Fiverr CEO兼创始人 Micha Kaufman：“如果你还不适应 AI，去你的。你完了！”

Why do you think it’s my responsibility to make you better as professionals? Seriously, if you’re not going to take the time to make yourself valuable, to embrace reality, you’re doomed. I’m not going to help you.

If you don’t want to do this, you’re either going to be poor or a burden on society. And again, this is a part of this awakening. Wake the up. Grow up.

Micha, it is so good to have you on the show today. I’ve been looking forward to this one, so thank you so much for joining me.

Thanks for having me, Harry.

Not at all. I’ve heard many good things for a long time. I had Adam from Bessemer on a while ago, and he sang your praises then, so I wanted to make it happen. Then, when you did your recent post, I was like, “We’ve got to make this show happen.”

1. AI Is Coming for Your Job

I want to start with this: AI is coming for your jobs. Heck, it’s coming for my job, too. This is a wake-up call. Can I ask what spawned this post? What was the reason or catalyst for it?

This was boiling for a while. I think that this is also echoing some of my team’s thoughts about their futures as professionals. For some reason, some of it might also be frustration at not seeing people go through this wake-up moment.

I really wanted to validate a lot of what people were thinking between themselves or in small groups, but be very blunt about it. I also wanted to offer some direction, or at least some thoughts about what you can do about it, and then have this as an invitation for a discussion, which we did have after the email. This was surprisingly large.

I asked my chief of staff to assign a room that could hold, I don’t know, 50 people, and I wrote an email saying, “Look, I’m going to spend 3 hours sitting in that room with my laptop, just working. If you want to have a conversation, I’m there. Just come sit beside me. Let’s chat.”

2. Confronting the AI Threat to Fiverr

We scheduled that time, and I came to the room. There were about 250 people crammed into it. I wasn’t prepared to give a lecture. The idea was that this would be a conversation, and it turned into what started with a brain dump on my part and then just having a conversation, which is pretty hard to do with 250 people.

Can we start with that brain dump? Bluntly, you’re a direct dude. I’m a direct dude. My question to you is this: We’ve seen Chegg be bluntly killed as a business by AI, and Fiverr is one that people say is very vulnerable. I’m sure people talk about it with that same sense of vulnerability. When you did the brain dump to people, what did you say?

It was, again, very blunt. What I told the team is this: I said, “In an ideal situation, my expectation of you is that each and every one of you, and it doesn’t matter what you do, is going to replace 100% of what you do with automation.”

Maybe it’s impossible. Maybe you can’t get to that number, but ideally, that would be the number. The obvious question is, “Great. If I can automate 100% of what I do, you don’t need me.” My answer to that was, “Actually, it’s exactly the opposite.”

If you can automate 100% of what you do, that means that now you have 100% of your time free to figure out what the things are that you can do that cannot be automated. You can spend more time having nonlinear thinking. You can work with me on strategy.

As I was writing this, it dawned on me that there’s something very interesting about this process. When you understand how AI augments who we are and how we do things, maybe, in a weird way, what’s happening with AI is forcing us to rediscover our humanity.

In many ways, if functions that we do as humans can be automated, then there’s no reason for a human being to do them. But what makes you, Harry, special? What makes you special is the fact that you have a very advanced taste for things and you can determine right from wrong.

Is it the fact that you are great at ideating and thinking in a way that is unconventional? What is making you special? This forces you to actually get down to understanding your humanity again.

There are so many things I want to unpack there. You mentioned finding your humanity again. An inherent part of humanity is also vulnerability and fear—running from things that can kill us.

I literally just came from a board 10 minutes ago where the CEO was like, “My team is not embracing AI because they’re scared that it will replace them over time, gradually. How do I, as a CEO, encourage adoption without fear?”

That gets me to another point that was super blunt. When you think about high-tech, the average age is very young. I’ve been building companies for almost 30 years, and when I started developing my career, it was very obvious that, as a professional, you need to be up to date. You need to be able to be valuable, and in order to do that, you need to constantly upgrade yourself.

But I think that what happened in the past 10, maybe 15 years, was this notion that your place of work has the responsibility to make you better. My point to the team was, “Look, guys, this is an aberration. You’re getting everything wrong. Let me tell you how things really are.”

Whose responsibility is it to make you a better person? Whose responsibility is it to make you a better spouse? Whose responsibility is it to make you a better dad, a better parent? And why do you think it’s my responsibility to make you better as professionals?

Seriously, if you’re not going to take the time to make yourself valuable, to embrace reality in a good way, you’re doomed. You’re doomed. I’m not going to help you. It’s not my responsibility.

But I will tell you what I will do: I will help those who want to go through that process. If you want to help yourself, I’ll be there for you. That’s my vow, okay? But if not, you’re done. You’re done. You’re not done at Fiverr. You’re done. Period.

There’s not going to be demand for people who cannot upgrade themselves, people who are having a bad time coping with reality and understanding what it means. I’m sorry. This changes now, and each person should figure out what they want to do with their time.

When you were saying that, it’s funny—I was like, “No, I disagree totally.” A husband or a wife should want the other person to be better, push them, and challenge them in ways that make them grow.

But to your point, it starts with you. The additional support is good if your spouse wants to help you become better, but if you’re not interested in doing that, it’s not going to happen. If you don’t put in the time, it’s not going to happen.

I’m putting in the time to provide the infrastructure for growth. That’s fine, but I’m not responsible for you. I’m not your dad.

I agree completely with that.

3. Do People Really Want to Work Anymore?

My subsequent question, without sounding bleak: I’m very good friends with [likely Jason Lemkin from SaaStr], who has quite a bleak summation of this. He says, “No one wants to work today. Young people don’t want to work like they used to.”

When you reflect and look at how much of society really wants to embrace that change versus how much doesn’t, what do you think?

I think that it’s going to be wrong to try and understand reality through the lens of the current present. Right now, we’re in a very fuzzy state in time where things change very, very fast.

I feel that we are in a period that is very similar to the dot-com era, but it’s happening now. It used to be Pets.com; now it’s Pets.ai or whatever. It’s the same.

In my feed, there are 40 startups every day. There is no need for so many. But everybody is trying to cope with the endless amount of information that comes in, and this is very discouraging.

I don’t think that this is going to be the situation in the future. I don’t think that the way people judge the present has too many implications for the future, because we’re in a very unstable world right now.

Now, the question of whether people want to work or not, in my view, is a personal question. If you don’t want to work, the exit from the building is on the ground floor. Bye-bye. That’s fine. I don’t think that it is my primary function to figure this question out with people.

My assumption is that those who want to come and engage in the activity of working do so either to make a living or to express themselves or in the search for meaning. Usually, it’s a combination: making a living, expressing yourself, and the search for meaning.

We spend the majority of our time—our awake time—at work. It needs to have meaning. If it doesn’t have meaning, then it’s very miserable.

If you don’t want to do this, you’re either going to be poor, a burden on society, or a burden on your parents. And again, this is a part of this awakening. Wake the fuck up. Grow up.

If you don’t want to work, don’t work. I think that, as human beings, we’re by nature competitive creatures. Why? Because we need to compete for food. We need to mate.

If you translate this into modern times, you want to have the best-looking girl at school. You want to have the fastest car. You want to fly business, not coach. These are basic things. They don’t come out of thin air. They come out of hard work.

In my view, it’s more important to actually look at the meaning part of it.

You know what I love about my job, honestly, is the fact that I admit to my woeful lack of experience in some respects. You have 30 years, and you draw the analogy to the dot-com era.

When you look at today and the dot-com era, where do you look at today and go, “That’s unsustainable and a sign of bubble-like times?”

Oh, look. I mentioned 40 new companies a day on my feed.

There is no market for so many companies. By the way, it started 8 or 10 years ago, when I started talking about bubbles. When you look at markets and the number of companies that are actually competing for the same customer, for the same function or need, you realize that there are just too many companies for a market that cannot entertain that number of companies.

I think it was boiling to a point where the market needed to have a cleanup. Sometimes cleanups happen with a lot of companies going bust, and sometimes they happen through consolidation. You say, “There are 5 companies competing for a market that is very specific.” In cybersecurity, this is a market that can probably entertain 1 or 2 companies in that subsegment. Someone is going to win, which by definition is not going to leave that much for the rest, which means that there’s going to be a cleanup either through consolidation or through companies going bust.

Right now, what we’re seeing is on steroids. I was reflecting about this a while ago. When I started building my first company, I referred to this as “time to copy.” You create something that is outstanding and you succeed. How fast would someone copy you? Maybe 25 years ago, this would have taken a year, depending on the sophistication.

When we launched Fiverr 15 years ago, time to copy was between 6 and 9 months. The time to copy right now is, I don’t know, 10 days, whatever. For simpler stuff, it’s probably nothing. The influx of the amount of things that are being built, when they don’t have a differentiator and don’t have a delta in what they’re adding to the existing product, is really marginal.

There’s another thing that is really interesting about this. You can see this, although it’s hard to measure exactly. Think about the horizon of building something. It used to be that you had Mount Everest in front of you, and this was the mountain of building something. You had to spend a tremendous amount of preparation and energy to climb the mountain. When you reached the peak, what you realized was that there were 10 other Everests in front of you that you needed to climb.

4. Why Your Startup Can Be Copied in 10 Days

What’s happening right now is that I think we’ve received a lift. Now you get a ride to, let’s call it, 300 feet from the peak. You can take a ride, and usually when you build stuff that could be Lovable or whatever, you get stuck at some point. Then you need to climb the rest of it with a professional or whatever, getting help. But when you reach the peak, what you understand is that now it’s really easy to climb. I’m still at ground zero, though, because who’s going to know about what I’ve built? How would I market this? So, it shortened that phase, but it didn’t change the actual tough issue. My sister builds apps. Great. This is what’s happening with AI: it’s so easy, everybody is building. But what’s the true value? It’s very small to none.

That was my question to you, which was on time to copy. Does that not just reduce the importance of time to copy? Respectfully, I would say that Fiverr’s value is built in its brand, its distribution channels, and the trust that it has with the supply side. It’s not the front end of the website or the payment system of the website. It’s all of the externalities and other products that make it an enterprise-valued company.

You’re right, but here’s the “but.” When the company was 9 months old, there were already 30 copies, and years after there were thousands of copies. We were having a debate between ourselves about what to do with it, and we had 2 camps.

There was one camp that said, “Let’s sue them all. Let’s go after everyone. Let’s fight it. It’s unfair. People should not be just copying our labor, our concepts, and all of our hard work.” In my camp, the way I was looking at it, I said, “Look, the advantage that we have is double. One, we’re first to market, and two, it’s speed. We’re just running really, really fast.”

When you think about the analogy of running, your head can point in only 1 direction. It’s either in front of you, or you look behind yourself. Try running while looking back. I’m telling you, you are going to slow down because you lose your balance. You’re afraid to hit a wall or a pole. You need to look straight and run as fast as you can.

The second advantage is that they don’t know what we know because we’re ahead of the game. We’re the ones who are innovating. We’re the ones who are actually figuring out how to solve the very fundamental, tough questions and challenges. As you said, there’s the front end and there’s the magic that happens behind the scenes.

But here’s the “but”: we had 9 months of head start, which were the 9 months where we started seeding the brand, creating something powerful, and creating a movement and a community around it. If the time to copy is 10 seconds, you don’t have that time to develop it. You leave more to luck. Everybody has luck—I mean, luck has a tremendous portion in success. I’ll be the first to admit it. But the portion of luck varies, and if you don’t have the head start, then you need to be super, super lucky. All the stars need to align for you, and I think it just makes things very hard.

What I see a lot of entrepreneurs doing these days is just building off of existing stuff and trying to introduce a marginal advancement on what was already built because their imagination is not large enough. The fact is that when you do incremental stuff, you get incremental results.

Does defensibility not shift in terms of where it occurs in the system? I don’t mean to sound like a VC with a load of words, but if you think about time to copy being where defensibility was before, now time to copy is almost nil, so there’s no defensibility there. But it’s in the mid-spectrum, where you’re at $1 million to $10 million in ARR, where you’re getting great data from high-quality customers, enterprise or consumer. The way you collect it, cleanse it, and utilize that data—maybe defensibility just shifted to different parts of the spectrum.

You’re absolutely right, and I think that this is where, again, as I was saying, AI forces us to discover our humanity. These are the things that are going to generate the most interesting startups. It’s that level of thinking and that level of building. I’m not just building for the sake of building. I’m taking the time to understand their real needs and their real challenges and trying to build off of that, instead of saying, “People like this, so let’s make X.1 or the next version of something.”

5. How Do You Invest When There’s 20 Startups in Every Category?

Another reason I love the show is because I get to ask advice from much smarter people like you. I’m an investor for a living, too, and there are 20 companies in every single category that I invest in. I just invested in an LLM optimization engine. We’ve both seen 30 of those. Don’t laugh—we’ll edit that out. My question is, how do you advise me as an investor to navigate and operate in a new world where every category has 20 companies, not 2 or 3 like it was 10 years ago?

This goes down to maybe a more fundamental view of what’s happening with technology right now. Technology by itself doesn’t give an advantage to anyone when it’s 100% democratized. Here’s the tricky part: the go-to-market of OpenAI in November 2022 was as democratized as possible because they gave it for free. You can’t be freer than free.

What happened with that is—and this is what people don’t get, and they definitely didn’t get at launch or the year after, but they’re realizing right now—everybody felt smarter. “I can write incredible stuff. I can do research much faster.” Yes, but so does everybody else. It’s super powerful, but it doesn’t give you an advantage because it’s giving the same advantage to everyone, which makes everyone get back to ground zero. It’s elevating the standard, but it’s just picking everyone up to the starting line.

As you think about how to apply this to VC thinking, it goes back to the oldest way of making judgments, which is to look at the people running the company. You make decisions based on people. Technology will evolve, ideas will change, and companies will pivot, but who are the people running the show? How talented are they? How passionate are they about what they’re doing? Are they 100% opportunistic, or are they missionaries? They’re not going to rest until they get it. Are they adaptable? Do they think about problems from first principles, thinking about fundamental truths? Do they have these things?

Invest in people. At the end, I’ve invested in tens of companies, and it’s the best people that get it. Period. I wrote off companies, saying, “Okay, this is money lost for sure,” and these were the most successful companies I’ve invested in because they had the best people. Period.

Which company did you write off that was the most successful?

This would be brutal. They were successful, so I’d be thrilled. It’s a public company, and I don’t know if telling the story would do well. This was a company at seed stage. It was a tough market. It was a tough nut to crack. The company had internal things between the founders—I mean, all the tough stuff. I said, “There are so many reasons why this company needs to die,” but it didn’t.

I mean, people—I think one of the investors in Etsy talked about them in the same way. He said, “You know why I invested in it?” I said, “No, why?” He said, “Because this company should have died at least 7 times, and it didn’t.” This, to me, speaks for the strength of the company and the strength of the team behind it, and this was the same story with this company.

I lost a little bit of connection with the company at some point, which is fine, and then I realized that they were going public. I was like, “Seriously?” I think this is probably the best investment I’ve ever, ever, ever made. The multiples were in the 50s—crazy, crazy story.

It’s all about the people. I always say seed is about 3 things: founder being directionally correct. In other words, you don’t make pillows or beds. And 3 is the deal. I’m not going to do 50 on 250 for a seed. If the 3 go, awesome. So there you have it. I don’t think that this principle has changed, and to be honest, I don’t think it’s going to change.

6. Is an AI Bubble Burst Coming?

You made the analogy to the dot-com era. Do you think we will have an AI bubble burst in the next few years, or do you think the exponential advancement in the core technology means that we’ll sustain excitement and financing in terms of the number of companies that are going to be around?

Absolutely. I don’t know, maybe 999 companies out of 1000 are not going to be around for a year, maybe 2. It’s going to be ruthless, and I’ve already explained why. I think there are just too many companies that are not introducing interesting advancements, and I think it’s opportunistic. By the way, there’s nothing wrong with being opportunistic, but the market can’t sustain it.

I think the market is going to start sorting itself out. You’re going to see, much like we’ve seen before, that the dot-com era also gave birth to Amazon, right? There are companies that have true value, a true mission, and the patience to actually build that and stay true to their core reason to exist. It’s going to be the same here. The market is going to arrange itself.

There are going to be very large players that are going to be foundational players, some of whom I think we know or have a good chance of being these companies—the Googles of the world, likely the Anthropics, the OpenAIs. I don’t know what the play is going to be exactly, but it might be reduced to the equivalent of paying for CPU cycles. It’s just going to be GPU cycles. It doesn’t matter, but it’s essentially a cloud play. Then there are going to be the other application layers on top of it.

Again, I don’t think that we can judge from where we are right now because it’s changing too fast for the market to get a chance to organize itself. I think it’s going to go through a cleanup process, and that cleanup is going to be triggered by investors. At some point, they’ll say, “You know what? I’m out. I’ll let it sort out, and then I’ll figure it out.” So that’s going to dry up.

I was on a call with a massive bank this morning, and they were like, “Harry, there’s no way that we can use Notion, let alone use AI tools, dude.” Enterprise adoption is so far complicated. Why? What was the reason? Compliance? You cannot, for data and privacy reasons, record calls and then take summarized notes. Not a chance. I mean, they laughed me out of the room. Micha, look, fine. So they shouldn’t use it, which is going to be another reason why banks are going to be left behind, whatever.

But look, it needs to be sorted out. If compliance is the issue, compliance is going to sort itself out. Some of it is going to be hosted on-premises, some of it—whatever. There are solutions for that.

Do banks use any cloud services? I guess they do. Is the cloud safe? Well, here and there. Do you have companies on the cloud with compliance? Yes, you do. So you can use them. Nothing is 100% safe. I don’t think that these are the interesting problems, to be honest.

What do you think are the most interesting problems?

I think that questions like, are we okay with the fact that truth is dead? We can’t even tell what’s real and what’s not. Are we okay with not being able to differentiate between things that are human and things that are not? When are we being manipulated and when are we not?

There’s an even much greater and much more interesting question: How do we consider ourselves as human beings and our place in the world? I’ll give you my view. My view is that humans—you and I—are the center of the universe. Everything revolves around us, acting from an assumption that we are in the center. Are we okay with losing that center?

What’s happening right now is going in that direction. If we continue to develop technology in a way that eventually is going to be unsupervised because it develops itself, and we feed it because it’s based on us, but it doesn’t care about us—AI is now eating the world. It is learning from everything everyone has ever created. It’s mixing it. It’s churning out new results without even saying, “I’ve just generated a nice podcast based on what I’ve learned from Harry.” No, Harry, right? It’s just generating stuff.

My actual concern is that if this goes on over time, what’s going to be our motivation to continue creating and sharing stuff if it has zero value? Some of the motivation to create is internal, but as human beings, we need feedback. We want to see how the things that we create impact others. This is why we share. This is why you have open-source code, right? People can use what you’ve done, but they can build off of it. If this has no value, then why would we create? And if we stop creating, why would it—

Why would it have no value? If you look at AI-generated content today, it has—

Yeah, not quite the same value, but it has a similar value.

No, but that wasn’t the question. The question is, what is the value for those who created the actual content on which AI is being trained? If AI generates something because it learned from me, I don’t get any credit for it. I don’t get compensated for it.

Copyright is dead. Essentially, it is dead. It’s a notion from 1710, and it’s dead overnight. There is no copyright. It’s done. It’s dead.

7. We’re Creating for AI, Not People

I’m sorry, can you help me understand why copyright is dead when so much of content and media is predicated on it? Just help me understand. I’m really naive.

If you look at the arc of advancement of humanity, Homo sapiens have been around for, what, 300,000 years? It was pretty linear for a long while. What changed the pace of advancement of humanity was the printing press, or the printing machine, because it allowed for the distribution of knowledge, which made more people knowledgeable, which inspired them to create stuff and then distribute it through the press so that more people could enjoy it.

Very close—maybe 100 years, maybe less—after Gutenberg, there was the 1710 Act in the UK called something like the Act for the Advancement of Learning, which was the basis of copyright. It said, “To motivate people to create and share, we want to give them protection.” One is their right for their creation to bear their name, and then to commercialize it if they want. One is a moral right, and another is a commercial right.

This is how we advance, because people wanted to both create stuff and share it because they got the recognition for doing this, and because it created connection. You knew who wrote this piece, so you could get in touch with them, and maybe it nurtures new ideas, maybe there are forks, and maybe it inspires other people to do stuff.

This was on steroids over a course of 500 years with modern media, with the internet, which is incredible, and with AI. So now the distribution of information is incredible. But what happened with AI is that it jumped over the hoop of attributing anything to anyone it is based on, which is why I say that essentially copyright is dead. Everybody calls it fair use. I’m fine, whatever. I’m not a regulator. I’m fine.

But my theory is that if people produce stuff and publish it, and it gets eaten and churned out without any recognition, without any connection to what they’ve done—

Their motivation to create and share is going to start slowing down and then decreasing because you lose the benefits of actually creating, getting the recognition, and maybe commercializing it, which to me is troubling. I don't know if this would be the case, but I'm raising a flag. I'm not raising the flag as the CEO of Fiverr.

Look, my place on this planet is finite. But I have kids. There's a future. I think about the fact that everything that happened in the past 500 years made our quality of life much, much higher. I don't want this to slow down, and I don't think, in my view, we should be the center. We shouldn't be right now.

Harry, I have news for you: You are working for AI, and so do I, because we're producing content, and that content is going to be eaten by a machine and used to produce new stuff. It sounds good. I don't know. Maybe fine. This is up to you. But I think we're at risk of losing our place in the center when we start working for something else.

8. Do Public CEOs Need an AI Take Now?

You're very vocal in a very refreshing way, in a very unguarded way. Do you feel the pressure to be vocal on AI given your role? I think a lot of public CEOs feel the pressure to have an AI strategy. Do you feel the pressure to have an AI opinion?

No. I wouldn't use the word pressure. I would say that I have a responsibility. It's not the same thing.

The garbage of once-great companies that are now no longer with us is overflowing. A lot of the reasons why these companies are not around is because they couldn't sustain transformational waves. Compare this to surfing: You have to start rowing to get a wave, and then you can ride that wave.

It used to be for 7 or 8 years, and then you needed to prepare yourself to catch the next wave so that you could ride it. Today, those waves are much shorter, with less distance between them. It is the responsibility of a leader of a company to ensure that you're able to be opportunistic about each one of these transformational waves.

In some ways, they force you to invent your act 2 and act 3 of a company. Most companies don't manage to go through even the successful ones—they do not manage to go through phase 1 of the company, even if they're a big company already. But there's always going to be a transformational wave. If you're unable to catch it and make use of its energy to surf for a period of time, then you're going to lose it by definition. You start getting left behind, waiting for the next wave, or just going past.

What are your biggest questions about whether you'll be able to transition to the next wave and stand on the board to ride it?

I was asked in another conversation, "What is it like to be a CEO right now?" And I said, "You know, it's like you're asking a captain of a ship in the middle of a storm, 'How is it to be a captain?'" The answer is, "It's wet, it's dark, and you can't see a mile ahead."

Does it feel like that? When you compare 15 years of Fiverr, how does today feel in context?

My dad was a very senior executive in the semiconductor business. The beauty about that business for many, many years was that it had Moore's law. Great. Unless it breaks, you can extrapolate and say, "I know exactly what's going to happen in 10 years to the power of computing and the size of chips." Period.

We can reminisce about those times and miss them, but that's not the case anymore. Right now, the cadence is super-fast. People that ask me, "What do you think the business is going to look like in 5 years?"—I start laughing. It's a laughable question.

You can have grand general ideas, but people that ask you to describe it are just out of touch with what we're seeing.

I think that, because of a lot of changes, let's just push on that. Why is that a laughable question? I think it's perfectly legitimate to try and stretch someone's thinking. Sergey Brin was asked the other day, "What will search on the internet look like in 2030?" And he said, quite understandably, "I don't know. We're taking every year as it comes." But I don't think it's a bad question to try and stretch.

I think it's a fair question, and I think it's a fair question that we should ask ourselves constantly, but also understand that right now we're in this intermediate phase of technology, where things are not sorted out. The signal-to-noise ratio is impossible. Then there is the question of whether it will get regulated and how that would change things. Will countries and governments actually take hold of it?

What the AI companies are doing right now—the foundational companies—is like privatizing the Manhattan Project. Imagine that they do build the atomic bomb, and now they're sitting and saying, "I don't know, maybe we'll do it nonprofit," or maybe they say, "We'll license the atomic bomb." Do you see a government allowing this? Seriously. Think about that for a second.

If you get to AI that cures the worst disease in the world, which actually is worth a lot of money, do you see a government leaving this in private hands? If this could create the ultimate weapon, if this could be the ultimate solution for infinite food and energy—no way, dude. No way.

So we're pre-all of this. To try and extrapolate 5 years from now? Seriously.

But do you not think we're just seeing the commoditization of all model knowledge and advancement when you look at [likely DeepSeek] and the 12 different open-source models in China? China versus the US is kind of a moot point at this stage because the—

No, I don't think so. Again, we're pre-expensive AI. We're pre-AGI. We're pre-superintelligence. We're not there yet. We're probably not that far, but we're not there yet.

Right now, companies are allowed to play. At some point, it's going to cross a line where governments are going to say, "I'm not comfortable with that level of power. Not in my hands." What happens then? They become nationalized. I don't know. Some functions of AI might be prohibited and kept only for governmental branches.

I don't know. I'm thinking about these questions. I don't have answers because these are equations with too many variables.

9. Tech’s Wealth Gap: Are We Creating a Monopoly of Everything?

How do you feel about value dispersion being concentrated? What I mean by that is, it feels like today's environment is more and more about the 0.00001%—Anthropic, OpenAI, NVIDIA, TSMC, Google, Facebook, Microsoft—all becoming the multi-$10-trillion companies, and then everything else kind of languishing in no man's land, in a way where that value chasm has never been greater. Do you worry about that?

But is it really different from AWS and Azure? How many cloud providers are there? Probably plenty. How many take 80% of the market? 2, 2.5, whatever. That level of concentration has existed for many, many years. How is this different?

Do you like being public?

Sure, why not? There are many reasons why I would advocate for being a public company, but there's time for that. If you're under the illusion that you can actually control the market or influence your share, don't do it. It's a bad idea.

For those who are long-term thinkers and understand that this is a long game, it's a great choice. It has a tremendous amount of benefits.

10. The HubSpot Problem: When Feature Overload Kills Execution

Help me understand that, because the feedback that I get is that the market relies on predictability and values predictability almost more than anything. Which means that your ability to invest in very long-term projects, which do not have concrete upside guaranteed and are timeline-variable, is less available to you. And so, if you have a long-term mindset, surely it would disincentivize you.

If your capital structure is very solid, if you generate free cash flow, you're your own engine. If you have very solid execution, which means that you have high trust and confidence from the market, and you decide to make bold moves, you can raise more money. That shouldn't be that hard.

Will you have more or fewer engineers in 5 years' time, do you think?

Probably more. I don't know how much more, but probably more.

Why more?

Because of the same reason I said that technology levels the playing field: Everybody has it. So it's square 1 all over. Great, now you can squeeze 3x out of each person. So does the next company. Great. How do you do more? How do you move faster?

In a market with a lot of uncertainty, one of the biggest strengths of a company is speed. It's just your ability to move super-fast.

When is more worse than better? What I mean by that is, HubSpot said recently that they're producing so much code they can't put it into production, and there are too many features to release. It's just too much.

The fact that you can generate more doesn't actually mean that you can just generate more. Which is why, when I talk with my team about the concept of moving fast, I never use the term speed. I always use the term velocity.

Speed is just the speed of movement. Velocity is speed plus direction. So speed is not enough. You need energy in a certain direction. If you're unable to push the code that you generate, it means that you didn't solve your infrastructure to be able to do this.

So it means that your priority is incorrect because you don't solve the things—the bottlenecks—that are actually keeping you from moving fast. So take the same energy and, instead of building an infinite amount of things, try to figure out, “How can I entertain that many ideas?” Solve your infrastructure first, then start building like crazy. Plus, you're probably building stupid things. Stop doing that as well.

Going back to this conversation that I had with the team, this meeting with 200 to 250 people in a cramped room, one of the things that I told them is, as I'm thinking about velocity, I'm thinking about how we reinterpret it. The result of it is just reducing the cost of failure. Essentially, all companies—the majority of what they do fails. The reason why you don't know this, or why some people are not aware, is because they don't talk about their failures. There's nothing to talk about; it failed.

If you take a company the size of Amazon, I'm guessing they have 5,000 tests running concurrently because they have enough traffic to sustain that. The vast majority—maybe at some point in time, all 5,000—failed. Who cares? But if building something took 10 engineers 3 months and you failed, that's unpleasant. If it took 1 person 3 days and it failed, great.

So we reduce the cost of failure so that we can test more things and have more progress. The way I was thinking about this is: How do we double or triple the output per unit of time, and the same for quality per unit of delivery? By doing so, we just reduce the cost of error.

Now, if you do this and everybody else is doing this, you're again at the starting line, and so you need to have those differentiators. Developers—the profession of being a developer—is going to change dramatically, but you always need brilliant people. Will you need many, many more of them? I don't know. But you asked me, “Is your engineering team going to be larger?” Yes, I think so.

Does AI help normal 1X developers become 10X developers, or does it help 10X developers become super-super-superhuman developers? The second one is an easy yes. The first one is probably an easy no.

What function, if engineering actually becomes more a part of our organization and more important, do we have today that we will not have in 5 years?

11. Which Department Took the AI Hit at Fiverr?

Everybody talks about customer support as an example. Which function has been most impacted at Fiverr by AI? Customer support.

The area that is going through the most fundamental disruption is actually marketing. If you think about coding and how developers are actually using AI, it's really interesting. When you look at the fundamental function of all types of copiloting, what a lot of them are doing is actually what developers hated doing: copy-pasting from open source, looking for a library, then reading the repository notes and figuring out if you want to use it as is, go through it and look for security issues, or fork it and build something off of it.

It's like doing research on Google, following blue links, copy-pasting stuff from different websites, and putting it into a document. Who likes to do this? It's robotic.

In marketing, it's a whole different situation because there is no AI at the level that we have in programming in marketing. You need to figure out how to use AI without screwing up or giving it too much freedom, because essentially, right now, we're still marketing to human beings. Whatever is being produced will be consumed by human eyes or ears, and therefore there is a much deeper sense of understanding human behavior when you do marketing.

I think this is where marketers get tempted to use synthetic technologies to automate their work. But I think it's not ripe yet, and they're not necessarily automating the right functions. I think this is less-charted territory in AI at this point.

It feels like there's no need to write copy, no need to manage social, and no need to create ads. Everything could be automated. I think what we've seen over the past year or so is an awakening from the illusion that this could be fully automated.

As someone who spends a huge amount of time in social and viral content, the truth is, you're absolutely right. For your Virgin, NatWest, Chase Bank social media manager, you're already, “We're thrilled to announce the launch of our new product.” See you later. Gone.

But if you're a master of social and viral social posts, and it includes nuanced context from current events, media, and very trending GIFs that are particular to today—not yesterday or tomorrow, today—the way it's structured, it is an art form. I can't even hire people today and pay them $1 million. Literally, I would pay $1 million for someone to do social like me.

So I think that in marketing, the entry-level juniors have been replaced the fastest, definitely faster than developers. Even if you use AI, at some point you still need to open the code and understand what the heck is written there. If you cannot do this, then you don't have a job, but even relatively young developers can do this. It's not the same with marketers. I think that this is the area where there's the most amount of disruption.

12. Leadership Adaptation in an AI World

Final one before we do a quick fire: In terms of your leadership style, have you changed anything in the way that you communicate—the tone with which you communicate, the style? Have you changed your leadership approach in a world of AI?

I don't think so. I don't use AI to write, which is very important for me. [likely Yamini] from HubSpot said that she writes her public updates with AI.

We've been talking to other public companies, friends from other companies, and saying, “We need to—we should—build our own investor-relations AI so that we don't need to do these calls every time. We don't need to answer the questions. We don't need to write the letters to shareholders.” All of that is—anyway, most of it is read by machines for algorithmic trading or whatever it is, and they analyze the tone, whatever. So let's feed them with their own.

When you have those CEO groups and conversations behind the scenes, is the consensus that AI is further along than people think and we're moving toward more unemployment more quickly? Or is it further away than we think and not as close as people assume?

I think most of them are just making the healthy assumption that we're closer to it than further from it. If it's further away, then great. Let's chill. It's fine. But if you assume that this is closer, then you at least think about what's necessary, or you think about the implications, which I think is the healthier thing to do.

I haven't seen companies go through massive layoffs. I think companies that had fat in them started doing that because they understand that it's very— This is what I wrote in my email as well. I said, “Look, before we figure out that we need more people, let's just make sure that we're extracting the most out of ourselves, out of our existing team.”

This was true 5 years ago, which is why, if you run a lean company—not too lean, so you can actually be aggressive, but lean enough so that if you say, “Oh, fuck, I need to cut costs,” there are, like, 50 people who are going to be the first. If you know this, you should say goodbye.

The message was: How do we maximize our capacity before we actually think about bringing more people? If we think about bringing more people, these people need to be AI natives in whatever they do. I don't care if you're joining the legal department, finance, or customer support; you need to be very well-versed in this new world because, again, I'm not going to teach you.

What cost would you most like to cut today but, for whatever reason, you can't? Shareholder pressure, whatever it is.

In general, if I could cut marketing costs, I would always do it, just because the only way you can actually do it is because you can grow without it. If your organic component in growth is crazy strong, then you have a good case for going even deeper into your marketing and saying, “Do I need the incremental additional customer that I'm actually paying money for, or not?”

But this is not because of pressure. This is just because we're not in that situation.

13. Quick-Fire Round

Dude, I want to do a quick fire. I say a short statement, and you give me your immediate thoughts. What one belief about AI have you changed your mind on?

The way modern AI burst into our lives was driven by a nonprofit organization, which made me think that this could be a global endeavor that would probably give more power to more people and not more power to fewer people. Obviously, with reality, I've obviously changed my mind, seeing how OpenAI changed its strategy. In terms of being in favor of them or against them, obviously Elon wants OpenAI closed. You were open, and now you're closed.

AI, as a very fundamental technology engine, I would probably rather have open than closed.

Tell me, you've got a child who's entering the workforce. They've just graduated from university. What advice do you give them?

I'd probably rather give them advice before they go to university, and that would be: don't go, because I think it's a waste of time in most cases. Look, if you're up for finding chicks and getting drunk, that's fine. If you need someone to tell you to wake up and get to a class, and someone to scare you with an exam because otherwise you're not going to learn, then this is a good institution to do it.

In 90-plus cases of modern professions, you can study on your own, and you can do it in a fraction of the time. By the way, if you don't like reading, great—we have video for that. You don't like video? There's voice for that. If there isn't, ask AI to produce something for you. Whatever. In my view, it's just a waste of time.

What strategic opportunity did you not take with Fiverr that you wish you had taken?

One of the strategic opportunities was to create OnlyFans many, many years before OnlyFans, and we decided not to do it.

Sorry, how does that come as a strategic opportunity, and why did you decide not to do it?

If you look at the size of OnlyFans, you understand why this could be a strategic opportunity. But from the point of view of our values as founders and the values of our board, we decided that this was not a direction that we wanted to venture into.

The reason why I'm saying that we could have done it is because this was actually bubbling within the first versions of Fiverr organically. We didn't have a category for that, obviously, but we started noticing that this was happening behind the scenes. There were code words that people spread on social media and said, “Oh, go to Fiverr, do a search for X,” and what it actually means is what you think it means.

We started seeing that in the system, and we decided that this was not a direction that we were going to go down.

Is Fiverr your last company?

I don't know. Right now, yes.

Look, I have this exercise that I do every year, where I take a day or two or three aside and do an accounting of how I feel about the next year. I'm asking myself, “Am I the right person to continue leading this company?” There's a process to it. I'm going through this process, measuring my energy, my contribution, my leadership, my interest, and my curiosity.

Do you not identify so strongly with your company that it is a part of you? I would have a very hard intellectual wrestle if I realized that I wasn't one. I started this when I was 18. I'm now 28. To not be a 20VC would be very difficult for me.

Maybe, yes. But at the same time, if your contribution to the success of the company is not maximized, I don't know if—

I have a rule: if I wake up 5 days in a row saying, “Oh, no,” then I'm done. It has never happened. You can have a day. Maybe you can have 2, but if you have them in a row, that's a bad sign.

Sometimes you fall out of love. Sometimes you lose your energy. Sometimes life causes you to attend to other things. If that's the case, yes, maybe you consider this to be your baby, but it's time to move on and find someone who's going to do justice to your baby and your creation. It is a hard exercise, but I think it's worth doing at least once a year.

When are you happiest, and is life about happiness? I know it sounds weird, but [likely Nikolay] from Revolut said to me, “I don't think life's about happiness. I think it's about success and winning, and that often leads to happiness, but it's not about happiness.” When are you happy, and is it about that?

I think it's about meaning. When you look for meaning, when you pursue meaning, you're going to have a mixture of misery and happiness by definition. But I think that as long as you're in your element, as long as you feel that there is very deep meaning to what you do, then by definition your core is going to be very happy.

Your life might have swings because we don't control everything in our lives—the surroundings, our family, our health, all of that. But generally speaking, I think meaning defines this for me more than the word happiness.

Final one. When you think about how you most want to be remembered—I know it sounds weird, but I think about it a lot with grandchildren, actually—what will my future generations say? How do you want to be remembered?

I think I'm a part of a legacy. There are a lot of things that have passed to me from my parents and grandparents that are at the very core of what defines me as a human being, what defines me as a husband, and what defines me as a dad.

A lot of it has to do with being a good citizen of this world. We've touched on some of these things: this idea that it's always better to be a giver than a taker. In general, you can't just take. I hate takers. Maybe they'll remember me as a giver.

The world doesn't owe you anything. That's one of the values that I instill in my kids. The other is that home is a safe place. It doesn't matter what happened in your life—I don't know, you killed someone by mistake—come home. We'll solve it together.

There's this idea, which doesn't mean that you're not going to be judgmental, but it means that you're not going to be judgmental to a point where you break them, because your motivation is to help them. I think this is also penetrating my management style. If you think about radical candor and these types of things, it's all encompassing.

My contribution? I'm probably an artist. That's what I do. I think I just express my art in the things I build. Sometimes you build a business, you build a company, you build a product, you build a community, you build a movement. I'm sort of a builder, an artist. I don't know.

Dude, I'm sure this show has taken some turns that you didn't expect. You thought about every question so hard.

Yeah. You know what? It was because of the review cycle in my head after a bad-question analysis.

Thank you so much for being so open. This has been fantastic.

Thank you very much for having me. This was really fun and thought-provoking.