Meet The Photographer Who Built a $10B Startup (Ivan Zhao)

Tonight, we're sitting down with one of the most influential founders in tech, Notion CEO Ivan Zhao, to talk about the future of AI agents and whether anyone can actually build a useful one.

Are you excited to meet the most tasteful man in the valley?

I'm genuinely nervous because I don't think my fashion or table manners is going to be up to scratch. Yeah, it's funny.

Like, I feel like 5 years ago, we would have dreamed about this opportunity to have dinner with Ivan, founder of Notion. Think of a $10 billion-plus company now, and there's so many things to ask me about.

Most founders we meet talk about speed, headcount, fundraising. Ivan talks about photography, Mozart, and beauty.

So beauty is on one end, utility on the other end. I really like the quote from Bell Labs back in the days, creating something beautiful and useful, right? We want to create tools that bring more of such utility and beauty to the world. And that's my energy source.

But something that's been really interesting to us is that there's been so much hype around agents, but honestly, I don't use a single agent day to day. I try them all out. I kind of forget about it a few weeks later. And it's gonna be really interesting. Can Notion build an AI agent or knowledge worker that we actually use?

It's very doable. We start with Notion trying to consolidate. There are too many SaaS products out there, so we have all the context in one place. Now you give this to language models, just let it cook. And it turns out worked quite well.

If an AI agent is only as strong as its context and tools, does that make Notion the best place company on earth to build a true knowledge work agent? Tonight we find out. So I'm really excited for it. Well, we're going to go walk up to the event now. It's at a fancy restaurant around the corner. It's going to be a bit of fun. Let's go.

And we're live.

Yeah, thanks for making the time. We've been spending the last few days and nights going through all of your photography online, all of the photography, as well as reading about Notion. But it was interesting to have a bit of a deep dive into your inner life. When I listen to you on podcasts and read some of your writing online, it doesn't seem to me like you're just a tech entrepreneur. I actually hear the soul of an artist. So do you think in another lifetime you would have been a photographer or an artist instead of a tech entrepreneur?

I wanted to become an artist. Like when I started exhibiting some photos in first, second year of college, I was thinking of doing art. Yeah, I actually set my mind on doing photography. I was in Vancouver at the time and realized like so much of doing art is like going to parties and there's too much people, social aspect of it versus meritocracy of doing science or doing building products. So I shifted away from art into tech.

What have been your favorite moments of photography that you can remember in your life?

I remember first and second years college, travel half China with my dad.

Yeah.

On road trip and taking photos, spend the summer with my dad. Those are pretty good moments. I remember the first time I ever been to New York, 'cause my photo was exhibited in New York Photo Festival. Wow. And actually at that time I thought maybe I should become a photographer, right? That's before I got into tech.

Do you think it's ever difficult to balance like the sensitivity of an artist? You obviously see the world in a very particular way. When we're setting up this set, I know you like things, you know, in certain ways with also maybe like the ruthlessness needed to build a multi-hundred billion dollar company. Do you ever find balancing those two sides hard?

It's really hard. It's kind of like Building a company is kind of like playing a sport, right? Building a business.

Yeah.

The scorecard is your revenue or profit. And it's like there's one way to do it. What's to make the profit number go up? Doing art is there's no one way to do it. There's so many different ways. That's your way. Too much of your personal art, you might not have something that's sustainable business model. If it's too much of a business model, only business model, then you're not creating something aligned with your value, become more like a commodity. It's something I every day I wake up, not every day, but oftentimes see which side I'm pulling me harder.

And I think that you can see in the way that Notion's built, how do you think about, like, you've gone through the journey of Notion, Notion 1.0, Notion 2.0, Notion 3.0. Is there any kind of thread in which art and beauty and this, you know, you said in another life you become a photographer, has carried across the whole Notion product.

And you use the word beauty. It's something we care a lot. I care a lot. So beauty is on one end, utility on the other end. I really like the quote from Bell Labs back in the days, creating something beautiful and useful, right? Either one or the other one. That's at the heart of what we're trying to do. You want to create something to amplify. You want to create tools that bring more of such utility and beauty to the world. And like, that's my energy source.

We talk to a lot of tech founders and very seldom do we hear people where their foremost concern about the world is bringing more beauty into the world. Why do you think you're different to other tech founders in that regard? Was maybe, is there something in your early life that sort of informed this sensibility? Towards aesthetics and beauty?

Maybe I did photography. My mom told me this story, I don't remember now. Like when I was 2 or 3, I won a drawing competition. And the reason why I won the competition, because like I drew something with perspective, like lines of soldiers and the soldiers on the front are bigger than the soldiers on the in the back. Maybe I'm attracted to something visual, to star. Maybe it's innate, maybe it's nature. I don't know.

Do you ever go one way too much and you feel like you're not being authentic to yourself?

Well, early days of Notion, we didn't have product market fit for first 3 to 5 years. That's kind of just too much of art. Then you're building something that only you appreciate, nobody else appreciates to pay your money. than too much on one side. So actually 2 years-ish ago we're trying to build Notion too much not like Legos, like Notion ethos very much is the Lego for software. For a while we're trying to build Notion too much like hard plastic things so we can go to market faster, more competitive in a specific, against a specific point solution. We did that for about a year and just feels like it's first it's Our customers don't like it. You're introducing a big new plastic thing to a box of Legos. It doesn't fit well with the rest of the system. Your community don't know what to do with it and your code base rejects it because your code feels like everything's modular. Now there's all of a sudden there's a big thing in there and your product team, engineer team doesn't love building it because they're coming to Notion to build Legos. Now, They have to build this point solution in the Lego set. It's an organ rejection and you just don't feel proud about it. So we changed course a year and a half ago. So only go all the way back to the building Notion in a Lego way.

I see Silicon Valley in some ways becoming more mimetic where people are copying more ideas, types of businesses to build. They're saying similar phrases. Something about you is that you've always had this— I think you've always had a unique approach to company building where there's been many unconventional things that Notion has done. I'm wondering now, How your sort of personal philosophy towards company building and taste is changing and whether that's leading to more unconventional practices that you can talk about, about how Notion is approaching the next 5 or 10 years.

Like what are the new ideas?

Yeah, any unconventional ideas that you're pursuing, you think that are as a result of your taste, personal philosophy?

I haven't thought too much about this. Trying to be— we're not trying— I'm not trying to be unconventional for the sake of it. You actually brought up an interesting one, like care about beauty. Probably most people in tech don't care about beauty as much. That could be one. Tech is very much of a— I actually like to quote this computing pioneer, Alan Kay. He always said tech is more like a pop culture. People, it's a tinker culture. People learn from people next to them.

Right.

People don't think back into history or learn from far away disciplines. And if you're zooming out, most of the body of work that people created is actually more about humanity, more about nature than about tinkering the latest technology trickeries. So if you're zooming out, there's more humanity to it. There's more photography, there's more art thing, idea you can steal from. Then how does that manifest into a company? I think that's an interesting angle, right?

Mm-hmm.

Treating, almost treating your company like an art. piece that has a business model, maybe that's an interesting angle.

It's a very Steve Jobs approach to company building, calligraphy and all these beautiful things coming together. Most people watching will never know what it's like to be the founder of a company that's worth $10 billion plus. What are some of the things people misunderstand about what it's like to be in your shoes and to be in the Silicon Valley power circles and to be the CEO of one of the fastest and biggest growing companies in this current era?

When people come join Notion, I like to tell them like, I'm no different than you. I just got here early. Don't take me too seriously. If you're just running company in Silicon Valley in tech, you sort of see, okay, some press piece going on TechCrunch. There's so much of media put on facade of something. make, put founders on pedestal, mystify everything. But end of day, I'm just another dude. It's kind of similar. Maybe luckier, maybe work a little bit harder, but not fundamentally different.

And staying on this tact about how you're building Notion now, we just saw you give a great talk, a lot talking about AI and how you're trying to build Notion super agents. And I'm wondering how you think about your philosophy to building in the age of AI when things are changing every week? It's new model releases, all these new companies and competitors. How do you build that enduring company in the age of AI while there's so much flux?

A lot of us are pretty lucky. Like we launched Notion 3.0. It's one of the best agents on the market for work. We did not start Notion to build for AI product. We started with Notion trying to consolidate. There are too many SaaS products out there trying to give people one tool that can do a lot, right through the Lego architecture. We heard a lot of buzzword of context engineering today, but just meaning they want the data to be in one place. And we, we've been building Notion 5+ years before consolidating all the context in one place.

So what does Notion actually need to do to achieve that vision? Because we see a lot of agent companies out there that are promising a lot. There's a lot of hype, but then they're not really getting the results that people think. And then there's a little bit of malaise after that.

Yeah, having context together is one, right? Like we're lucky that we've been consolidating context either using Notion first party to store all the knowledge and document and on one end for unstructured content and database for all those structured context inside Notion. The space we're playing in, which is general knowledge work, require teamwork, right? So you have seen specialized agents like customer support or like coding been more successful because, you know, it's first the context and tools are kind of in one place. Second, it's very much a solo work. And we sort of have been building this layer for some time now. And layer on top of that is this kind of UI engagement layer. You still, yes, AI agent can do more and more work, but humans still has to supervise, still have to say good or bad, right? So you need a layer to interact with the language model. That UI layer still matters and then you want a UI to be flexible.

So kind of summarizing that a little bit, you guys think you're best placed to build this AI super agent, one, because by luck, the way the product has been built, you know, in this modular way is actually very good for building good agents. But 2, because so many companies, you know, we rely on it, so many other companies rely on Notion for their second brain almost, that allows you to then kind of go up and have the right context to build the teammates and understand how they, how do you actually think about evals as well? And how do you actually evaluate how these teammates perform?

Actually, you just inspired me. Maybe we should eval based on how does the Notion agent do a certain class of task, right? Another way to think about Notion is as a box of Legos. Each Lego is different kind of tool, right? There's a text editing, there's you manage a relational database. creating different views to look at the database. We teach language models to use all different tools in Notion. Then you just eval against how well it's using those tools. And because Notion Lego has been able to encompass most knowledge workflows within Notion, so the agent that's building, that knowing how to use the Notion Lego block can do most knowledge work.

Do you think you're giving away some of your advantage by owning that context through APIs where people, other people could build on top of? the Notion context?

Yeah, like historically we actually hasn't been that open as a platform. Like we're more of a design-driven culture rather than a developer-driven culture. We're actually changing that. More recent people joining the company brings us more of a developer-first mindset. Like we're one of the first major players in embracing MCP, for example, the beliefs that it's When other people get scared, it's also a moment to open up and just need to make sure you regulate how information flows and hopefully create more value for the customer. They'll come back to you for good reasons.

You mentioned the term supervising before, that if Notion's vision comes to life, there'll be— we will be in Notion and we'll be supervising a team of agents. That's a very different state of work to what knowledge workers currently do. So what do you think are the different required skills to succeed in the sort of white-collar domains in the future where you do have this team of supervised agents?

Yeah, I think the need for capability probably will lower. You no longer need to be a great writer for a memo, for strategy doc. You need to have a taste if that's a good one or a bad one. So the requirement for capability will go down. Requirement for taste will remain the same, if anything higher, because you spend more time using your judgment rather than doing the work. I would say the third dimension besides capability and taste is agency. Do you have the drive and will to constantly try different things? Are you hungry enough? So capability, this needs to get democratized, can go down. Taste, Same level, if not anything go up, agents become what's the most uniquely human at this point.

It's like that archetype that's maybe creative and kind of lazy may actually do quite well, right? Like if you can direct an agent in the right way, but you don't necessarily are the person that can make sure every line is perfect or you don't cross the i's and dot the t's, then you could actually be a lot more valuable.

Like more people become less of a bricklayer but more of an architect.

If we kind of like dream for a second, what do you think this Notion super agent actually looks like in its best form?

It should feel like a teammate, literally has a teammate that you can design and build. So 2 of you, if you start a company tomorrow, can have the capability of a team of 20 people and you can design your team in a way as if you're writing a document or manage your project task tracker as easy as that. Uh, it's very doable. Almost technology is there. Notion can do quite a big part of that today.

Running a company is both a sprint as well as a marathon. And normally running a company, a startup for a few years is an extreme sort of journey to go on. And you've been now doing it for 10 years. I'm interested in what has really kept you in the game? Like what has kept your energy high and how do you stay showing up? year in, year out after such a long journey already?

It's treated more like a journey rather than destination. So focus on what's in front of you. Do a better job than you can done last week, last year. Pick up new craft, like a craft mindset. To me, it's a craft mindset and a journey mindset is kind of similar. That gave me a very durable energy source.

If we kind of went inside your internal monologue, what do you think people would be surprised about?

I usually think my inside and outside are kind of similar. I'm a pretty direct person. Yeah. Probably surprised I'm not that sophisticated because I express everything already. There's nothing more. And it's a lot of just intuition.

As we start to wrap up the episode, I always love to know about people's influences and inspirations. And I think something that is quite apparent about you is that you're very particular about the influences you do take on. It sounds like you've got a range of people you look up to, but you don't talk much about contemporary Silicon Valley hype. So I'm curious, how do you think about what advice to take on and what influences to actually really dig deep into?

Yeah, my general advice, even I talked about in the earlier meeting, is like advice are so contextual. A lot of advice can become very self-helpy. It's like there's one meaning, one way to gain meaning in life, which is that's often not the case. Most advice are very contextual. You need to think about where the context come from. Does the context still apply? Like a lot of go-to-market advice, each go-to-market method only works for a few years until the channel got saturated, right? So you need to think for yourself. I think that's probably the meta advice. Think for yourself.

That brings up an interesting question. Like, if you could have dinner with, you know, 3 people dead or alive, who would you pick?

Mozart would be interesting. Yeah. Mozart, he must be a very strange person. So probably one artist. I feel like you'd have a good conversation with Rick Rubin. Rick Rubin?

Yeah.

Somehow I'm not too interested in that.

We'll just see.

Yeah, I don't know why. I never got into it. I flipped his book a little bit.

Yeah.

Not Rick Rubin. Darwin would be interesting. Hard to find a third one, but Mozart definitely.

Yeah, I love that.

And if you could send a text message to everyone in the world right now and it would kind of get delivered on their phone, what would you write in that text message?

Oh, I really believe and really, I usually say this over and over again, at least for me, the journey is the reward. Your hair that's growing up, But it's a good feeling when you actually treat this that way, at least for me.

Yeah, yeah, awesome. I think we can start to wrap up there.

Yeah, thanks a lot for the conversation, Ivan, and enjoy the rest of the time in Sydney. You've come at a beautiful time.

It's a beautiful day. Yeah, is it always this nice? It's first morning, first day.

Yeah.

Wow, you should get up and go for a long walk outside.

I think this is my last session today, so I think, well, I'll do that.

Yeah, go catch the sunset at the beach. I think you'll like it.

Should we do that?

Yeah, take him to the beach, go for another swim.

Okay, thank you for having me.

Yeah, thank you.

Thanks for that, that was really great.