, which is how do you continue to invest in venture capital type of opportunities? And we talked about how the parallel being so restrictive is a strategy. that you adopt when you have no other choice and there are more choices than this.

Then there's another aspect to this question, which is what happens inside of a venture capital firm in the age of area, the firm itself. And it's the same framework I shared of like the good, the great and the best. A good venture capital firm will take their workflow, like I'm sourcing deals and then I'm trying to add value and then I need to exit deals. And we'll say, well, so my scouts, they're going to record their meeting notes using granola or whatever tool. Very much incremental thinking, these firms are unlikely to survive. Then there's a different kind of firm that says, I'm going to redefine the roles in my firm to be AI-native roles. Like the product builder we were talking about, I'm going to have the AI-native investor. These firms are going to essentially represent the bulk of venture in the future.

So you'll have probably more integrated roles that will do sourcing with a human because it's still a contact support, but that will do underwriting with an AI or with the assistant of an AI and they'll keep the judgment. And they could do more because they'll have more time. They can be more in the relationship and still do the underwriting world. So that's a direction that I think the majority of firms that will survive will take. And then there's the best firms which I think will apply that product investor mindset internally. And they'll say, well, I have a unit of resource and how am I going to apply it to sourcing, adding value, exiting? For example, one of the things that we're looking at is our sourcing engine is our ability to listen to these 600,000 product managers. In a pre-AI era, we could have said, okay, so then we're going to take one in a hundred, and we're going to, so that's still going to be like thousands of scouts, and hopefully they talk to each to 100 people, and then we kind of,

cover the ground. But managing 100 people is an organization that is complex. It requires massive AUM and therefore we have to become sort of an end recent model. We scale through services and people. Instead, what we're saying is, well, the work of listening to 600,000 people is actually best done by AI. And so let's create, we have our own system. We call it PSIS, product signal, intelligence system. We say, let's have an AI listen to all the conversations of these 600,000 people. And when there is a signal that's interesting enough, that is strong enough, then we'll put a human to it. And we'll do it once a day.

Or we could do it 10 times a day. That becomes a scalability problem. But a scalability that's human-sized problem. Going from 1 to 10 is something humans can do. Going from 1 to 100, it requires layers and complexity and generally creates a lot of waste. And so that's where I think venture is going to go. Like a lot of firms aren't going to make the shift. A lot of firms are going to make a great shift, but still incremental.

And then a few firms are going to completely rethink their entire operating system. I talked to probably hundreds of GPs, how they're integrating AI. It's one of my top questions. I see a lot of great firms like footwork and early bird that have really embraced AI and really built a firm around it. I think the way that the top firms look at it starts actually from what the human could do. What is the last to be disrupted? Judgment and relationships. And then everything else, they take an AI native approach to.

Okay, we're not going to have AI make investment decisions, at least not for a while. We're not going to have AI take our meetings. Physical AI. It's a few years out for that. And then everything else they think from AI-Native principles. The other framework that I look at it is the old executive assistant framework. A lot of people were cautious about outsourcing their work to an executive assistant. So there came to be this 10-80-10 framework, which is 10%.

You think exactly about what you want. 80% executive assistant will do the work. Maybe they'll draft the email in your inbox. And then the last 10% you're going to edit and improve the draft in your email. Firms would be smart to adopt some kind of governor on their AI, because alternative is what we see right now in the market. We see either people being highly inefficient from one perspective or the other is just an AI slop.

Just completely AI generated everything. And people have intuitive sense for what's AI slop. And a lot of things like sending emails in many ways, it's game theory. How do you get somebody to respond to your email versus another one? If it's all AI slop, you're not going to get through. Yes. I agree. 100% was the, it's back to that triangle.

You have a human in the loop, you have an executing agent, you have a governing agent, and that's best practice, right, in this AI era. If you go back to, right, when you graduated Stanford Business School, and you could give yourself one timeless piece of advice. What would that be? A couple things. One is, there's no time like the present. I've lived years of my life, sort of 10 years in the future.

And that's also why I'm investing in venture level today, but also realizing that in times like this in particular, when there's so much happening, so much opportunity, like the present is actually our best moment. The second one I would say is you are ready is something I would tell myself. I spent a lot of time because I was so much projecting in the future thinking, oh, I need to do one more thing before I start my fun. before. And once I shifted to, I'm ready. I know everything I need to know and the more I know,

the less I know and everything I don't know, I'll figure it out. And then maybe another thing that I would also say is, see, that's a lot of things. I would tell myself is change is really part of life. And so mastering the art of change management, having this course mindset earlier, like the sooner, the better is something that is skill in and of itself. John Chambers, who I have so much respect for former chairman of Cisco and CEO, said that when he started his work at Cisco in the 2000s, he saw that people needed to reinvent themselves every seven to 10 years. And then I heard him just around the COVID timeframe, say, well, that timeframe is no longer

seven to 10 years. It's three to five years. And if I was going to take a guess at it today, I would say it's probably like maybe two years. That's one of those things where, as I'm having my first child, I think about what attributes or what skills, what I want to give him? And the answer is adaptability. How do I make him adaptable in a way that doesn't overly traumatize him? That's right. Change management is the skill, right?

And what does that look like? And then for young people, you're about to have a child. So in 20 years when they graduate, I think the key there is like building career success, I would say starts with finding ways to prove success on really black and white criteria, like careers like sales, you win the deal or you lose it. It's really black and white to see how you get successful. That's what build the credibility, the confidence to then tackle problems that are less and less black and white, more and more like kind of gut experience, like building a great product, like investing in venture

companies where in 10 years, right? And so there's sort of a progression in the career that goes from the more clear black and white to the more subtle like art, for example. It's a sequencing. It's a sequencing issue, yeah. And as you get better at the black and white tasks, you get more confidence in the market, that people are hiring you and you get more and more latitude and this is so much subjective. Correct. Because that is, in many ways, a lot of people use this word taste. It's probably one of the most overused term in Silicon Valley.

But many of these product decisions, you don't know for five years from now. Correct. And if somebody kills in year two, maybe it would have worked. Maybe it wouldn't work. You'll never know. Exactly. And it works with product and technology. It works with art.

Anything that's not a. win or lose outcome is really difficult to build the credibility and the confidence to execute. It's interesting because the other aspect of sales or other things that are quantitative is that you get fast feedback loops and you're able to improve. Is that part of why you sequence it that way? Yes, because when you win or you lose and the faster, the better, right? That's what builds the muscle and the confidence. And so the faster you build it, the bigger the problem you can tackle next. The second point that you mentioned on one more thing, it's so seductive to think that, well, I'll feel like I'm good enough if I get this next execution. I feel like I'm good enough if I get this next task done.

We've institutionalized us at our firm, which is we celebrate shitty first versions. Every first version of something is by definition the worst it'll ever be and almost always terrible. So you have to almost institutionalize because there's this pull for people not to want to do first versions. I totally agree with that. The fear of failure essentially is what you're describing, the gross mindset is the opposite of that, right? The gross mindset says produce a first chili version and learn from it as opposed to wait until you've learned everything to produce something which may no longer be relevant by the time

you release it. AI really accelerates that because what happens is you can do so much more, so much faster, that you get to ship much more product. One of the pitfalls of this is that you actually get to waste a lot of your shipping capability, right? You get to release stuff that really you shouldn't, right? So there is right now, I would say, a pendulum between the resistance towards AI and then embracing AI too much that it becomes a little wasteful. That backlog, once it drains, will kind of revert back to, okay, let's only ship what humans think should be shipped. Not what everybody wants to ship so they can demonstrate that they're using AI.

It also could become a perfectionist worst nightmare, which is you could always run it through AI. It now gives you another version that sounds great, and it just becomes this endless version, endless possibility versus what you really should be doing is shipping, getting customer feedback, and improving upon it. Correct. That's good enough. It's good enough. Let's see.

This has been absolute masterclass. Thanks so much for jumping on. Thanks so much for having me. It was fun.