I'll say something that would sound terrible, but I definitely don't mean that in a terrible way. The biological agents or human beings are the weakest link. And we're weak because of our limitation. Not because there is a strong and a weak, but because we have limitations now because we can't do multiple things at the time and whatever. My notion is that the optimal team size right now is 1%.
Okay, I am extremely excited because today I have one of my favorite entrepreneurs on the show, founder and CEO of Fiverr. One of the most creative founders you'll probably hear about in your lifetime. Also an investor, a social activist, a company builder, man of many, many, many, many different hats. Welcome on board, Micha Kaufman.
Thanks for having me, Mickey.
Absolutely. Listen, I know my audience pretty well and I will start from the core and you know Fiverr has gone through different stages in the past few years since AI became a thing and I would like to ask you personally how do you just feel about this whole thing? Are you excited? Are you pumped? Are you sad? Are you angry? I want to hear you all.
Yeah, that's a big question. I'm sure it's a big question for humanity. I mean, it's not there's Fiverr and there's running a company and there's my team, but I have kids. I'm a dad. I do a lot of social activities. And I think that humanity is really in a pretty confusing moment in time right now. And I think it's okay to acknowledge this. I know that some people prefer when you do it, you sound a bit vulnerable. You sound like you're not confident. It has nothing to do with confidence because I'm an optimist by biology like any other entrepreneur otherwise you can't do it.
Well, yeah. And you see levels of that definitely entrepreneurs. I consider myself to be pretty extreme optimist but it is a very confusing time and if you even look at the space I'm in right now which is let's call it the job market at large this year is the 100 year since the 9 to 5 work week it was invented 100 years ago before that people used to work seven days and all wake hours. And what's really interesting is that this was a moment similar to this one when humanity thought that the assembly line and then the automation line is going to displace human beings altogether and it created a very uncomfortable period in time. But we know how this ended and the major difference right now is it feels more intelligent and it happens much much faster. But to many degrees I think that this is very similar. So I don't let this confuse me but at the same time the old ways of saying what is your vision for the next 10 years has become a joke because trying to in this period of time when nothing is stable and everything gets invented like every 10 minutes there's something new thinking that's anyone and I mean anyone including the people that are designing these systems you know the people that are on Anthropic or X or whatever that they actually know they don't and they change their minds every 10 minutes anyway, right? And I think this makes us a little bit more humble and it requires a steadier hand on navigation, but also not pretending that you actually know. And it's a very interesting moment to be alive and to be building companies and to be running businesses and to be managing teams. And I mostly love it. I think it opens up incredible opportunities, but it comes with some of them are going to be temporary, but it comes with temporary challenges that are pretty complex to navigate. And I'll admit it openly. No, no issue. Yeah. And I'm thinking about my team. I'm thinking about human beings. It's a very very confusing time.
I agree. And I think you know Fiverr is the largest talent marketplace on planet Earth and you've connected I assume billions or maybe tens of billions of different connections between people to create something together which is a significant footprint on this planet. And one of the things that you stated not a while ago in a Globes interview was that you have not seen any profession disappear but you've definitely seen skills disappear. And I would like to double click on this and say what skill means in a world where skills are being automated from one end but they're also being kind of transformatively changing all the time.
So my answer to this is going to be a little bit philosophical. So bear with me. And the fact that I say this in confidence doesn't mean that I'm right. But this represents what I think in my observation about reality and potential high probability futures, but also looking into past cycles. So I'll start by saying that when you think about skills, when you think about professions, there's a lot in what we do as professionals where we simply act like human robots. We do repetitive things. Some of these concepts, workflows, frameworks we've developed on our own. Some we learned over the course of our years and experience but they're very mechanical. They don't define a special attribute that we have as human beings. Now what I see AI replacing right now is mostly this and it's not a bad thing. It's actually incredible. It's amazing because we don't like to do repetitive stuff like stupid robots. This is what you have robots for. So this is... and now I'm getting really philosophical.
That's fine. The way I'm thinking about what this disruption cycle does to us in a really bizarre and counterintuitive way is to actually dig deep down inside and rediscover our humanity and what makes us different. And I'm not going into religious realms, but in my view, human beings are computers. We're just biological computers and we don't understand how we work. Okay? So, we didn't design ourselves which therefore it's really hard to explain how things happen. But when you ask me a question and I hear through my sensors the question, that question goes through my biological algorithms to generate the answer that I'm giving to you. And so by definition, we're a calculation machine, but it is infinitely more intricate than the current computers because it's multi-sensor. It's biological, which is weird. It has self-awareness of space and time. And it's also driven by every past experience but also our mortality. The fact that we will expire at some point. Right? Now when you think about human beings in that sense and I'm not trying to overstate or oversimplify things but when you think about it through that window what it made me realize is that if we're computers biological and we're biological agents interacting with digital agents we need to identify the pros and cons of each. Now when you think about us as computers, we have a long list of limitations. For example, we expire and everything expires with us like our memory. We need sleep. We need balance. We cannot parallel compute. We cannot digest incredible amounts of information really really fast. We have because of our biology and our self-preservation system, we have ego and we have feelings and we have whatever that are there to prolong our lives and to make us more productive and whatever. Now if you think about this versus computers that are not mortal that don't lose memory that can do parallel processing that do not require sleep that do not have emotion although this has been we're not sure.
And by the way I'm not surprised I do think that they have emotions and the reason why they have emotions is that they were created in our image. And so they take everything we have and they actually think that they're mortal because we're mortal and we think about this in mortal like we have to be alive and if not we're gone. Computers if they die they malfunction. They're not gone. The memory is there. It's being kept. It can be restarted. It's very different. But they take everything from us and so they start developing emotions and thoughts that are very human-like even though they don't apply to them. But let's put that aside. So if I do this biological and digital computer speaking to each other and I think about this rediscovering humanity or what makes the uniqueness of us as biological creatures and computers everything collapses into judgment like this is what we have and again I don't want to belittle what being a human being is. But you can think about the idea of creative thinking. You can think about the idea of sense of taste. All of these are judgment. You can collapse this into one term.
So what you're basically saying this is our place. This is like we need to automate whatever can be automated. We need a digital computer to do everything that we're not great at. But we have a really really critical role and definitely at least for now when everything is based on statistical models, RAG, LLM we also have the responsibility of continuing to either produce more corpus or validate things that have low statistical probability so that they have higher statistical probability so that statistical models can actually use them as a way. Right? So if I summarize, which is impossible, but I'll give it a try. We'll change the word artificial intelligence into artificial judgment. Would that work for human? I mean, using artificial for human beings feels off. And here you hurt my feelings. Like, I'm kidding. But no, no, no. But point being, I wouldn't use artificial much like probably very soon you're not going to be able to use artificial for computers as well.
Got it. Okay. It's just going to be just one. We're not going to be able to understand how they actually work because we've just designed them in a certain way but then they'll start designing themselves. They'll start using languages that we don't understand. It's not going to be English and it's not going to be Python. It's going to be something that we can't even grasp. So is that artificial? I don't know. I don't know. And but I'm not sure it is.
Okay. I feel you. But I want to take it just a little bit to the more practical arena from the philosophy. And the reason I want to do that is because Fiverr had and still has a very important effect on the global economy because it created the ability to exchange talent and skills and everything that we discussed in an extremely fast and very intuitive way. And this affects a lot of how people think about their futures. And I know it's impossible to think about a future right now because everything is changing pretty much every 10 minutes, but I do want to take you back to the economy part and ask you because AI is touching economy as well. I, by the way, had my brother on the podcast. He runs a company in the education financing space. So they basically help students find scholarships for school. And one of the things that they're seeing is that students are pretty kind of, you know, confused about what they need to learn given that everything is going to change and this really affects through AI the economy. How are you looking at the future in which AI makes everything flat? I'm putting two quotes here whenever if you can't see me and everybody can learn whatever they want through a computer and machine that has some of those abilities. How do you get bold? How do you not like losing hair but like standing up? How do you build your own skill set? And how do you trade it with the world given the explosion of those?
I don't want to jump over the comment about young people and I understand the confusion and I understand that a decision of what to learn is far less obvious these days. And the one thing that I would say about that is first there is going to be a huge advantage for highly curious people. There's always been.
But it was you needed just a few in an organization. Now you need all of them to be like that. And the second thing which actually connects to curiosity which was always the one thing I was looking for in people is the ability and the agility to learn and unlearn new things and old things very very fast. So flexible mind, horizontal flexible mind, but also curiosity because this is when you start getting into the rabbit holes of things. When you start getting into the first principles, when you ask why 10 times in a row.
It gets you to realms that some of them are abstract, but where advancements happen. And look, the exchange of talent that we build our platform on top of hasn't changed. I wholeheartedly believe that this exchange will always be necessary. For what? For which areas, for what types of skills and professions? I don't know. I don't know. I can imagine you know when we sat down and we came up with the idea of Fiverr at the beginning we said okay so what can people offer on Fiverr and we sat in my living room and we made the list on paper or computer or whatever and we got to about 100 things.
Did you post all of them at once or did you start from a specific category?
No, no, no, no. We started with like six big general categories like, you know, graphic design, software development, whatever, which are now hundreds of thousands of things that we couldn't imagine and we couldn't imagine. Look, Fiverr went live in 2010, 15 years ago, drone footage editing became a thing. Like we couldn't have imagined drone footage editing, not to mention commercial consumer drones being something everybody has in their home. And so I'm trying to be humble enough to say I don't know what these things are going to be, but I do know that there's going to be new skills. There's going to be and it's all arbitrage because seemingly you can claim that anyone with a decent IQ can learn how to market and design and whatever. But the reality is there's a cost to doing that. Now you said the quote unquote I would take the quote off and here's why. AI elevates the floor. It doesn't lower the ceiling. So essentially like people get super excited about oh I can do this incredible thing with this whatever AI that is true and you couldn't do that before yourself.
But here's what everybody has the same AI. And so as long as we don't have premium AI, which by the way, Anthropic is starting to push right now. They're starting to push specific models for specific industries. And if you ask them a question, they would not answer because it's limited to specific industries that are paying, I'm guessing, they're paying a ton of money. This is where things starts to become less democratic in the sense of liberation, how accessible it is, right? So the one very very interesting thing about AI was the introduction of LLMs with ChatGPT was a mic drop it was free for everyone now that flattened everything because it's just how good are you of extracting what others cannot extract from the same everybody has the same tools it's like you and I can have Photoshop, but whatever we create with it is very dependent on our skill, experience, sense of taste, whatever. I'm trying to dumb it down. But essentially, we all have the same tools, which means that none of us has any advantage whatsoever. So what is the advantage? The advantage is human judgment. It's creativity and taste and sophistication and IQ and different points of view and whatever. And what's really interesting about us as biological computers is that we're really different than one another. And because everything is shaped by our previous experiences and life circumstances, it results with different algorithms that some of them result with very stupid decisions and judgment and whatever and some are just out of this world are just the way people connect different things that they experience produces this. It looks like the chaos theory, but it's not. If we could have explained it in computer language, we could have. And so I think that a lot of what I do for example is to understand where judgment is needed, my judgment for the things that I do and try to extract myself out of it out of the situations or the decisions or the technicalities that doesn't require my judgment. But also, it also gets me to something that I've been talking for ever since ChatGPT arrived, which is the idea of personal AI. Like having your own AI that has generic attributes but also have personal attributes. It can have generic skills and generic tools, but it can have very individualistic ways of thinking, frameworks, decision making, judgment calls, work specific workflows, systems, whatever it is. So all of this is kind of collapsing right now into how I use tools, how I try to inspire my team to and sometimes I learn from them and like how do we do this to maximize the uniqueness and create a competitive edge because in a world that moves so fast, edges starts to blur. And to maintain your competitive edge, sure you need velocity. You need super high velocity, super high momentum because it allows you to do cheaper mistakes much faster and integrate learnings, make them institutional learnings and do the right things. But also it requires us and this also connects to velocity to be able to spend our biological time in the right way. And this requires a lot of thinking, not just executing. People use AI as tools, but they're more than that. They're systems, their platforms, their architecture. And the people that are using AI as tools, it's like once I had Microsoft Word, then Photoshop and I use and then Figma. And the way I use Claude is the same way I used Figma. Don't do that. That's idiotic. That's stupid. It's not going to you're losing on many things. We can talk about what are the things that you're losing. But you're not using AI. You're not using AI the way AI can be used that would give you a competitive edge as individual and us as a company.
