SaaS末日：谁能活下来、谁会出局？| Insight Partners联合创始人 Jerry Murdock

Most of the companies, their view, as they've told me, is that Cursor is obsolete. That's where the product is today. What we have with the tsunami happening is a wake-up call to move to higher ground. Don't get caught on the beach when the damn thing hits the beach.

Jerry Murdock is one of the most influential venture capitalists of the last 3 decades. He's a co-founder of Insight, which now manages more than $90 billion. He led one of the most eminent rounds for Twitter, and today we sit down to discuss his biggest lessons from 30 years of investing.

If I have any wisdom at all, it's because I fucked up so much and I've learned from it. You really don't know the edge unless you go over it. Here's the secret. The real secret. I never left the game. I was lucky because I was losing money every year I've been in this business. I failed. Money does not come with instructions.

Jerry, I've known you for years. I've wanted to make this happen literally since we first had that lunch, so thank you so much for joining me today.

1. Autonomous Agents: The Tsunami Analogy for AI

Hey, happy to be here.

2. Is Now the Best Time Ever to Start a New Fund?

I think it's such an interesting time because, bluntly, me and a generation of investors are going, "Hang on a minute. Is everything that we've been taught for the last 5 to 10 years completely irrelevant?" Now I get to sup on your wisdom for the next hour. When we chatted on the call Friday, you talked about tsunamis as an analogy for where we are with the AI wave that we currently face, right? What did you mean by the tsunami?

3. How to Invest When Anthropic Can Kill Your Startup

Well, the first thing I think about a tsunami is that it's harmless when it's out at sea. It's only dangerous when it hits the beach. That's number one. Number two, it's messy. You're going to get maybe an earthquake or two notifying you that there's something possibly coming.

It's not just one wave. There's these pre-peak waves and post-peak waves, but there's an event coming that is more than a single product. In this case, it's autonomous agents. Autonomous agents, in my opinion, are what the tsunami is about—not just AI in general. Autonomous agents are the big wave that's coming. So where are we now? We're in the anticipatory period, where we can see it coming, and that's why we're seeing the SaaSpocalypse.

Is that the case?

I'm not a doomsayer, but I am one to say, look, change is coming fast and you need to anticipate that. You need to be on top of that. The idea of, "Oh, I can just bolt on AI to my company" is possibly true. You can maybe get an exit, but being AI-native and thinking that way is going to make you a better company.

You need to think deeply about what is going on in the communities that are driving the tsunami, right? Those open-source communities that are popping up in massive amounts—they're the ones that are going to have the big impact.

We're going to get to the bolt-on AI strategies, but you have an incredible portfolio of companies and you work with many incredible founders. What do you see in the portfolio—as we touched on when we chatted on the call Friday—that maybe people aren't seeing today when they read the news or think about investing?

If I look at real, true AI startups like E2B, Eventual, Lotus AI, GetDynasty, and Aven, these are all native AI companies that are up and running. What's happening with them is they're all using OpenClaw, NanoClaw, or one of their own homemade autonomous agents.

I think that's not obvious yet in the marketplace, but it's because it's only about 2 months old. These guys have been doing it for anywhere from 2 weeks to 6 weeks, roughly, and have brought in autonomous agents to actually write code. That's the thing that I think is mind-blowing.

They have autonomous agents to actually write code. What does that mean for Cursor, a $27 billion to $30 billion company that's raised a lot of money?

4. The Future AI Orchestration Layer

I'll tell you. Most of the companies I just mentioned, their view, as they've told me, is that Cursor is obsolete. That's where the product is today. My view is, yes, but that team is really smart. They've got a lot of money and a lot of customers. They've got time to embrace autonomous agents, which is what I think they'll do, and they've got a shot to pivot and figure out what the next direction is.

In the AI business, you've got to be going where things are going to be. You can't be thinking about yesterday. So I think those guys are going to have to quickly embrace autonomous agents.

You mentioned the utilization of OpenClaw. What do you think is the impact of OpenClaw more broadly that we're maybe not considering enough?

Great question. Look, the first thing you have to do is look at OpenClaw and look at the community. I mean, the commitment to open source and the number of people developing for it, right? You can look at huge companies like OpenAI and Anthropic, with massive resources and super-talented people, and then you look at open source, where you've got sheer numbers of people doing integrations—massive amounts.

If that community keeps accelerating and growing, we're going to see agents do incredible things that they don't have today. When I think of an autonomous agent, first of all, we're going to have to come up with what we call the Claw stack, or some form of stack for the autonomous agent.

Back in 2003–2004—you may not remember—we were post-9/11, and the world was miserable. People couldn't afford to build websites with Sun servers and Oracle databases, which were expensive. But the LAMP stack came out, which was Linux, Apache web server, MySQL database, and PHP front-end development tool. That led to the explosion in 2004–2005 of websites and, ultimately, commerce. Google went public in 2004 and rode that wave brilliantly.

I think the same thing with autonomous agents is that you're going to find the open-source community coming out with the stack, right? Right now, you've got a reasoning layer that's dominated by Claude, Codex, and Gemini. I think what's going to happen with autonomous agents is, ultimately, they're going to have an orchestration layer where they can have multiple different LLMs that they can orchestrate and use to triage workflows.

They'll triage workflows and say, "Hey, for this part of the workflow, let's use Claude. They're expensive tokens, but they're going to get the job done better. For this other part of the work, let's use an open-source model like DeepSeek, Llama 3, whatever." They're going to start from a big picture. Maybe this is going to happen as soon as orchestration gets solid enough. They're going to determine where the workloads go for reasoning models, which is going to be super powerful.

5. The Rise of ASIC Chips & Nvidia's Threat

My guess is this will lead to the rise of open-source models more proficiently. That's going to lead to the rise of ASIC chips, because what's going to happen with ASICs is you're going to put the model on the chip. ASICs are going to be a lot cheaper and a lot more tunable for a specific workload than an expensive chip from Jensen. I think we're going to see massive development of autonomous agents influencing open-source models and ASIC chips. From the big picture, that's a revolution.

One of my dear friends is Rory O'Driscoll from Scale, and he always says, "The thing I love about Harry is the statement, 'That's great, Jerry, but what about me?'" [laughter] Which I specialize in as a venture capitalist. I have a shitload of NVIDIA, and I'm thinking through this as I listen to you. Can you just play out for me that migration from Jensen chips to ASIC chips and how that actually plays out in reality?

Right. Well, it's the thing that wasn't mentioned as to why he bought Groq. He needs that capability so that he can handle ASIC chips eventually, right? Groq—those guys know how to put memory right on the chip, and so that's automatically an ASIC chip today, right?

I think the Groq acquisition is not just about handling different types of workloads and getting memory on the chip. It's about making sure that CUDA can also support ASIC chips. They're going to have to go out and make sure they know what's coming. They absolutely know what's coming, and so I think the Groq acquisition is going to help them make sure that CUDA is viable for the ASIC explosion that's coming.

If the ASIC explosion comes and CUDA does migrate with it, does NVIDIA retain its value, or does it still denigrate because they have a leakage of value with the ASIC chip expansion?

It depends on execution. This is the thing about the game: who's going to out-execute, and who's going to get more of it? A lot of people like to criticize Meta here and there for being behind the game, but they had the balls to say no to Jensen. "Sorry, Jensen. We don't need you."

So why did he do that? Because he's betting on ASIC chips. No question about it.

You also mentioned the routing and triaging between different models. It makes me think of a couple of different things. One, is that not just a commoditization of models, where it's a race to the bottom on price and who can deliver the cheapest and fastest? And two, aligned to that, what if we then just see models do what we're seeing, which is eat into the application stack? How do you think about those 2 elements?

I think the answer to that question is going to be decided by the autonomous agent, not developers. The agent is different from a developer. If you and I are developers, we're going to go and do something based on what our experience tells us, what we think.

Okay, let’s go build something here, and we’re going to use an ASIC chip or whatever. What’s the difference? An autonomous agent is probabilistic. The probabilistic nature of an agent is going to say, “I don’t know which is better, ASIC or one of Jensen’s chips. Why don’t I just go out and get 10 Python libraries, run them in 10 different sandboxes, write the workload, and then see which one performs better?” That’s what’s going to happen.

Look for autonomous agents to have more of a say in that.

But when I think about it again, that’s great, but what about me, Jerry? I’m investing in the agent layer today, like many other investors are. I’m also seeing Anthropic release incredible application-layer products very quickly, whether it’s legal or their Cowork.

Yeah.

How do we determine what’s safe for us to invest in versus what’s in the path of an Anthropic product update?

You know, that’s why you get paid the big bucks, Harry: no one’s going to tell you what’s safe.

I thought that was the point of this.

Listen, I’ll tell you: about 80% of the investments I’ve made have returned less than 1.3x. It’s the 20% that made all the money in my life and made all the impact. Money is just a sort of batting score for how much impact you had as an investor. That’s the way I see it.

There is nothing safe right now. You’re going to have to watch who’s going to execute and who’s going to execute the best.

Do you think the public markets overreact when it comes to price responsiveness on these updates? You saw Cloudflare and CrowdStrike hit by about 10% on the back of an Anthropic security update. Is that an overreaction?

Look, to answer your question about whether the markets overreact, if you go to Wall Street, you’ll see the giant bull statue that’s there. It’s because of a herd mentality. Investors are like, “Hey, I’m not a technologist. I’m just going to sit back and watch this.” Someone panicked. Maybe the others are just sitting on the sidelines.

What you’re seeing is less about panic selling—because it’d be down a hell of a lot more than it is—and more about cautious, on-the-sidelines buyers who are saying, “Hey, I don’t see a deal here. I’m not sure I have enough information to jump back into CrowdStrike and bring the stock price up.”

6. Parallel to the Dot-Com Crash of 2000

Is it an overreaction, or is it just a pause where a lot of investors are sitting back and saying, “Hey, let’s get more information about who’s going to be the winners and who’s going to be the losers?” That’s the way I see it.

Is there a parallel to a prior time where you can remember this feeling of, “Whoa, whoa, whoa—immense uncertainty. I do not know what happens. I’m better off sitting out and watching because I don’t want to see Monday drop another 40%”?

Yeah. March 2000, almost 26 years ago: tech stocks dropped between 30% and 40% across the board. If you missed a quarter, you were down over 50% or 60%. We lived in that malaise from March through the summer.

By the way, we were still able to get a couple of IPOs out at lower multiples during that time. We just sat in misery until 9/11, which I call the coup de grâce to finish us off. The market got destroyed, and we were thinking, “Hey, we don’t do dot-com investments at Insight. We’re going to be fine,” because we knew then, in ’99, that there was a bubble and that it was going to blow up.

We said, “Look, it just can’t be sustainable. There’s not enough commerce. Not enough people are on dial-up. You can’t do commerce on dial-up, and there’s not enough fiber in the ground.” Sure enough, there was a burst. However, the tsunami came in, took out the dot-coms, and then took out all software. We all went down, and we were in misery for years.

Do you think we’re in that same pattern? Do you think there will be a depressed public-markets atmosphere for years, or will it be different now?

Well, it’s always going to be a little bit different. What’s happening now is the speed at which things are changing. We have this company, E2B, and you think, “Okay, they do sandboxes. Great.” If you talk to most technology people, they don’t particularly put much emphasis on differences between sandboxes. It’s just a thing to keep your agent safe, or to keep whatever code you have safe.

In reality, it’s also a productivity tool. For agents, if you’re making sandboxes for an agent, it’s going to have a qualitatively different speed. About 400 milliseconds is a time period at which humans can recognize a delay. A lot of mobile systems operate on that 400-millisecond time frame, and that’s what most sandboxes are.

E2B has sandboxes that respond in 80 milliseconds. I’m like, “Wow, no one would ever notice it.” He said, “The agents notice it, and that’s what matters.” When an agent spins up literally 100,000 sandboxes in seconds, that response time is critical.

We’ve spoken a lot about autonomous agents. When you have autonomous agents, do they make your systems of record valueless, or do they make them much more valuable because they have existing distribution, which they can incorporate those agents into and leverage? Which one?

Well, it depends. I’m an investor in Carta and have been for a long, long time. That’s a system of record of sorts for stocks, and it has a lot of potential.

If tokenization of stocks comes and they use Carta, then Carta is going to be infinitely more valuable, right? They have the cap table, so they can help manage the tokenization. However, if tokenization comes and bypasses Carta and a new system of record gets built, then there’s not much to say that that system of record is going to be worth much in the future.

It depends on the execution of the Carta management team and their ability to capture the trends that are coming forward.

If you were to look at Salesforce today, which camp would you put it in?

There are a lot of companies that sit on top of it, like nCino. There are dozens and dozens of companies that have built on top of the Salesforce system of record.

The real question is, let’s look at their health. Let’s see how they do. If those guys start getting knocked off one by one, and enough of them get knocked off that the underlying value of Salesforce becomes less, then you’re going to say Salesforce is going to be worth less.

But in the history of companies, Salesforce is like Mount Everest. It’s an 8,000-meter peak sitting there. It’s not going to melt overnight. That thing is going to be around for a long time.

The question is, how valuable is it? The way to analyze that question is: how valuable are all those companies that build themselves on top of it? If those guys start going down, then you’re going to see some issues.

7. Billion-Dollar One-Person Companies

I’m a little bit stuck here because our job is to ascribe where we see value and where we think there will be more value. The things where we used to ascribe value—revenue, growth rate, margin—have become transient or questionable. In that time, how do you think about where you ascribe value?

It depends on what the time frame is. If you look at most software companies that are out there, normally when a new technology comes, it sort of expands the market before it contracts, because people are already in their momentum. They’re doing what they’re doing.

I wouldn’t expect normal software companies to see software just fall off the radar. That’s not going to happen. What happens is that people get fearful and prices change, and then the underlying quality of the business changes based on how well the management team adapts to the situation.

You can imagine some companies that have systems of record with great context actually increasing in value if they’re able to put that context to work with agents, with autonomous agents. Then you have other people who don’t adjust fast enough and don’t really understand what the new market is like. Their system of record, or their software, declines because they haven’t really moved fast enough.

What we have with the tsunami happening is a wake-up call to move to higher ground. Don’t get caught on the beach when the damn thing hits the beach. Move to higher ground.

That kind of adaptation of your business—some people will do it, and some people won’t.

8. Agents as Employees: The New Reality

You said to move fast. You’ve been doing this a long time. I’ve shockingly been doing this for 10 years now, which shows on my wrinkled face. My question to you is: we’re seeing growth rates like we’ve never seen before, ever. I mean, zero to 100 million in revenue, actually quite frequently or semi-frequently in some of these cases.

Is triple-triple-double-double dead? Have growth expectations been completely blown out of the water?

Think about this: right now, all software is eventually purchased by human beings. Software is going to be purchased or used by agents.

What’s going to happen is an autonomous agent becomes an employee. You give it credentials. You give it identity, and then it’s up to the agent to make the decision. You will review them like an employee and say, “Hey, what did you buy? What did you spend? What did you accomplish?”

You will manage that autonomous agent just like an employee. You need to prepare for that. Whatever you’ve got invested in, is that future going to be done and controlled by agents, and am I in the right position for that or not?

How does the world change when you’re not selling to people but you’re selling to agents? How does that interaction change the interface? What changes when there’s that core change in the buyer?

Well, first of all, it’s never happened before, right? So it’s happening now.

My guess is that the way pricing models work will improve, and it’s been in process for a few years. One of my companies, Docker, has been moving dramatically toward this as they move into AI, which is a consumption-based model. If the agent has access and authorization to use something, say a sandbox, you’re just going to pay based on consumption.

The sandbox is nothing more than buying memory or buying compute. You just open it up—it’s free—start using it, and then ultimately you have to pay. Then the agent will come and tell you, “Hey, we’ve reached our limit with the sandbox company here. What do we need to do?” That’s how it’s going to work.

Going back to what we said earlier about bolt-on AI strategies, there are a lot of public SaaS companies—I don’t want to name any of them—and even some late-private companies that are trying to bolt on AI and trying to pivot. Does that work? How do you think about the bolt-on strategy as an effective strategy?

Well, look, I was just at the Olympics, and there are a lot of people trying to win a gold medal, but not many people do. You can try all you want, but you better be world-class at what you’re trying to do if you’re going to get the medal.

I have the CEO of monday.com on the show tomorrow. When you think about the question of whether enterprise software—and software in general—will just be entirely vibe-coded, personalized, and customized, how do you think about that? Because that is the kind of eradication of a lot of their market caps.

First of all, autonomous agents—and it may take a year or longer for most enterprises to actually enable and be committed to autonomous agents—but they’re going to write software faster and cheaper than any human being on the planet. They already are.

Again, if you’re making your software for autonomous agents, and they have a reason to use it and it’s valuable, great. You’re going to do fine. But if you’re not making your software for autonomous agents today, you’re going to be challenged in the future. Maybe it’s 6 months, maybe a year, maybe 18 months, but you’re going to be severely challenged if you still think human beings are going to buy your software.

9. AI & the Labor Market: Which Jobs Go First?

We have agents buying software, and we review their decisions. Great. I don’t have to deal with sick leave. I don’t have to deal with entitled millennials. I sound incredibly old-school, but this is the truth, as you know.

Are you worried about what happens to labor forces, given what you said—most importantly, that the speed of change is so fast?

I’m glad you brought that up. We’re about 2½ years from the next presidential election, and I’m here to say that this is going to be a major issue in the next presidential election. It could decide the election. It will be one of the most important things because there’s no easy answer.

There’s no question that anyone who inputs data into a computer or does scheduling—an executive assistant, for example—or people doing marketing, those jobs are probably ultimately going to be better done by autonomous agents. White-collar jobs—people have been saying that for years. This is nothing new. The question is, when was it going to happen?

Now, with the advent of OpenClaw, NanoClaw, and all the other autonomous agents that are coming, there are direct threats. First of all, not to the people with the job, but to the next person in line: the next junior developer you want to hire, the next executive assistant, the next marketing person.

The first thing you see is people stopping or slowing down hiring white-collar employees who input data and use a computer. That’s the first thing. Second, depending on the speed at which these autonomous agents evolve—because there’s a lot that they need to do—it’s been proven that they can write code and do anything you want from a scheduling perspective for your life. They already can replace a human being. They will.

I think the job market is very uneven. I think the first companies that are going to go with this are going to be small and medium-sized businesses, because they’re the ones for whom one secretary makes a huge difference in how they handle customer service. For a 2-, 3-, or 4-person company, that can be transformative.

I know one company, Retro App, that’s out there today. I think they’re using autonomous agents to help triage their support email. All of a sudden, their customer support can shift immediately. You don’t need humans doing it; you’ve got the autonomous agent doing it.

I think we have to look at the speed at which autonomous agents grow in 3 sectors: consumer, small business, and enterprise. My guess is that enterprise is last. They have been last in this whole AI revolution. They may catch up in a year or 2. We’ll see.

That’s what I think ChatGPT, Google, and Anthropic are absolutely going to be focused on: getting their own versions of autonomous agents adopted rapidly by enterprises. When this happens, we’re going to see dramatic change.

Politically, I think minimum viable income may become a reality, or at least a ballot question, in 2½ years.

10. UBI as the Policy Response to AI

What does that mean, minimum viable income? UBI?

It means that you’re going to get, effectively, a certain amount of money guaranteed to you every month. If you’re a white-collar person, they’ll change the unemployment program, because no administration is going to want to sit there and say, “Yeah, I’ve presided over 10% or 15% of the country unemployed.”

What they’ll do is have an option to put people into a program and retrain them to do something different, maybe to have a better quality of life. If you take the optimistic view, the jobs that are going to be going away are not going to be jobs that get you a vacation in Brazil anytime soon. They’re the jobs where you’re suffering every day, making a living and struggling with whether you can pay for health care.

Maybe those people say, “I’m going to move out of the city, move out of the suburb, and maybe I’m going to move back to the country and grow food. I’m going to get this grant to do so.” There are already innovative businesses in Wyoming where they’re taking veterans and giving them branches to lease, because now, with technology, 1 or 2 people can run a massive branch instead of 8 people. New things like that are happening everywhere.

Does this labor displacement help or hurt a Trump administration?

It depends on when it happens.

In the next 2 years, I think we both agree that customer support—we’re already seeing bookkeepers, legal, and coding would be the 4 really vulnerable areas.

I think health care may hit that before the labor market hits it, because it’s going to happen. We’ll find out this fall what the impact of health care is on the election. If it’s a big impact, then you can assume that labor will be a major impact on any administration.

When we think about the impact of agents on companies, I had Seb from Klarna on the show, and he said that at their peak they were 7,000. By 2030, they’re going to be fewer than 2,000. Is headcount a bug, not a feature, now in companies?

Look, people never really talk about it, but it’s all about culture. What’s your culture? If you’re looking to have a culture of A players—and Steve Jobs talked about this at Apple—you want to continue to grow with A players.

You’re going to have a hell of a lot more agents than people. So it depends on what those 2,000 people who are left at Cluely are doing. What’s their culture like with those 2,000? If they’re all able to have a greater quality of life and have more time with their kids, maybe it’s an awesome win. But I’m not there. I’m not him, so I can’t comment.

Do you think we will have billion-dollar, single-person companies? It’s often said. Do you think it’s actually realistic?

Absolutely. It’s all about how smart your agent is, how smart you are at deploying the agents, and how willing you are to listen.

I think the thing that’s shocking is that autonomous agents work—that OpenClaw worked. That’s why so many developers are passionate about OpenClaw, because it’s something that’s working on its own without having to be reviewed like an employee. You’ve gone from an assistant to an actual employee. Think about that. That’s a major difference in life.

I think those kinds of changes are going to be really important to notice.

I have Orlando Bravo on the show tomorrow, and I’m sitting looking at my book. I’ve got Anaplan, and I’ve got Coupa. I’m not picking on him, but I’m picking on you, right? How does it impact me? I need these assets to go out. They’re 15 to 20 years old, typical enterprise SaaS companies. What happens to traditional private equity in this environment?

Let’s go back to 9/11. There were several of these traditional buyout firms. They were called hedge funds. Now they’re called something more sophisticated: private equity. But they were kind of buyout shops.

Ted Forstmann, who I knew, ran a great one. Except in 2000, he went all in on telecom—virtually telecom. Then it was over for him after 9/11. Done. No more Forstmann Little.

There will be stories like that. I don’t know about Bravo or anybody else, but there will definitely be private equity firms that end up like Forstmann Little.

And then there’ll be other ones that had the right bets, did the right creativity, and will be bigger, better, and more important than they were before. Life is basically a set of experiences you have and the total sum of the decisions you make. Any of these firms is about the decisions you make.

Now we’re at a point in time when you need to review what your assumptions are and what decisions you’re going to make going forward.

If you were to make a decision about where you would place a firm today, seeing the atmosphere and the landscape that you have, where would you choose your investing shop to be?

The problem is you’ve got an old man here who’s going to want to have it up in the mountains in Aspen or in the Alps. That’s my new thing. I’m not like Elon or Jensen, who are going to do this for the rest of their lives. I’m just doing this for impact. I can’t even imagine going out there and doing it for the same reasons I did it years ago.

It’s like Bob Dylan got interviewed, and they asked him, “What was it like for you back in 1964 at Newport?” He says, “I don’t even know who that guy was,” and I feel the same way. I don’t even know what that guy was like back in 1995. It’s moved on.

I was always into this business for the creative impact and to play a very small role on a team that was doing something meaningful. I want to do that. That’s going to be important. But it’s just a matter of having the right amount of wisdom about how the change is going to impact people and how you’re going to do it.

I mean, VC firms should have their own autonomous agents. Everybody should. Can you imagine the amount of analytical work? Certainly, the PE firms are all going to have autonomous agents analyzing markets and analyzing where the opportunity is.

If I was starting from scratch, the 1 thing I could imagine is having incredible data I could vet on a new opportunity in the white space where these new AI companies are going to go and say, “Wow, the data says this guy’s on the right market. He just needs to execute.”

Then you evaluate not just the person, but the quality of how they use autonomous agents. That’s going to be the deciding factor for venture capitalists and startups. We’re going to be on the same level playing field. How well do we use autonomous agents in our job?

Going back to Bob Dylan—I don’t remember that—and I think, actually, money makes people better investors. I’ve spoken to some of my most successful friends, and I say, “Has becoming rich made you a better investor?” I’m intrigued to hear your answer. Have you become better the more money you’ve made?

Well, if we use money as a proxy for success, it’s funny. With some of our mediocre companies, we thought about swapping out the CEO, and it’s like, well, the problem is it’s a mediocre company, and anybody we would put in would have to already be worth a couple hundred million bucks, so why would they do it?

You’ve been forced to limp along with the existing team because you’re just not going to recruit someone better. They wouldn’t be successful. There’s no validation unless they’ve made a lot of money.

Number 1, it’s a proxy for success, and success is a combination of experience and validation of when and how to make decisions. For me, there are 2 components to every decision. There’s the logic, and the second is the intuition, and you need the intuition.

The 1 thing that the autonomous agents aren’t going to have is intuition. Not anytime soon. We need people. We need good intuition to make good decisions. Without that component, you’re not going to succeed.

In my humble opinion, maybe I’ve got better intuition, and that’s the reason I’ve been more successful.

Jerry, when’s your intuition been most wrong? And what did you learn from it?

11. Intuition vs Wishful Thinking in Investing

The first thing is, how do you know it’s really intuition and not wishful thinking? That’s the first thing I’ve learned over time. Intuition was almost never wrong. What I was wrong about was thinking it was intuition. It was nothing but wishful thinking.

When did you wishful-think too much?

When I found an entrepreneur that I liked who was smart but maybe too comfortable. Maybe a person I liked more as a human being, and they weren’t crazy enough. They weren’t really driven enough. They didn’t have that chip on their shoulder. They didn’t have that obsession.

You want to work with people you like, and so most of the people that I liked were the ones that let me down and let themselves down. It’s the people that you struggle with, who are a little sharp-edged and a little aggressive, and who you may not like as people, that are like, “Wow, okay, but they might be the ones that succeed.”

You have to find that core of humanity in them so that you can like them and help them. But most of the most successful people are going to be challenged socially.

It makes me feel incredibly validated because I’ve never felt more awkward than being at a birthday party for another 30-year-old. I’m much more akin to 60-year-olds. My best friend is 63 years old, and I feel much more comfortable with him. But Peter Fenton once told me the best founders make you feel uncomfortable, and I think that’s very true. Have you changed as an investor, Jerry?

Yes. If you don’t change, you die. Changing is what allows you to adapt to the world.

The most interesting thing, if I look at future stuff, is Fei-Fei Li’s new startup about this visual representation of the world. What she’s clearly articulated is that language is a very subpar descriptor of the world. It’s just limited.

For autonomous agents, we need a new system where we can visually recognize the world to be able to have an objective view of reality for the autonomous agent. That’s why we have poets, because we can’t really describe the world in enough detail with language, but we can give you a feeling about it.

I think what we have to do in the world is definitely be aware that we have limitations in understanding it with language.

I think limitations are interesting, dude. I often reflect on regret. I love being in Europe. I have family. I make a lot of money in Europe, which is great, but I never tried to go to Hollywood as an actor, and I’m just wondering if I’ll regret that. What would you say to me as a wise OG?

There’s a great British actor I had lunch with once named Kenneth Branagh, and he was reflecting on exactly that question. He said he’s made the Hollywood movies, but he’s done a lot in theater, a lot with Shakespeare. And he said, “I could have been more like Olivier, who went to Hollywood and did Hollywood films late in life, but I didn’t. I chose a different path.”

For Branagh, I think he was very happy he didn’t move to Hollywood. He could have had that career, and he’s 65 or 66 now, looking back and saying, “Yeah, I could have been that, but you know what? I’m really happy I didn’t do that.”

So maybe you’re like Kenneth Branagh. You’re going to be happy you didn’t go to Hollywood.

I didn’t. We started in New York. I’ve never left Aspen. I lived in New York part-time. Jeff Horing and I founded Insight Partners in New York City, and I covered the West Coast because it was a lot quicker for me.

Jeff and I just lived on airplanes for the first 10 years, and that’s how we handled it. Back in the day, that was pretty revolutionary. Remember, Silicon Valley was all about people walking down the street. There were VCs that said, “I don’t invest in anything more than 5 miles from my office.”

If you can’t cycle there, it’s not a deal. That was the thing, I think.

Yeah. You choose your life, you choose your path. You shouldn’t regret it if you’re getting what you want and you’re growing as a person.

The most important thing looking back is, what have you learned about yourself? When I would talk to our younger partners, like you had Jeff Lieberman on your show and some of our other partners, I’d say, “Hey guys, it doesn’t help to get super rich early like you have unless internally you learn you have a sense of self and you’re comfortable with the decisions you make, because that’s the key thing.”

You can become super successful early, but you’ve got to internalize that success. Recognize where the failures are. If I have any wisdom at all, it’s because I fucked up so much and I’ve learned from it.

I made money very young, and I was also an alcoholic. The combination of failing badly, passing out drunk many times, and my mother being disappointed in the drunk son that she had actually showed me what life’s really about. I’m almost grateful for failing and being like that.

Absolutely. You’ve got to go. Hunter Thompson was a writer who lived near the Aspen area whose works I used to love. He said, “You really don’t know the edge unless you go over it.” And clearly you did. [laughter]

Dude, I’ve been over it so many fucking times.

Knowing the edge, Harry, is what it’s all about. We’ve got to all go over the edge and hope to live to tell about it. That’s what gives you the sense of knowing that gives you joy in life and gives you a sense of ease and comfort as you move through it. When did you go over the edge?

Well, I did a different kind at 19. I got accepted to United States International University in Nairobi, Kenya. So, in 1977, I got on a plane, went to Europe, hitchhiked around for a couple of months, and then I spent 6 months in Africa. I spent weekends documenting traditional dances with the director of Bomas.

I was out in the hinterland, and Kenya is the size of Texas, with 13 major tribes, 76 individual clans and dialects, and we wanted to document it all. Going out and living with these people on the weekends was going over the edge. I ate things I'd never believe I'd eat again. I lived in little huts with goats. It was not a pretty sight. [laughter]

Dude, you've built a firm that many, including me, would aspire to. I just want to understand some wisdom, knowing all you know now. What strategy did you do with Insight, or what project did you do with Insight, that, with the benefit of hindsight, you maybe wish you hadn't done? What did you learn?

Wow. One thing we hadn't done in the early days of Insight was hire some famous guy to start Insight Europe, and after 6 months we shut it down. It was a mistake and a half. We were not mature enough to do that.

We loved Europe early on. Fund 1 had an investment called SLP in Paris. Fund 2 had a company called Meta4 in Spain. It went public. That was a big win. We did deals in Portugal, and we did deals in the Czech Republic in the '90s. So we thought, “Oh, let's do a fund.” Big mistake. Too much difference. Too much.

We needed to stay in New York together. We needed to focus. We needed to learn from each other. Having people in Europe was too challenging. That was a mistake that we made. So I always say the product that we sell our LPs is the quality of our investment decisions, and that reduces with remote decision-making.

Do you agree, or do you think that today remote decisions work just as well?

That's a function of the team, that's a function of who's involved, and that's a function of experience. Overall, if I were to invest in a company, I would be very suspicious of remote management, because human beings today need it.

As we move to autonomous agents and human beings, that's a good question. That's a very good question. I don't know how the future is going to play out, but in the past, remotely managed situations—I know for me, because I didn't live in New York as a co-founder, it was really hard on my team for me to fly in, fly out, fly in, fly out.

12. The Single Most Important Factor in VC

Because you have to make decisions remotely, you have to sit with an entrepreneur. You're in San Francisco and you have to commit now, and you've got a whole team in New York. It's really tough being that remote guy. If you're all local and you're all sitting in the same office, you've got to make a decision tomorrow, it's a lot easier. So I would argue that remote decision-making is very difficult.

You mentioned momentum and the importance of momentum in some ways.

Yeah.

Critics have suggested that momentum was a challenge for the $20 billion fund, given the speed with which it was deployed and the high prices. With the wisdom and experience you have, Jerry, I'm fascinated: How do you reflect on that fund?

One correlation you can look at across all VC funds, if you look at their vintage—whether it's a '99 vintage or a 2004–2005 vintage—is timing. The one correlation you can make to success is timing.

If you did a fund in 2005 or 2006, you were in great shape to take advantage of the mobile thing that started happening in 2008, when Steve Jobs got 10 million subscribers from AT&T. That's when mobile became real. If you had a 2005–2006 fund, you had a chance to get some of the best, most iconic companies, right?

But if you started a fund in 2009, you're going to miss being early in Twitter, you're going to miss being early in Facebook, you're going to miss being in Uber, you're going to miss all that. So timing is the most important thing. Whether you're early stage or late stage, we can't apply it to just an overall general strategy. It depends on the timing of when you did the investment.

Do you think now is the best time ever to be starting a new fund?

Absolutely. Absolutely the best time, because you're in a sea change. Humans are no longer going to be the decision-makers about software. It's going to be autonomous agents.

This sea change, this tsunami, is so different than anything that's come in the past that people who embrace this new model can embrace it from scratch right now and have a huge advantage over people who have been successful with older models that don't quite move fast enough because they're already rich. They already made a lot of money. I don't know. You know, it's hard to get old dogs to move fast.

How do you feel about getting back in the game then, Jerry? You're doing this show with me. You're talking about autonomous agents. It seems like you're here.

Here's the secret. Here's the real secret.

I never left the game. I retired from Insight. I retired from negotiating term sheets. I retired from board seats. I retired from internal investment committee meetings. I retired from having dinners with LPs and answering LP phone calls, but I didn't retire.

I haven't done a single investment on my own that someone didn't bring me, that I love and trust. Every investment that I've done, which is over 100, has come from someone close to me who said, “Jerry, I need your help to help me think about this thing.”

13. The Twitter Bet: What Jerry Saw in 2009

And through that process of trying to help someone else, I end up in it. That's been fun because it's an experience that you're not doing on your own. It's an experience you're doing with people you care about.

Final one before we do a quick-fire round. Do you believe the future of venture is like “go big or go home”? We see Andreessen raise $15 billion. We have Insight, a megaplatform. We have the megaplatforms of General Catalyst and Lightspeed. Is it go big or go home now?

That's the only way I played the game. [laughter] I went big on my things. When I invested in Twitter in 2009, there were 30-something people and no revenue. My partners were like, “Are you kidding me?”

It was a watershed moment that my partners allowed me to make that investment.

What did you see?

What I saw—and this is really interesting—was that the idea of Twitter, which I call the status update for the world, was so brilliant, so over the top. It was far better than anybody on the team's ability to execute against it.

I think Twitter could have been infinitely more interesting as a company, but VCs got involved and created politics, and they fired Jack Dorsey early. I don't know if that was right or wrong, but they created issues, and the company was pretty messed up from the trauma of firing one of the founders.

At the same time, they were very excited, the community was growing, and the community was helping it. For me, I saw an incredible idea that wasn't easily replicated. There was enough momentum to it that this was, for me, a no-brainer.

It was the bet that I put my entire reputation on the line for. I got Jeff Horing and Deven Parekh to jump in with me on this, and I pushed really hard. We did that investment in less than 30 days from the time I met the team to the time we closed, and it turned out to be a watershed investment for the company and got us into lots of things like consumer and things we were doing at the time. That was just because the quality of that idea was insanely great.

Did you ever have a trough in confidence?

I was lucky because I was losing money every year I've been in this business. I failed. So I was always balanced in that I knew I was going to fail. I knew it was going to be miserable, but I knew I was going to win some, and we were fortunate to do what we did.

We were very, very fortunate, but failure to me was always part of my job and I had to just suck it up. So there was nothing that came down the pipe that caused me to lose confidence, because I'd been taking and dealing with it from day 1.

14. Quick-Fire Round

Listen, I'm going to do a quick-fire round with you. I say a short statement, you give me your immediate thoughts.

Yeah.

What advice would you give to a graduate entering the workforce today, looking for a job for the first time?

I would have him go buy a Mac mini, have an OpenClaw, and go to the job interview with his OpenClaw. [laughter]

Who is the best sourcer? When they bring you a deal, you pay most attention to it.

Oh, that's easy: the partners at Insight. What we've really built is one of the great sourcing engines of the world. Mike Triplett, Jeff Lieberman, Deven—these guys all just were focused on sourcing and managing their teams, and to get through the gauntlet of Insight due diligence and get to something that could be funded was a pretty big accomplishment.

I think those guys were the absolute best sources in the world.

Who is the single best picker that you know? They see value where others don't.

I don't want to build egos up, but Peter Fenton and John Doerr early on. Those guys were astounding at picking deals.

What is the single most memorable first M&A founder meeting you've had?

Oh, wow. This is not going to be a quick thing.

We edit the shows, dude.

Okay. I had the founders of Truecaller in Sweden come to Aspen. I said, “I'm not going to invest in them.” But they took their last dollars to get a plane ticket to Aspen, and they came to me. This was written about in The Information by Amir Efrati—the story about it.

Basically, these guys were nice guys, and I couldn't let them down. I told them, “I don't want to invest, but I'll help you.”

I ended up helping them and giving them some money and making it a successful company. It went public a few years ago, and it was a great company for a number of years. But that's one I did not want to go into, and I didn't want to go to Sweden, but I liked the guy so much. They came to Aspen, and I said, “Okay, I've got to do it.”

What's the difference between a used-car salesman and a software salesman? [laughter]

The used-car salesman knows that he's lying. [laughter]

I love it. You've got OpenAI, you've got Anthropic, and you've got Grok.

Which would you most want to be in as an ambassador?

It depends what stage you get me a shot in.

I'm getting you a shot in OpenAI at 500, and I'm giving you Anthropic at 380.

I take them both. [laughter] But I think, for me, it's OpenAI because I've got 800 million people using it.

Wow. Really?

800 million. You've got 800 million people doing it. When you cross a billion, then you're there. I'm not that big a Bitcoin fan. I'm big in a couple of crypto things, but until you get to a billion users, I'm suspicious of any consumer technology.

I get you, but do you not think Gemini and Google can still wipe the floor with the distribution advantage that they have and the speed at which they're moving?

They can, and they are long-term better assets. If I have an autonomous agent business and I've got Gmail, you're going to have a phenomenal asset that OpenAI doesn't have. But the issue is who's got 800 million people getting to a billion first.

You can say there's a billion users on their other properties, like YouTube and things. If you can convert OpenAI's chatbot to my own Jerry Murdock autonomous agent from ChatGPT, which is, I think, what Peter Steinberger is going to do, you're going to have an insanely great business.

What do people not know about having money that they should know?

Very easy answer: money does not come with instructions. Just remember that. Everything else you buy comes with instructions. You get a Rivian, it comes with instructions. You get an iPhone, it comes with instructions. Money doesn't come with instructions.

So you need to learn to respect it. It's energy. Money is the equivalent of energy, and you need to respect it. You don't just leave the lights on at home and the air conditioner on in your home all day long while you're gone. You respect energy. You treat it with respect, and you use it for good and use it to make things happen.

You don't waste it foolishly being an idiot. You don't want to go to Las Vegas and hand out $100 bills to just random people. Take that energy and do something great. I don't know. As Larry Page has said, if he had extra money, he'd just give it to Elon to let him build things to change the world.

I think that if you've got money, you've got energy. You can give it to entrepreneurs that'll do amazing things.

You had kids quite late in life.

Yeah. [snorts]

You wish you'd had them earlier?

Well, it's like, if you've had a fantastic life, you look back and you say, "It's already great. Why would I change it?" Number 1, that guy, when I was younger, wasn't ready for kids. All he was doing was building a business.

Would it have been possible to build Insight if you'd had kids during that time?

I would have been divorced and a pretty sad character, probably. Look, when Jeff and I went out to raise money and we were getting sand kicked in our face, he was 30 and I was 35 or 36. We both had acne. Nobody wanted to give us money. Me, that guy with kids? I can't imagine him. I imagine he's a disaster.

What was the breakout moment? You said there, you know, sand kicked in your face, money, and then you raised $20 billion funds in single moments. What was the breakout? Was it Twitter? Was it a—

It was the fact that we survived 9/11.

I mean, if you want to know, Insight had incredible funds. Funds I and II—these incredible successes—are still some of our best funds ever. And then Fund III, our '99 fund, was challenged, and it did pretty well, all things considered.

But the fact that we survived that collapse—a lot of VC firms were zombies, the walking dead, with partners splitting up and people going in different directions—we hung in there. We hung in there and survived. That's the breakthrough moment.

What's the biggest parenting advice you can give me now, knowing what you know as a papa?

Your kids are always, always watching you. So use that as an opportunity to be your best self. Put your heart first and be your best self every moment of the day that you can with those kids.

Final one: what are you most excited for in the next 10 years? Jerry, I like optimism. We sit in this incredible moment in technology in terms of development and adoption. What are you most excited for?

There was a movie in the 1950s called The Long, Hot Summer, and the character Big Daddy was dying of cancer. At the end of the movie, things were working out with the family, and he says, "I feel like I'm going to live forever."

I think about autonomous agents and AI, and maybe it's going to help me live forever. [laughter]

That's what I think. I actually do think it will increase longevity significantly in the next 20 years. Do you agree?

Yeah. What I've heard is, if you can live another 5 years, you're going to extend another 5 years. So if you're in whatever state of health you're in today, if you make it 5 years, you're going to at least get 5 years in the back end.

Jerry, you're a star. Thank you so much for this.

It's great, and great catching up with you again, Harry.