Wix创始人：Base44能赢得“氛围编程”战争吗？｜氛围编程经济学的真相

You're not going to vibe-code Shopify, no matter how good you are. Today, we're trading on other companies' news. We're not trading on Wix news. I really don't care. We make about $400 million a year in free cash flow. So we make cash, and we invest about $200 million and something into Base44.

We all give too much credit to AI and what it can do. If I were in university studying today, I would be terrified. If I'm here, I need to be fully committed to what I'm doing because I've chosen to be here.

This is 20VC with me, Harry Stebbings. Now, it's a tough time for some public company CEOs with the SaaSpocalypse. That is more true for this CEO than almost any other. The Wix market cap is about $2.8 billion. They are doing 2.1 billion in revenue. They have Base44, which is their vibe coding product, which is now well over 150 million in ARR, growing incredibly quickly. And quite frankly, the public markets just won't give them a break. Whether it's justified or not, it's been pretty brutal. The buyback has been about as successful as the Titanic voyage, and poor Avishai, their CEO, is going through it. I wanted to sit down with him and hear how he thought about how Base44 makes such a big part of their business today and what that looks like moving forwards. How he reflects on the buyback. What public market investors don't see that they should see. This and so much more in what I think is probably the most frank and honest conversation Avishai has had in recent years about the Wix business.

Avishai, it is so good to have you on the show, dude. I've wanted to make this one happen for a while. Thank you so much for joining me.

My pleasure. I'm so happy to be here. I've watched a lot of episodes and, as I said, I learned a lot, so I'm very excited to be in the spotlight.

Dude, I've just got to start with a weird thing, but I like to bring people to personality and life a little bit. I spoke to Eran on Monday before the show, and he told me that you stay up insanely late. Is that true? What does your sleep schedule look like, just before we dive in?

I usually go to sleep at around 5:00 or 6:00 AM, and I wake up around 11:00 or 11:30. It's been my routine pretty much since I can remember. I think there's something good about it for a CEO, surprisingly, because I have a few hours where I can actually think, work on things, plan things, and do things, as opposed to always being in a meeting where I have to fight to get a clear schedule.

I have probably 4 hours where I can do real work and not just talk to people.

Do you mind what time people come in in the morning? If you're sleeping until 11:30, you're going to be in the office at 12:30 or 1:00. Do you mind if people have the same schedule?

Actually, I don't mind. But most people don't like that schedule, right? Most people will not agree to it or volunteer for it. It also created something else that I like: when you're a smaller company, people want the CEO to solve their problems. They're like, "How do we mediate between arguments?" But by the time I arrived, most of the things would be solved by themselves.

The things that were left for me to handle would be the harder things to solve and between people communication skills and other things. I think that actually helps the organization be more mature.

Your WHOOP or your Oura must absolutely hate you, and your sleep score must be shit, respectively.

Actually, Oura did, at some point, add a sleep profile. They had this theory, which is, I would say, very Northern California: you have to go to sleep at 9:00 PM or something, and that was the beginning. Then, at some point, they added, "Well, that's an early sleeper."

They had a normal sleeper, which I think is around 12:00 AM. Then they added very late sleepers. They said there are different profiles. Apparently, by the way, those are genetic profiles, so it's very hard to move between them.

Well, thank God. Otherwise, your sleep score would have been in the basement.

What time do you go to sleep?

I go to sleep at about 2:00 AM, so I'm pretty late, but I'm not quite as late as you.

Oura would probably now classify you as a late sleeper.

I don't use it, dude. I don't need another fucking device to tell me that I'm shit. Life is hard enough. I've got a girlfriend to tell me that I'm not doing well enough. I don't need a fucking Oura.

1. The SaaS Valuation Reset

But I want to be direct. You have a market cap today of about $2.8 billion, and you do $2.1 billion in ARR, give or take. What does the market not see, Avishai, that they should see?

I think that's a thing now with all SaaS companies, right? The multiples went to shit, and I think a lot of it is because the market doesn't know how to calculate the danger for SaaS companies with AI. That's the reason. In some cases, I agree. In some cases, I disagree. But overall, I think that's what we're experiencing now. Time will prove who's right, but we are working hard.

Yes and no. A lot of the companies from the SaaSpocalypse have come back, and some of them haven't.

What did come back all the way up?

Atlassian has come up again. CrowdStrike never went down. Palo Alto Networks is up, obviously. Box is up.

Salesforce is down. ServiceNow is down. Figma would be down. So there are a lot that, at their core, I think are still down. Atlassian, if it went up, is actually the one I'm more concerned about, and I don't have a lot of shares, so I didn't notice that it went up.

Well, there you go. I've provided you with some good news for today. I'm a bringer of good news.

2. Base44 Changes The Equation

I had Maor on the show from Base44. Base44 was doing $150 million in revenue, last announced. I'm sure it's more now. It kind of ascribes zero value to the core business. Is that just completely nuts to you, or are you like, "Yeah, I get that. In a new world where technology creation is easier than ever, I understand that"? How do you think about that?

First of all, I think it ascribes less than a zero value to the current business, because if you look at peers that are doing vibe coding, the multiple is much higher than the one we would get for Base44 to be at $2 billion, right? So Base44 on its own should probably be at $8 billion.

I'm not trying to guess the stock market logic. When you follow companies for a long time, you understand that the market can get them completely wrong or completely right. There's always a mix, and it's a cycle. This cycle, I understand where it's coming from. I disagree with it, and I disagree with it strongly on a few companies, less on other companies.

For example, Salesforce. What is the biggest value Salesforce provides? It's not the CRM. It's the fact that you trust a SaaS provider with your data. You have huge banks and huge companies that actually allow Salesforce to hold all of their customer data because they trust Salesforce. You're not going to vibe-code that because the trust is massive, right? It's a massive part of the value that Salesforce gives.

And yes, they have the marketplace, which has a lot of functionality and other things, but the CRM itself is a very small part of the value.

In my mind, I don't see how you compete with that with some kind of a Claude Code project.

If Salesforce becomes a valueless database that agents are able to scrape and then provide in different interfaces, then you basically become a commodity piping provider, which would mean that you could almost just replace it with—

What other platform will JPMorgan trust for their customers' data outside of the JPMorgan closed-garden DMZ secure environment? None. The answer is none. They don't trust anybody else with that, but they do trust Salesforce. That's huge. It's not a small thing. It's huge. It means that it's extremely hard to change that.

I get that, and actually, I was thinking of the government contracts that Salesforce has and has won. Again, I would say, I trust your point, but is that a one-of-one? If we go to the other providers, the other people who've been hurt by the SaaSpocalypse, they're not in that same gilded trust.

Oh, I agree. I agree. It depends which one.

By the way, Atlassian, for me, is actually the one where—again, I'm a shareholder, so I want to be clear, okay? I love the company. I've been a shareholder for a very long time. When I look at Atlassian, I think that its customers are the ones who are more likely to build things with Claude Code because they are developers. I'm always a bit concerned there because I'm thinking somebody would really replace the management stack of the tickets and workflow. Those would be those guys, okay? And yet, you said this is the one that recovered, so obviously I'm missing something, right?

For Wix, look, most of our customers will be the pizza store next to your house, your hairdresser, or a lot of small businesses, okay? They're not going to use Claude Code to build their business stack. They're not going to do that for all of what they need. It's just not in the realm of possibility now. Maybe in a few years AI will be able to do that, but we are very far from it, right?

We tried it at Wix. One of the first things we did with Base44 was ask, “Should we integrate the business logic of our vertical offering into Base44?” We ran a test and said, “Can we build the business logic to manage a hairdresser business in Base44?” Professional developers tried to do it for a week. They ended up doing a little bit, so we said, “Okay, we're going to bring in a stronger team,” specifically the team that's already at Wix. Two weeks later, they still hadn't achieved that. Those things are complicated.

I don't think the guy who has the pizza place next to your house is going to use Claude Code to build his delivery, scheduling, and staff management systems, and then build a website to do that. Some will, okay? Some will, and we see it. Claude Code is much harder than Base44. Base44 gives you a lot out of the box, so you can do way more with it. But we see that some do, and I don't think the majority will. So I think this is where Wix is, right? These are our customers. To assume that they're going to one day just say, “Well, you know what? I'm going to start building all of those things,” maybe, but we don't see it yet.

I totally get that, but then the counter would be, okay, so it's just a TAM reduction. Some will, and maybe that's 20%, which means the TAM that was 100% is now 80%, so it's just a smaller market. Is that fair?

Well, yeah. Maybe they are, but then where are they going to move to? Vibe coding.

Yeah, move to Lovable, Replit, or Base44.

We own Base44. They might move from here to here. That's fine.

If you say, “Hey, the hairdresser's not going to do it, the construction worker's not going to do it,” I get you. Do you not think they will in tomorrow's model or the model of the year after? I had Sam Altman on the show, and he said you have to build for the model of tomorrow or next year.

We are. That's Base44. That's why we invest so much into Base44. If that happens in that way, that's why we have Base44.

Essentially, the way I look at it today is that there are 3 different options. One, the guy who has a pizza place and wants to build the website and the management tools for it is going to stay a Wix customer because he doesn't want to put the effort into that. His business in life is making pizza. That's option number 1. A lot of them will stay here.

Option number 3 is that they're going to say, “Well, I'm actually really interested in that. Base44 has made it so easy for me to vibe-code everything I need. It's custom-made for me, so I'm going to use Base44 for that.” There is, of course, the middle ground, where some people will be here and some people will be there, which I'm guessing is the more likely scenario to happen in the next 5 or 6 years. Not everybody will be on this side or on that side, which is why we have both options.

Can I ask you, when you think about Base44's margin profile, a lot of the providers—we've had Lovable, we've had Replit, we've had Maor on the show—and their margin is hidden in a lot of cases. It leads to negativity and skepticism around whether they're actually good businesses. How does the margin profile of Base44 compare to the margin profile of Wix?

Wix is way more profitable. The cost of generating a Wix website is pretty much zero in comparison, and hosting is pretty much zero in comparison. Wix is way more profitable than Base44. The margins are much better, and the retention is also better because when you start building something, you need to put a lot of effort into finishing it. It's way more challenging.

But we have gotten the cost to go down. We're continuously seeing a decline in the cost at Base44. There's a brilliant team leading that, and we just released our own AI model. We fine-tune a few models, combine them, and use that. We're actually achieving performance equivalent to the top-tier models, so the cost is dramatically lower. I think that in 2 years or less, this is not going to be an important question.

Can you talk to me about the decision to build your own model? It's a big technological decision to choose to do that versus using open-source, off-the-shelf, or frontier models. Can you talk to me about the decision to do that and why you thought it was strategically right?

Yeah, of course. I think there are 2 parts to that. Number 1 is quality. We have a huge amount of training data that we're getting on Base44. When we see what users are trying to do, we see where it fails. We understand better and better what they're trying to say when they say it with the wrong prompt.

By learning all those things, we think we can actually create a model that is better for Base44 than the frontier models, which are way more generic. We don't need the knowledge that a lot of them have about Chinese poetry, because they do. They've been trained on that. But we do need to focus and increase the knowledge they have about what you mean when you say, “I want to create a task manager for me to manage my husband,” or “to tell my boyfriend where he's wrong.” We have that, right? All of those things.

We have a huge amount of data that we can use to create a model that is better and better. It's not trying to replace the frontier models; it's trying to make us better at Base44. It's a different target, which is why it's easier than building a Feibel or Mittus, right? We're not trying to build Mittus.

The other thing is that we already did it on Wix. On Wix, when you generate a website today, we're running a model that we trained. It's faster, it's cheaper, but it's also better and makes fewer errors. We keep improving it because we have a continuous improvement loop, a feedback loop. We see when customers like something, so every week we're trying to run it again and train it again to get to a place where it's better.

How much cheaper is your own model versus using a frontier model?

It depends, but I would say between 1% and 30%.

Wow. That's not nearly as much as I thought it was. That might be surprising, but Chamath came out and said that when he uses open source, it's 14 to 16 times cheaper.

Yeah, but he's using it on a smaller task. Yes, that's absolutely true. By the way, if you're using open-source models and you can use the really basic ones for a lot of different things, they work great, and it would cost you even less. But when you want to use a complicated model—for example, what you need for something like Base44—you're not going to get to that much of a level. But you can achieve 5% to 10%.

Do you think we're in the cost-optimization period for Base44 yet, given you have Replit doing multi-agent and multi-model modes? They're not cost-optimizing at all, but they are delivering a good customer experience. We're not in the cost-optimization stage for these businesses yet, are we?

No, I think it's more about quality. If you let me have 20% more quality and 20% less cost, I'll definitely go for the quality.

It's a new market. It's just happening. It's starting. We should make it better.

Why do you not publish Base44's numbers? They're amazing. Would it not be helpful to tell the story?

Well, we do publish them every once in a while. We publish the ARRs. Maor told you, and we do publish them every once in a while. So we do publish them. I mean, none of the players in this market are actually publishing on a weekly or monthly basis. I know of them publishing only when they raise money. We do it quite a bit more often.

There's no need for us to start breaking down the numbers because, with Wix, when you do the quarterly report, we tend to do things like move customers between one platform and the other, right? We're pushing customers in between.

Do you encourage customers to try each platform?

Yes. But more than that, we encourage our partners and agencies to use Base44 now. We do that. We actually spend time on it.

And you do that because you think it's the future? You have a worse margin profile on that. It would make sense financially to push them to the core business.

Yeah, I think that it's part of the future. I want to be clear here: I think the future is what I said. It's going to be a combination, and I think it's a bit more complicated than that. If you ever tried to use any vibe-coding tools and build anything very complicated, it's hard. When you're trying to do visual editing, it's really annoying.

So I think the future is a combination of both, and we want our agencies to have all the tools they need for the future. I'm 100% sure that Base44 would be part of the future. But I'm also almost as confident, if not more confident, that things like Wix will also be part of the future. You're not going to vibe-code Shopify, no matter how good you are. The business stack is too hard.

What happens then? Do stock markets just appreciate that trust is important, people want Salesforce, and Wix's market cap rises again? I'm genuinely being naive here. Help me out.

I have no idea. I'm not a stock trader, all right? This is the truth. I do that for fun. As I said, I don't even make sense because I own a lot of Atlassian, even though I'm like, "Oh my God, all the developers will replace it with their own thing," which obviously doesn't happen, right? So I was wrong on both sides, except for being lazy and not selling my Atlassian stock, yes.

Are you able to detach your happiness from the market cap or the stock price?

Yes, I really don't care. I wake up in the morning and have 1 thing that I need to do: focus on my business. I can control some of it, not all of it, and I focus on that. Today, we're trading on other companies' news. We're not trading on Wix news. We're trading mostly on what OpenAI, Anthropic, or Google are saying. The reality is that I cannot influence that. I cannot influence what people are thinking.

To be honest, I care less about that. I really don't wake up in the morning and say, "Oh my God, it went 20% up or 20% down." Even when we went up, I was more concerned about the issues that we still have to solve, which is the same thing I do today.

I completely understand that, and I think that's the right and very healthy mindset, one that very few CEOs are able to have. I'm thrilled for you that you're able to do that. Eran from Monday said he was not able to do that.

He has more white hair. You see that?

Yeah, slightly. He's got great surfer hair, I have to admit.

3. Acquisitions Become Execution Tests

Where it does become an actual tangible challenge is in M&A. One of the joys of being public is that you have a tradable currency that you can buy companies with, and when you are priced lower, that currency is less valuable for buying other assets, like the next Base44. With the price today, does it make M&A impossible?

Well, we still have cash, and we can get more cash if we need it. I think our issue today is the M&A of Base44, right? We bought the company with 1 person. I want to be clear: we put a lot of people from Wix into it, right? So that's actually complicated M&A because you have to build a company around it. We didn't get a company.

Luckily, Maor is a brilliant person and an amazing leader. Essentially, I don't know how to do another one of those at the same time and have the same quality when we merge the operations together. So it's very hard. We're about to launch another product next month, right? That's a different product that we built.

I think that our challenge today is actually execution, not finance. If we want to do an M&A, it will fall on the execution challenge and not on the finance.

You said 1-person company there. It did make news when it happened. It was a great story: 1 person, $80 million. Since then, it's been an unbelievable acquisition. Obviously, it's now probably double the revenue you bought it for in terms of price—it's $160 million, give or take.

More than that.

Yeah.

Yes, more.

Wild. Talk to me about bringing that into a board meeting or an executive meeting and being like, "Hey, guys, I found this 1 guy, Maor. He's great. It's a 1-person company. We should buy it for $80 million." What did that decision-making look like, and how did the discussion go?

To be fair, it was very easy. Actually, I never thought about it in the way that you described it, so you're right. Maybe it was surprising that it was that easy. We have a really good board, really smart people. We explained the logic behind it, and they were like, "Yeah, it makes sense."

They asked us a lot of really smart questions, but none of them asked, "Why are you willing to pay so much for 1 person?" That was not part of the questions. They actually focused on the business: the business logic, what we're going to do with internal marketing, and how we're going to build a team around it. We got a lot of questions in regard to that, also about why we thought it was better than Replit or Lovable. So we explained it, and until now I never thought about it like that.

Can I ask you—you said that, "Well, we have a lot of cash." How much cash do you have? I should know this, and it's public, but—

It's complicated. We have debt here, a loan there. We make about $400 million a year in free cash flow. So we make cash, and we invest about $200-something million into Base44, or a bit more than that. Overall, I think we're okay.

If I were sitting with you and we were in a pub in the UK, I'm not sure if you have pubs in Israel—you probably have bars. If we were sitting in a pub, I'd say to you, "Avishai, dude, let's take this private. What a mispriced asset." Would you like to take it private?

I think it would be pretty much illegal for me to even answer that, right? I'm focusing on the company and building the company. We're doing incredible things. We have super-smart talent. It's an exciting journey, so I love that, and we're going to continue to do that.

I'm not buying the company, I can tell you. I think the stock market should be a place where smart people make money, and if they think like you do, that it's a mispriced asset, then that's a great opportunity. If they think that AI is going to kill everything, then obviously that's a terrible opportunity. Everybody has to make their own decisions and plans.

Totally get you. In terms of how people make money from the stock market, you did do the buyback. How did you think about the buyback, and what are your thoughts and reflections on that?

Well, obviously, we had terrible timing.

Dude, I didn't say it.

But it's very hard to time the stock market. Very hard to time the stock market. I think we had too much float, and we had a lot of money in the bank that was sitting there. We knew that we weren't going to do a major acquisition because we were too focused on the new product coming and on Base44.

We couldn't do both things. We couldn't do another acquisition, so we had $1.5 billion and said, "Okay, it doesn't make sense for us just to hold it." Then we said, "Okay, let's use that opportunity, where the stock is very low, to buy it." Obviously, the stock market jumped like this when we did it, and then went crashing down. All right, that's part of life.

I think the big question isn't what happened in the last 3 months; it's where it's going to be in 3 years. I hope that by that time, it will have been proven to be a smart decision.

4. The Buyback Disaster

I mean, if you want real bulls, then you buy back more now. Is that conversation like, "No, no, no, no, no, we had that discussion"? How do you reflect on future buybacks given what you know now?

I think buybacks are a fantastic tool for a company, and companies should use them. I think they don't use them enough. The concept that companies always issue more and more shares, decreasing their dilution by issuing more shares to employees and for other things, is something that should be balanced by buyback when the company can.

I'm really in favor of that. I think it's another way of giving a dividend to all of your stockholders. That's my opinion. Again, I'm not a professional stock market person.

You know, that’s not my profession. I’m a geek, an engineer, a bit of a marketeer, and a product guy.

I’m worried about the level of stock-based compensation that we’re seeing. It seems like we’re normalizing insane levels, a la Snapchat, that make it very difficult for me as an investor to see the same attractiveness in terms of returns that I had before. How do you think about SBC being good to employees but also not screwing investors?

Well, I don’t think you have a choice, right? I think that you want to get the best talent, and you want to keep and maintain the best talent. Part of it is allowing them to be invested in the dream, because they built the company. You should give stock to employees. That’s my opinion.

The fact that it’s only in technology companies and not in many other companies is a different issue, right? I would actually love to see every fashion company give the whole team stock options, because I think that’s way more fair. Maybe if the U.S. government would do that, or the British government, we’d actually have better people running the country and somehow finding that their compensation is related to their achievements. I think it’s a really good mechanism.

When companies start making cash, when you collect cash, you essentially have a few options for what to do with it. Number one is acquisitions, but you can only do so many acquisitions and maintain the company. Some people master that, and I look at them and admire them. I wouldn’t be able to do that, but the majority of companies didn’t, and the majority of acquisitions fail.

I think that’s a path that some companies should pursue. Every company probably should pursue it a little bit. The other thing you can do with your money is a buyback, which is essentially the same as giving dividends, and then you balance your stock-based compensation. The third thing you could do with your money is invest in hiring more engineers, maybe, but then your EBITDA gets destroyed, so the stock market hates you.

At this point, though, do you give a shit if the stock market hates you? One of my own neighbors is a public company CEO, and he’s one of the companies that have been hit. We haven’t mentioned them, but he’s like, “Honestly, dude, at this point, it can’t do any worse, can it?”

Well, I’m sure you can, but it’s hard, okay? Again, I have a huge amount of respect for everybody who put money in Wix and bought stocks, and I know and understand that my job is to do my best to create a wonderful company. I have a bigger obligation to my customers. That’s another big obligation that I have.

I don’t think they’re one against the other. I think I should balance both, and, too, for my employees. There’s no stock market you can talk to, right? It’s millions of people with different opinions. I don’t know how to take that into consideration. I don’t even know how to listen to that.

5. Retaining Talent Through Turbulence

I’m friends with many of these enterprise SaaS company CEOs. We’ve mentioned some of them. When I chat to them, all of them are like, “Oh my God, my talent is leaving.” Their stock isn’t worth what it was, and morale is down. How do you advise them on how to retain talent when you’re in an environment like yours?

I think it’s a huge challenge. First of all, I don’t think there’s a magical solution. There’s no magic here. Everybody is going to lose talent, and we are losing talent. Everybody has lost talent; everybody will lose talent. I think a lot of times it’s a healthy thing for a company.

As long as you don’t compromise on who you keep, and you make sure that the people you allow to stay are really great people, you’re going to find that you have a lot of talent that you never noticed before. That’s going to refresh a lot of the organization. It’s not about how you prevent people from leaving. It’s more about how you make sure the organization keeps being an organization where the vast majority are top talent.

A lot of us got very spoiled by saying, “We have people here working for 15 years who are always top talent, and nobody ever leaves because our stock is at whatever.” That kind of made us spoiled to the fact that that’s not a natural thing.

I highly appreciate that there are a lot of super-smart people around me who I would be extremely sad and worried about if they left. But, saying that, I still think that we got a bit spoiled, and we need to remember that when we got those people, they didn’t have 15 years of experience working on our problems. They became super, and the reason they’re here and they’re super is because they had talent to begin with.

We need to find a new generation and bring a lot of those people in. I think that’s the challenge for me. Having said all of that, just to be clear, I’m working extremely hard to keep the top talent in the company. But that’s life.

How many people are in Wix today, and how many are in Base44?

Base44 is around 400 people.

Wow.

Wix in total is about 3,500. You have to remember, we have a lot of small businesses. We need to support them.

3,500.

Yes.

Well, hello, Jeff Bezos. What are 3,500 people? Where’s the majority of them? In customer support?

Customer support is our biggest department.

Wow. Are you using AI? That must get taken to, like, zero pretty soon.

For real, I don’t think it will ever be zero.

Not quite zero, but maybe 10?

What, 10 customer support representatives, you mean?

Yeah.

We’re offering our services in 192 countries. It’s a bit tricky to have 10 people do that. We just have to remember that it’s a different scale of operation when it comes to that.

Did you build your own customer support AI, or do you use off-the-shelf providers?

We built our own. It doesn’t work.

Avishai, this is what I don’t get. You see all these hyped customer support companies, and I’m going to get in trouble here, but I’m too old to give a shit, so who cares? Everyone I speak to, whether it’s Airwallex, Klarna, Navan, or Wix, builds their own.

We actually tried not to build our own. We tried and tested many different things. It doesn’t work.

I don’t get this market.

To be fair, we’re looking at young companies that built AI agents. They prove something very quickly, then hire a bunch of salespeople and try to sell it. Then they go to one of those companies you mentioned, where they have super-smart engineers who are really focused on that. They have more engineers supporting their support than all of the HR on those companies that are trying to sell them, and better engineers a lot of times. So, it’s not that simple. It will take time. I think in 5 years it’s going to be different.

Do you? It’s going to be different in 5 years? Again, dude, I have no idea. I’m a fricking podcaster in London.

Maybe. I’m not sure. To be fair, I don’t know. But we tried. It didn’t work. We tried again. It didn’t work. We tried again. It didn’t work. We kept working on our solution on the side. At some point, it was like, “Okay, that’s enough.”

6. AI Changes Small Business

In 3 years’ time, how big will the Wix team be then?

It’s very hard to say. I would love to say, “The Wix team in 3 years will be 1,000 people in Base44 and 1,000 people in Wix, and everyone else will be AI agents.” I’m not so sure it’s possible. We all give too much credit to AI and what it can do, and then when you start looking at the details, it’s not really doing that well for a lot of things.

It’s amazing. It’s incredible. It’s super powerful. But the “but” is pretty big.

You need to be more American, okay?

All right, all right. AI is the next big thing.

Like you believe it?

You know it. It’s going to change everything. Everything’s going to happen and blah blah, whatever. But I’m going to be by myself in the office, and most of the time we’re going to play PlayStation. All right?

Now you’re like Sam Altman, dude. You sell the dream. It’s much easier in life to be Walt Disney—“Tell me a story”—than it is Jerry Maguire—“Show me the money.” The trouble with Israeli entrepreneurs, I find, is you’re all like, “Show me the money.” You’re generally not that excitable. You’re very realistic and very direct.

I could tell you what I dream about here, but it’s not about AI replacing my customer support, okay? That’s a nice thing to have. I think that, with AI and technology, small businesses will be able to achieve so much more than they do today. They’ll be more efficient, grow faster, and be way more profitable. I’m sure that will happen.

If you run a gym or you’re a personal trainer, your ability to use AI to manage your business better, sell additional products, and reach out to new customers will be massively enhanced—not in 10 years, not in 5 years.

Next year. And that will really have a massive effect on how you do things as a small business.

But dude, is that not a little bit paradoxical? You're like, “It'll have a massive effect next year.” Great, I'm a venture capitalist. That means massive revenue growth and my companies grow fast. Cool. But you just said AI's not that great, and it doesn't actually do that much, and the questionability of the value is there.

No, no. I actually said it's really doing a lot of amazing things, but the “but” is very big. We all tend to… How many times do you go to ChatGPT and say, “I want to know what this thing in medicine is,” or, “I want to know what that is,” or, “Tell me what the chance is?” I can tell you, I needed to test something, and I wanted ChatGPT to write me the safety protocol.

I actually use Claude, not ChatGPT. Sorry, Sam. I got 6 gates, okay? And he said, “Those are the gates. You have to make sure to test every one of them.” So I started to ask, “Why this one?” We started to analyze it, and we worked on it. From 6, it became 1. All of the other ones said, “Oh, okay. Good question. You know what? You really caught me here. I might have overstepped.” Then he emphasized that. He writes, “Well, I'm removing it from something you must test to something that you might need in the future to take a look at.”

And we all tend to do that, right? We all tend to have this massive trust in AI models. I think that's one of the things where the “but” is. When you take an AI model and make it work really well for something, it can be incredible.

I'll give you an example. The most basic thing we all take for granted now: you can actually have a conversation with an AI model, right? You can actually have it talk to you on the phone, talk to you on your computer. We couldn't do that 5 years ago, 4 years ago. I actually do that. I do that when I drive. If I have a thing I want to do, I click on it, and I love the way it works in ChatGPT.

There is a big difference between taking AI and trying to do customer support for a very complicated company or helping you run your customers' orders and your schedule. If you build the right skill, the right tools for it, then it can become super efficient. It can also call your customer and talk to them and do things. So yes, you can have that massive extra layer.

Do you find that SMBs are excited or terrified by AI, both in the product and when you speak to them? There's a lot of skepticism. It's not always welcomed.

Well, I didn't see that with SMBs, actually. I think we see it more with everybody else. SMBs, when they wake up in the morning, have so many issues running their business that they really want something to help them run their business better.

As opposed to if I were in university studying today, I would be terrified of AI. That's the truth of AI, because why am I learning this? Will it even exist in a year?

Do you want to hear something wild? To change a university curriculum—in the UK, this is, so forgive me; I'm sure it's pretty similar around the world—it takes 3 years to change the material that goes through a yearly program. 3 years to change it. Imagine the changes that have happened in Claude. If you're doing computer science, it's almost pre-ChatGPT.

I think that, actually, I don't know how long it takes here, but it takes time, and it's challenging, I think.

So if you were genuinely advising a 22-year-old—imagine a 20-year-old son of yours. I know you've got a newborn, so just skip a few years—what would you say to him, genuinely?

Well, I would not be that concerned with LLMs replacing every kind of white-collar job. That's not going to happen. It's not that smart. It keeps making silly mistakes. It's not very good. Actually, it's not very good at reasoning. It's very good at collecting data and reframing it.

I would be less afraid of that. However, if we get another 2 or 3 jumps in algorithms, then we all become a bit like monkeys, right?

The core question for me, actually, is that we're an investor in Agora, which is an AI legal service, and the core question for me is, do we see the same rate of improvement in adjacent functions to coding that we saw in coding? Andrej Karpathy went from doing 20% of his coding work with coding tools to 80% in 6 months. Do we see that same jump in the amount of legal work done by AI?

I think it depends on a few things. Law might be a very good example for it, but I do believe that in law, when it comes to laws, because laws, a bit like code, are mostly about textual context. There's a tremendous amount, a huge amount, an endless amount—way too much—of legal written documents and cases. So I think this is one of the perfect places for it to be useful.

But every time I see a doctor trying to do research by using LLMs, I get horrified, because that's the wrong tool. It will probably answer you with what it read somewhere on Reddit, where somebody wrote something. From the LLM's perspective, it's equivalent to the best research done in Nature, right? There's no difference. It holds them all in this giant thing that it has. It'll try to verify it somewhat. Most of the time, it doesn't read the research when it actually answers you, even if you use the highest level.

For legal stuff, I think that's really good. For a lot of other things, I truly think that it's not the right algorithm.

Do you buy that for health? I'm just being honest here. I think OpenAI and ChatGPT are a lot better than the majority of doctors at this point, actually. And on top of that, health is now the second-biggest usage of ChatGPT.

Well, I'm invested in a company called K Health. I don't know if you're familiar with it.

Yeah, yeah, yeah. Allon.

Yeah, Allon.

Yeah.

And they've proven that they diagnose better than most doctors. All right? But that's not the same as doing research, right?

No, it's very different from doing research.

For that, absolutely. Because most of the time you go to a doctor, they diagnose you with the flu and tell you to go home. They give you some pill and tell you, “Take the pill, and you'll get healthy in 5 days.” And you would get healthy without a pill in 5 days. So nobody knows, you know?

When it comes to specific cases, I think it makes sense that AI, when it comes to understanding the natural scenarios that you go through, can help. Then, yes, absolutely. The legal system will have a lot of those. But there's no research in legal. Well, what is research? But you do have research in medicine. I think when it comes to research, it's not a good tool. So it's a bit different.

Do you think frontier models will really only be used for a very small number of use cases, like science research, molecular biology, astrophysics—very, very difficult, complex stuff?

They cannot do it. AI cannot do those things. You cannot do research on astrophysics with an LLM. They don't have the capacity for that. It's not what they do. They might answer you, but they don't run the models, the simulations, in the background. If you want to have the simulation in the background, you have to move to something like DeepMind. They do some of that, but not ChatGPT.

Fine, but model providers—you don't think they'll do that in 3 years' time?

Maybe, but then they have to connect different kinds of models and different kinds of architectures to their LLMs. It's not that trivial.

I don't think it is, but I do think your Isomorphic Labs, your Latent Labs, and your Chai Discovery, who specialize in very, very complex reasoning around scientific—

Oh, yeah, I agree, but that's probably not just an LLM, unless they just write—

No, I totally agree with that. Yes, 100%.

So those are completely different models. Yeah, absolutely. Yes, you can. Obviously, you did protein folding, right? DeepMind did that. An amazing achievement, of course.

7. Resilience Beyond The Market

You come across as quite chilled, in a nice way—composed, calm. Have you always been like that?

No, I'm half Italian. From the south of Italy. I know.

Does money make you more chilled?

Well, I think that a Slavic wife—

What makes you more chilled?

Yeah. They make you more resilient.

What does it take to have a successful marriage?

I'll tell you a story about my wife, which is one of the things that helped me understand that this is something different. I came home, and she looked at me and was like, “Listen, I'm in a terrible mood. You did nothing wrong. I'm going to the room to relax. You go play PlayStation.”

And I was like, “Wait, what? You're in a terrible mood, and it's not my fault?” And she's like, “Yeah, you did nothing. It's not you. I need to relax a bit. You go play PlayStation.” I was like, “Oh, my God.”

So I think that's the number-one quality I learned, and for me, it works, right?

Sorry, what's the number-one quality? Distance?

No, no, no. It's the fact that she's able to know that you don't load your emotions onto somebody else.

And for some men, the number one thing you have to learn is to say, “Yes, I agree. Whatever. It doesn’t matter what happened. Yes, you’re right.”

How did making money change you as a leader, though? I have a theory, and the reason for this is I think richer investors and richer founders make better decisions. You’re not operating from fear. You’re not operating from losing homes, losing school fees, losing core pillars of life. I actually make much more rational decisions now that I’m not doing it wholly for the money.

I tend to agree. I think that, overall, when you’re less afraid of things, you’re more reasonable, but you’re also a bit more lazy, right?

I think a really interesting question is: What does no one tell you about making money that you wish someone had told you?

The way I look at it, and that’s my personal opinion, is that money gives you one thing beyond everything else, and that’s freedom. Freedom from fear that you won’t have food tomorrow, and then freedom also to know that you are here because you’ve chosen to be here.

Because I can always, and you can always, say, “You know what? I want to be in Costa Rica on the beach. I want to be in Brazil, dancing at Carnival.” Or I can change completely what I’m doing, and you have the ability to do it. You don’t have to wake up every morning, go to work at a job that you don’t like, and do it just to maintain your life and have food.

That is the number one value you get from having enough money or a lot of money: that freedom and that knowledge that I’m here because I’ve chosen to be here. There’s no other reason except that. I think, for me at least, it’s freedom to understand that if I’m here, I need to be fully committed to what I’m doing because I’ve chosen to be here. Nobody made me. There’s no other reason except that I chose to be here.

The BTC of work, see the stock crashing like that, see other thing happening. You know, all those things. I’ve chosen to be in the middle of that, and I think that’s a tremendous amount of power.

What would you say to someone who wants to build resilience for a shitstorm that’s about to come? Like you said, dude: the pressure, the fucking headlines, the stock crashing, this, that—it’s just hit, hit, hit. What would you advise me, or someone listening, who wants to learn how to be resilient for the hits that are going to come?

I think the first thing to accept is that it will happen. We tend to think that if yesterday was easy and the day before that was easy, then tomorrow will be easy. That’s not how life works, right? You never know when the next storm is coming from. You never know when COVID will arrive. You never know when war will arrive. You never know when the stock market will crash. You never know, and it will happen—for sure, 100%, it will happen.

We tend to believe reality is going like this: if you’re here, it’s going to go slowly up and slowly down. But it doesn’t. It goes at an extreme speed in some places, right? It will go like this, and then boom, and then boom. This is how reality is.

The worst thing to happen is probably going to be some random thing happening on some random Wednesday. It’s not something that you get a lot of warning for. In this sense, I think AI is the exception, because it was an extreme thing that we’ve seen happening in slow motion, right? It took a few years. It didn’t happen with a boom.

The first thing is to accept that it will happen. The second thing—and this is something I remember from Babylon 5. Do you remember that TV series?

Yeah.

There was this time when they were flying to a conflict somewhere, and the captain of that ship was very nervous. The guy told her, “Why are you nervous?” She said, “Well, we’re going there, and I want to make sure.” He said, “Listen, we’re going to get there when we get there, and the weapons that we have are the weapons that we have, okay? This is the most you can do, so you just have to accept that this is how it is.”

This actually runs in my mind. It’s not very smart, not a very inspiring sentence, but for me it runs: Why am I doing the best that I can? That’s a huge question, right? I always try to make sure that my answer for that is yes.

Everything else is: What can I control? I cannot control more than what I can, so I try to do the best with what I can control. I know it’s not a deep answer, but I don’t think there is a deep answer here. The big question is: Are you doing the best that you can in the situation that you are in?

If you’re able to be honest with yourself and say, “I think I am,” then that’s a lot of the pressure going away. The other thing is that I believe in meditation. I believe in neuro-linguistic programming, NLP, and I did a lot of that. It’s really helpful.

Totally. One of the things I always think, too, is that a mentor said to me, “You’re never wrong to do the right thing, and the right thing is often the hard thing.”

I agree with that 100%. Then I’ll tell you the last thing I always remind myself: You go around and you see happy people, right? A lot of it is places. They are not necessarily correlated with how much money they have or where they are or whatever.

I always remind myself that the happiest people I’ve ever seen on the planet are in Brazil. I can always move to Brazil. Are you surprised it’s not England?

I mean, obviously, right? After 5:00 p.m., people in England become happier when they go to the pub.

Dude, you haven’t been to England in a while. We go to the pub at 3:00 p.m.

Two hours happier. Oh, yeah.

Yeah, yeah, yeah. Finland is actually the happiest country in the world, which is astonishing to me when you look at the hours of light in Finland. But there you go.

I’m guessing the person who wrote that has never been to Brazil. For me, in my experience, that’s the case. All of these things matter to some degree. At the end of the day, I can always go to Brazil.

Do you find it difficult to context-switch in your relationship? My biggest problem now in my relationship is that I come home and I’m stressed. I’ve been bashed all day by whatever’s happening, and then you come home and they want to talk about life and normal things, and you’re still in your zone. Do you struggle with that, and how do you—

I used to. It was horrible. That was very hard for me. I remember that I would come home with this massive list of things in my mind, and I would either not be patient with my partner or not really be attentive, and that was very hard for me.

I learned with time, as I got older, that I stop most of it now and then I’ll do the thinking later. That’s why I go to sleep at 5:00, right?

You time-bound your thinking, and it allows you to turn off for the time being with your wife?

Yes. I’m just doing that, and then I’m turning it back on. Sometimes I write myself things that I need to think about in 2 hours.

That’s perfection, because I cannot do that at all. Final one, and then we’ll do a quick fire. I care very much about being a great parent. I don’t have kids yet, but I very much want to. You do have kids, and you’re a public company CEO under a fuckload of pressure, let’s just be honest. What advice would you have for me on how to be the best father that I can be?

Kids don’t care as much about the amount of time that you’re with them as they care about the quality of that time. If you give yourself, let’s say, a few hours every day or every few days, make sure that you are there and happy when you’re there, and have fun with them when you’re there—not coming because you’re forcing yourself, then being on your phone and fighting with everybody.

Remember, it’s more about how you’re there than how often you’re there or how long you’re there.

It goes back to what we said about being present, not stuck in your own head.

Easy to say. I’m not sure it’s so easy to do, but yes.

Impossible to do. It’s like tying happiness to stock price. I remember interviewing Vlad from Robinhood when it was worth 20 billion, and I interviewed him when it was worth 120 billion. I can tell you, he was legitimately 6 times happier when it was worth 120 billion compared to when it was worth 20 billion.

It was a different fucking person. He looked different. He sounded different. He was different.

Oh, yeah.

Okay, we’re going to do a quick fire. What have you changed your mind on in the last 12 months?

Actually, a bit on AI. I’ll say it like that. I was very concerned—very concerned—about how quickly we were going to find humans being replaced with AI. I think I’m now looking at the same thing. I think it’s going to take much longer than I anticipated.

Why?

Because I remember that a year ago, we would hear Elon Musk speaking about how he’s waiting for AGI, because you can have a conversation with it about how to solve big problems in physics and how to do a lot of things and everything.

And now everybody’s talking, “Well, you know, we kind of achieved AGI, but it still cannot do those big questions, right?” So we kind of changed the definition to fit something that doesn’t do the critical things. The more I know about AI, the more I see the frontier, the more I see things that I think are tremendous, as I said, but I don’t think they’re going to replace us that quickly.

Yeah, I think you’re being far too nice. I had a picture of me and my mother on the beach from a holiday last week, and I said to Gemini, “Remove the people from the background.” They changed my mother’s face. They put a different woman in there. I was like, “What?”

I think solving that might happen faster than you think.

Which competitor do you most respect, and why them?

It’s not direct. I think the one company that I really, really appreciate is Figma. I think they have brilliant product understanding and a brilliant understanding of their market. When I look at it, I’m just like, I really admire some of the things they can do and the quality they can bring to that.

I really like them. In terms of not just respect, it’s almost admiration for some of the things they can do. People are almost religious about Figma, right? I know very few products that are like that. My designers love it. They go crazy about it. They’ll do their grocery lists on Figma. They’ll do everything they can on Figma just because they love using it so much. That takes massive skill from Figma to do.

What’s your single biggest concern today?

It’s always the same: execution. I think we have good plans and good ideas, and I think the biggest challenge is always execution for me. It’s how we get a team to be super-efficient and how we get them to be more ambitious. I think it’s about going for a bigger goal.

You said that Base44 is 400 people today. How many people is it in 2 years’ time?

That’s one of the things that’s very hard to predict, because it depends on how well AI does and how many models they’ve got. I’m a huge believer in Base44. I’m a huge believer in Base44. I hugely believe in what Wix does. I see the way it scales.

I would imagine probably 800 to 1,000. The majority of them would actually be engineers, right? So I think that is another thing.

Final one for you, my friend. What is the kindest thing that anyone’s ever done for you?

I think there are a lot of really kind things that people have done for me. When I started, I saw how often people who were way ahead of me in this business would take time to sit with me and talk to me about mistakes they made, give me advice, and think with me about other things.

I think that’s extremely kind, because this is not about giving you money. It’s about giving you the most precious currency those people had, which is time. When you see that, I think it’s amazing that they did that. The advice was amazing, too.

In 12 months’ time, who will have a higher market cap, Wix or Monday?

It’s kind of like asking me to compete between which kid of mine I like more.

Your eldest.

And so my eldest, Wix, right? But I’ve been lucky to be involved in a few amazing products: Wix, Monday, and Base44. I think I got this rare opportunity to see this happening 3 times, and I think that’s an amazing experience.

I do believe that all 3 companies—all 3 products—have tremendous market potential and the potential to be helpful to many different people. It’s so hard for me to choose. I think Eran and Roy are really incredible CEOs, probably better than me, so maybe I’ll bet on them—or me. No, actually. But then Maor is a great CEO as well. I don’t know. It’s very complicated.

Avishai, it’s been fantastic to have you on the show. Thank you so much for being so open and for being so wise in your love, family, and parental CEO advice. It’s been fantastic, man.

All right. Cheers, man. Thank you so much for having me. I love what you’re doing.