No Priors: Artificial Intelligence | Technology | Startups - RAG is the key for smarter productivity tools with Notion CEO Ivan Zhao

Episode Date: February 15, 2024

Notion is a productivity app that has invested heavily in AI to create products that enable workers to access information instantly without having to search through their own countless notes. Today on No Priors, Sarah and Elad are joined by Ivan Zhao, the co-founder and CEO of Notion, to talk about Notions Q&A interface and calendar applications. They also get into how using RAG models means better retrieval, longer memory, and the user can be less organized and how Notion is leading the charge in this era of SaaS bundling products. Sign up for new podcasts every week. Email feedback to [e-mail] Follow us on Twitter: @NoPriorsPod | @Saranormous | @EladGil | @ivanhzhao Show Notes:  (0:00) Introduction (2:09) AI and Computing literacy (5:39) Building the Notion AI team (8:43) Notion as an application company (12:09) Prioritizing AI investment (14:53) The rapid evolution cycle of AI development (17:46) Notion Q&A (20:00) Workflow and AI for calendars (22:43) Moving past the need for organization (24:36) History of SaaS doesn’t repeat, it rhymes (30:14) Design at Notion (34:26) Notion office design (36:52) How RAG will change the future (38:30) Building our the software in the Notionscape

Hi listeners, welcome to NoPriars. Today we have Ivan Zahou, co-founder and CEO of Notion, the beloved productivity application for notes, tasks, and knowledge base. They recently launched an AI Q&A interface as well as a calendar application. We're super excited to have Ivan. Thanks for being here again, Ivan. We are going to start with the hardest question, which is what is Notion? Notion is always pretty hard to define because it can do so many different things.

But that's also our goal. We want to give people one tool that they can do their most work with. For a personal user, that means all your personal notes, all the planning for a trip for your wedding. For business, for enterprise, for a company, that means all your documents, all your tasks, our issues, calendaring, knowledge-based in one tool. The reason we want to do that because there's just so much fragmentation in the market today, we wish, like, it wouldn't be nice as one place to do your most work. And our approach here is rather than try to cram all different use cases into one product,

what are the underlying software building blocks? What are the Legals that power those use cases? Can we give users those Legos so they can be creative with software themselves? They can create and tinker their perfect workflow for their personal life or for their company. And none of this is new, by the way. people back in the 80s, even 70s, tried this kind of building blocks approach to software, which you're trying to take a modern spin with cloud web, with AI, to what it's like to break the prison of application-based software.

It's dramatic to think we've been living in a prison of SaaS fragmentation for the last two decades. But I do think it's actually, you know, surprising to hear eight points of view that is so, which is like, of course, we want one tool where the data was interconnected. Why do you think more people don't try that to have unified tools and unified data underneath? I think people try for different angles. Like even fairly recently, there's this thing called no-code, right? No-co is like coming from this kind of like power user-developer angle of wouldn't be nice everybody can modify this underlying software they use every day.

every day. That's one angle. It wasn't coming from the angle of the knowledge and data wants to be in one place, right? And language models sort of give another angle is the underlying knowledge in the betting space wants to be one place. Wouldn't it be nice in one place, right? And the macro is also coming from the budget place. Wouldn't be nice rather than pay for five different vendors and all sea-based business just to pay one vendor and save some money. So there have different angles from different times. I would say we are more come from this kind of computing and medium and literacy angle. Like, you and me go through school to learn how to read and write, English and Chinese.

We've spent years to do that. We all know how to do that. The world, the same MacBook for most people, are very rich as more like a machine to do typewriting or watching YouTube, not much more beyond that. It's not very creative, right? Wouldn't be more nice than more people can use their software more creatively, right? Because there's a separation between people who can make software and people who use software, that's why SF's rent is so expensive, because we're the modern-day Detroit or Manchester, right?

We're the factory of the world. So notions largely come from that angle, which is the original angle. We were inspired by early computing pioneer. They thought about that angle, right? They thought about computing could just be like literacy. One day everybody can do it. I guess they didn't expect AI might make that even give a really interesting twist to it. Because now language model AI can not only to create software, but also do a lot of thinking, working for you.

So the future is pretty interesting. So for someone who thinks on, you know, span of, like, decades of, you know, what should computing look like and what were the most ambitious plans for personal computing, you know, three, four decades ago, like, what are you most excited about seeing from AI broadly over the next decade? I think three, four decades a bit too long. If AI happened at that time, like computing might not be necessary. For this decade, I think one sleeper category is the dragged embedding space. The decades might be too long. I would say in the next year or two. Now the language model can understand what you put into a computer, understanding.

So rather than you do the organization to make you retrieve the understanding more easily, machine can do that better than anybody else can. So before that, we use keyword-based search where you find your own. a coworker who will remember that, that queue, where does that information sit? Now, just ask notion AI, and you get that in seconds. So that's one. I'm personally really excited about it. I think not enough people talk about it.

And of course, the other one is like the agent, the workflow side, that has a lot of buzz already. So that's interesting, too. You and Simon said bet the company on AI and are, you know, have real conviction. And as you are building out the team, like what does the talent look like you have or need to make notion an AI-first company? And I kind of argue you folks are one of the earliest adopters of AI at scale as an application. So part of the question in some sense is you've built so many interesting things,

like what are the people that you now need to sort of build the next level stuff in addition to what you already have on the team. In the early days, kind of just move forth. Slymer's a really good thing. I've built a lot of things. I learned really quickly, right? I would say notions the company were largely people interested in interface and design a lot of full stack. From them heavy folks and back and people who scale, we have somewhat a small team of search,

but we don't have too many ML folks, almost nothing. And at least my learning, our learning in the past year or so building for AI is, okay, ML folks are important. It's kind of like you no longer do it. a this deterministic thing that you can see how it works it's almost like a I don't bake but feels like a baking right you have to like do something get the thing ingredients ready run through rinse press the button away for a while and see oh does it come up or it's a different way a different sense of patience and different type of personality to do that well a lot of massaging

a lot of preparing so my friends call that probabilistic software engineering kind of like i think it's it's morphed into this sort of stochastic world or at least partially stochastic Yeah, so one is like, maybe gardening feels like that way. I don't garden either. So that category of people are, to me, is pretty necessary. The other category is like people who are curious and learn really fast, right? It's like, okay, like the group of prompt engineer, language model sort of make everybody like a real-time machine learn, learning engineer. you just prompt right, and then you can get your stuff, right?

And there's a lot of trick and techniques. And how does that plug into user interface? I think there's a category of people call AI engineer or something. There's a terminology form. They tend to be pretty young. They tend to be like we have someone like under drinking age working at the ocean. They fit into that bucket. And I think both things to work quite well.

We don't have too many researchers at Notion. another one I think will be important, but we're fundamentally set in the application layer. So it's more about apply side of things. So we're manipulating the models and making sure you can scale them to, like user outcomes and think about that. And another part is like the scaling part.

How do you scale to like tens of millions, 100 million users? It's a problem on its own too. It's beyond just a demo on Twitter, right? It's tricky. Yeah. So you have often said that Notion is less a productivity company than an application-building company. How do you think about the initial use case and what makes you believe people want to build more applications?

I don't think people want to build more applications. What got me started in Notion, Godda started in Notion, it's last year of college. I read a paper by one of the computing pioneers Douglas Angobar. He talked about his paper's name, Augmenting Human Intellect. So every day we use software today, very much like application. When you go into one application, do one thing. But for that generation of computing people in the 60, 70, 80s, computers are a lot more, software are a lot more malleable.

You can actually tinker and modify, small talk. You can go into it and change the hot openings and work on the fly. That really inspired me. It's like today, people's software are so rigid. can we create a new breed of software that people can modify, can change, can customize, and bring back some original ethos of those early computing pioneers.

That's why we start a notion. The hard lesson for us is like, like you mentioned, most people don't want to create software. They don't wake up and say, hey, and you want to create my perfect project management toward my project, perfect knowledge base. The boss asks for something, they just have to get the work done, right?

So in some sense, our learning and pivot is instead of giving people those software building tool, we have to package the software building blocks together as ready to use templates, as ready to use cases, then people can adopt really quickly. So you were one of the earliest adopters of AI in terms of an application with any real scale. And I think it's impressive how quickly NoShan ended up starting to work in this area. How do you think about how that impacts different aspects of what you built and what you're building going forward. And how does that impact that vision of saying,

okay, we have this effective platform that allows people to both interact with documents or core use cases in simple ways, add things like calendar, but then also go in very interesting directions in terms of both the set of applications and templates they can use. Yeah, I think we're lucky. Like I mentioned, we're not trying to build specific use cases.

We're trying to build a Lego bricks that power those use cases. What are those Lego bricks? Text editing is the one fundamental Lego bricks. Most software have that piece. Relational database, a table, is one fundamental Lego bricks, right? Different form of permission, commenting. So we've been spent five plus years building those Lego bricks

and feels like, boom, AI just drops in almost like a brand new car engine and can power those Lego bricks in brand new ways. So it feels very lucky in that way. And that, because we've been building those Lego bricks and refining those, allow us to ship features plugging with AI really quickly. where one of the earlier is one to launch AI writing for productivity software at scale

because we've been spent years building a text editor. We can do AI-powered database table features really quickly because we've been building relational databases. We've been building a knowledge base for a long time, so we launched AIQ&A really quickly, fairly quickly, the Rack system on top of the notion because we have those Lego bricks. So in some sense, kind of like just right moment, right time. for us. How did you begin to resource and prioritize this effort?

Because you're like, magic. We have this engine. It applies for our Lego bricks. And then you start shipping pretty quickly. But I think there are a lot of organizations right now trying to figure out what to do with AI. And so, you know, in terms of like designing the features, prioritizing that effort versus everything you're already doing in a rapidly growing software company.

Yeah. I think I had the conviction. My co-founder, Simon, actually, they all had the conviction. Funny because we all live in the mission, right? And Open AI initially still is in the mission. And some of them are friends, especially Simon's friends, work at Open AI. I remember we go to their office, they do in the Dota days. They were like, what is this company doing?

Kind of interesting. And Simon and some people in the ocean saw early demos of GPT. It's like, what is this thing? Spin out text, sometimes gibberish, sometimes useful. I personally, I have to admit I slept on it. On GPT3, even saw a GPT3 feels like, what is this thing useful for? It's like, yes, for marketing, for content writing, for the first draft. It didn't really click for me.

Personally, for me, it was fortunate enough to saw early preview GPT4. And that's like, oh wow, this thing can think. It can reason. You can know how to do things. this little bit workflow power to it. That's a big aha for me. And like it just personally gives me so much conviction, like this is going to change everything. If you think about what knowledge work is in, right, why do we use software? Fundamentally, SaaS, software, is all we're all in the same information, people, paper pushing activity, right? It's like a piece

of paper coming in front of you, a human, like, changed a couple of bits, push to another human. Language model can do some form of this now. So that just, like, give me the conviction. Like, this is going to completely change everything we do with the computer. And after that, we sort of just bet the company on it. Like, we're lucky enough to have those Lego bricks.

And then which Lego bricks can work well with AI, which doesn't, we're trying to figure that out. Who inside a company are good with this technology? We have search, but it's not like, we don't have a lot of ML folks. So you need to hire more ML folks, need to get people inside a company to have similar convictions so we can move in the same direction. It's quite interesting. It's kind of like it must have a dinosaur feels like when Astro hit the Earth and it's like, what do they do? Yeah, yeah. There's a lot of change coming for sure. It's a lot of change. Yeah. What do you think

is missing from the capability set? Because to your point, I think a lot of people weren't really thinking about AI too much until Chat GPT and GPT4 came out. And there was a period of time. where 3 and then 3.5 and you start to see the capabilities incrementing up and entirely new businesses are suddenly enabled with each sort of step with the next GPT level model you know GPT 5 or to your point rag adds a lot of capabilities what are the biggest missing gaps for you to take full advantage of this technology is that future reasoning is it better uh thinking and knowledge like what's the yeah i think all about that above to me feels like technology is fundamentally we're in the business, technology is fundamentally about trade-offs, right? It's like the plastic can do things

that wood cannot do. We discover plastic and then we figure out new things. We can bottle water like this before you cannot bottle water with the wood table, right? So all of a sudden we have this thing called language model. They have a new characteristic that deterministic software cannot do in the past. And we don't really know how it's made fully. So every month, every week, if you're on Twitter, people discover new techniques to get more out of this. And for companies, for entrepreneurs, they're also making tradeoff discover how the market react to this capabilities of this new word, this new language model. And so it's a constant evolutional cycle happening really, really fast right now, right?

I think with that mindset, what are the dimensions? On the technical side, on the technology side itself, yes, the model gets larger context windows, more reasoning, better speed, smaller for print. Those are all great. Like for a notion to power the workforce that we really need, like we learn like GPD4 is smart, a cloud to a smart, we need that intelligent to do reasoning, or for the tech summarization, cheap fast, it's better, right? That's the technology side.

And in my opinion, there's so much about human's behavior as well. Just like inertia in our personal behavior, companies risk tolerance, and that's slowly evolving as well, right? Like what Steve Jobs, I was talking about, you cannot make something too new. You have to be largely the same and change one thing and two things, right? Virtual Abelow, the off-white guys, like 3% difference. Just push the boundaries so people can accept it,

still also new, right? To me, it feels like language model, power, application are kind of in the same phase. If it's too different, people like, what do they do with it, right? Such alien behavior, it has to, the rag is pretty nice because largely existing behavior but better output, right? Can you describe the AI Q&A product for people who have experienced it? Right. Essentially everything you put in notion, notion help you remember.

And this is not just applied to Notion, pretty applied to most Rack systems. But like, why do we use computer? We need to store things and you to recall things. Before language model and Rack, the recall largely happened based on keywords, right? The keyword has to be precise, or there's some lexical tricks that you can recall easily,

imprec precisely. What Rack happens, language model can actually understand what you're put into there. So you no longer need to organize your information in Notion. Whatever you throw in there, you can find it later. What that means is for a person or for a company, for a team, you can have perfect memory. And not only have perfect memory, the right piece of information,

if we design our software right, can push to the right person at the right time, right? That's probably more than 50% knowledge work, right? We're still perfecting the system. I think we're one of the first on the market that apply at scale. We still have somewhat a waiting list, on a waiting list because it's hard to do this at scale still. But for a company, for a team,

before searches to one of a weaker point, but with the rag, you completely change that. I changed how I use notion. I can just ask a question to notion, like how large, when we're moving out of the SF office to a new office, as someone in the company wrote in some documents, I don't have to ping three different people

to find the answers, if it's a notion, go find it for me, right? Every day engineers, designers, operation people just keep asking each other on Slack or in email, such a question. Each question is 10 minutes writing the answers, 20 minutes to find answers, and there's the delay in the middle. With notion Q&A, you can completely cut that into in seconds. We're just at the beginning of a rack and do for work. It's pretty amazing. I feel like rag and embeddings are very under-discussed or underappreciated in some sense, right?

in some sense relative to the impact that they really seem to be having or starting to have. And I think Notion Q&A is a great example of that. I guess the other thing that you folks just launched is calendaring. And if you can't talk about it or if there's nothing to talk about this point too, but I feel like one of the really interesting things that people are talking increasingly about is agents and sort of the agentic world. And there's a lot of capabilities missing to really make those valuable. But in the context of the calendaring application, you could think of all sorts of ways that

having AI act on your behalf or help understand things can be incredibly valuable. And so I was just curious how you think about the application of AI relative to calendar versus, you know, some of the core information-related things that you just talked about. Maybe we can group AI stuff into, at least my mental model, it's the rack, the knowledge, information retrieval is one bucket, knowledge bucket. Then there's this workflow bucket, right? You use the word agent. That's in that bucket. Counter somewhere in that bucket. Why do we need to meet? Why do we need to Canada? Because we need to meet and when you schedule time, we need to figure out, exchange some kind of bits between my brain to

your brain, right? Can that bit, can I exchange be done by a language model? Maybe. And can the meeting time be done by scheduling be done? That's like a baby step, right? And most things we do has this kind of time dimension to it. Can language model help us shuffling our schedule? Yeah, It feels like there's also the information retrieval piece of it because, you know, if my calendar auto-populated everything I needed to know about the meeting or the people attending or other things, that's incredibly valuable as a user of a calendar. And so I just feel like there's a lot of these things that kind of tying together both in terms of the coordination, which you mentioned, and the workflow and then separate

from that, there's just, what do I know about this person? The calendar part is the simpler part of the workflow. Like the Holy Grail is kind of like, can just the agents, robots, do our knowledge work for us, right? It's a really interesting framing that I didn't have before, of a bunch of the work you're doing a notion actually eating into like communication, right? It's sort of obvious in retrospect, but like if you look at what you describe of like, um, like, am I really going to slack back and forth about this thing about, you know, when we're moving, if I can just know, I'm in motion, help me know, or with calendaring, like, you know, the most intelligent version of it is like, well, do I need to have that meeting or can you tell me what I was going to tell me? I know. Like, why do you need to communicate? Because, there's something, the work cannot be done asynchronously by the software itself, right?

And that's why you talk. Yeah, it's kind of interesting, maybe it's an interesting question, like, are we kind of communicate more or less with language model? I probably feel it's probably less. The agent side essentially bet on language model, that's the communication. One question I have for you,

just going back to like the implication of RAG and, like, you can be my brain and do my organization. for me, like, what if my brain is really disorganized? Like, do you think that this changes the amount of work people should do in systems like notion input, right? Like, you know, should I be designing my knowledge base in the same structured way? Or kind of just dump it all in a stream of consciousness in the future? I think organization might be, we might be moving away from the organization in the world.

why do you need to organize because you can retrieve? Why do you have index? Like, index initially are file cabinets, and the little index are sitting on top of us so you can find things quickly, right? And they're indexed based on certain names or certain dimensions. But embedding and rag, sort of,

you have semantically connection of all the things. You throw into this file bag, and you can find bringing that out however you like. So I think we might be moving past the need for organization. That's really liberating. That means on my phone, imagine this experience. I'll have a new idea where I see a whiteboard behind me. I just take a picture or write something, dump it,

and no she's going to organize for you, right? So then that's become my perfect memory to start. Later it could be my perfect assistant to help me do something with this knowledge. And that's the vision we're moving towards. That's super exciting to me. So you are, this is a question for Mike Rahl, we've been long-time friends with, you know, you are a student of history. You mentioned that Engelbart earlier.

I know you think about, you know, the transition in terms of like Alan Kane, what he did in terms of simplifying many of those concepts for, like, a broader audience around computing. Bernal's question was, what lessons in history do you take that, in form your point of view of, like, how to treat AI strategy with notion now? Like from a prior revolution in computing, you know, how does that help you decide what to do? A lot of intuition. I think understanding history gives you a sense of history doesn't repeat itself but rhymes. So, like, okay, which phase are we in?

I personally think we're sort of in this kind of bundling phase. Like, who said this? Like, there's only two ways to do business. bundling and unbundling, right? And actually, during the breakouts, you're reading a Chinese novel, Romance of Three Kingdoms. And the opening sentence for that is, the empire long divided must unite, long united must divide. That's as always been. Business is the same way too, right? We're in the bundling phase. I would say the SaaS, it's sort of this unbundle fragmentation phase.

If it treats back to SaaS, why is SaaS happening? In the mid-2000s, before that, everything is running on Microsoft. That was like a bundling phase. Early days of PC, there are so many different applications. The first version of the world-perfect, different text editors, D-Based, different database software. The funny fact of D-Base, it's like they start with D-Base 2. Because there's so many company go busts that it sounds like if they start with D-Base 2,

people have more credibility. It feels like this product has been around for a while. So that's the 80s. 90s was this kind of bundling phase because Microsoft has OS layer underlying it. And the SaaS is because the web becomes good enough to run software. Right. Then so then we have this unbundling phase, a fragmentation phase. And then with the last 10, 15 years,

it's really, the money is cheap, easy to create a company. There's so much. Too much now, it feels like. There's like information so fragmented. And now the new technology is happening is AI language model. And if you build more with it or just think more with it, language model wants information to be one place, wants the endpoints to be connected,

so it's easier to, It's hard enough to look at a current version of language model to do what you want, but imagine top with different endpoint. That's even harder, right? And so we're in the bundling phase because the macro, but we're also in the bundling phase because language model, I believe, wants the things to be together. I think that makes sense.

I also feel like we're in the bundling phase because the nature of how founders think about their businesses shifted. How so? I think that it's interesting because I remember, I don't know, 10 years ago I used to argue with people about, oh, you should really buy other companies or integrate or sort of pull all these things together. And in consumer, that actually happened, right? Like Facebook bought Instagram and WhatsApp and other things.

And they effectively created like a bundle of social products that they could cross use in different ways for distribution or other things. But I feel like what happened is we had a series of highly technical founders because we shifted in the Facebook era from Cheryl becoming CEO to COO. And you went from business-centric CEOs in the 90s in some cases, although there's people like Bill Gates who learned and adopted as technologists. to very technology and product-driven founders who often thought no matter what product I build, it always has to be better. And so I can't just think of distribution as my wedge. I need to think of every product as being superior, and so I'm not going to build certain things. And now I feel like people are both building great products as well as bundling them, but also they're much more aggressive about saying it can be 80% as good. It could be 50% as good, but I'm going to

have a bundle. And that's HubSpot, and that's Ripling. And they have very high quality to their products. It's just they realize they don't need every single edge case and every feature. as long as they're able to cross out. Yeah. I think the YC school's philosophy built one thing, use Internet to find the distribution. That was, I think, overlap quite a bit with the rise of Internet, right?

And it feels like there's a value to create on the other dimension, which is, like you mentioned, it doesn't have to be as good, 90% as good. But because the synergy of things just make a lot easier, a lot cheaper, less tabs opening your browser.

Yeah, it's all integrated. You have the information flow or the system of record for whatever thing that you're dealing with. I think a lot of people also just perhaps lack that sort of historical context. If you look at the strategy of companies like Oracle, right? It was very much for a decade and a half, like a dominant, at least commercially, attitude of like, okay, we're going to buy the second best product in this additional software category we want to be in and then sell the heck out of it.

Works great, right, actually, because it was very hard for customers to do. deploy these different things, or they're just advantages to everything being attached to a single database at some point. And I do think there is some analogy to, as you said, language models because having things in the same embedding space is very useful. Very useful, yeah. I think there's bundling of distribution and bundling of information. What you're describing to me is more of Microsoft, more like bundling of distribution. Languantan model wants the bundling of data, bundling information. So I remember hearing from Dylan at Figma early on that, there was one crazy user who was in the product like 14 hours a day.

It was you early on in the Notion journey, being really design-obsessed. I think the company has a reputation for that. Do you think of Notion as like a design-centric company? And is it important? How do you scale it? I think it depends on what you mean by design. Design is to us, at least to me, it's less about how it looks. It's how the system plucked together.

And then in that case, the trade-off you make, do you centralize that thing or do you decentralize that thing? Certain company work while, or certain business of product work while being decentralized. Like, operation-heavy company could work that way. And Notion, like you mentioned, we're sort of in the bundling business, our value provided having this one information space, one workspace for people do all different kinds of things. So things need to be work-while together. It's almost building notion. It feels like building an operating system. building a programming language, right?

You don't farm out to like 50 people to design a programming language. Usually programming language are done by one person. So that means the design here is very much a centralized energy. It's kind of like Apple, how they build OS integrated with their hardware and, right, like what is the Apple for software? It doesn't quite exist today. It truly doesn't quite exist.

So that's what I'm interesting, what we're interested in. So in that case, means to build a good product, a good customer user experience, we need to think things more horizontally, more holistically. That means the decision making tend to be centralized or design team or like, tend to be centralized, right?

So less like, more Apple-like, less Amazon-like. It's funny because when I first met you was just you, starting notion, and is before you brought on Simon. And you talked about things that way even then. And I felt that one of the reasons I was lucky enough to invest or you know I came on board was because you had such a cohesive view of how you wanted to build software and you had such a cohesive design aesthetic

and it was your mocks but it was also how you were dressed and how that reflected into the product I felt like it was extremely striking you know like you're one of a very small number of people I've ever seen where that design aesthetic has just kind of permeated everything in a very cohesive way and so that's one of the things that got me excited at the time I was like wow, this is capturing a aesthetic that could be an incredible product platform. But you also talked about things. Even then, I remember in terms of like, okay, what's the cohesive Apple-like thing that you can do for software and things?

I think it's kind of amazing to see that consistent thread. So I was just stricken while you were talking by that. Thank you. Yeah. Like I studied cognitive system, cognitive science, which is kind of just like a degree for everything in sometimes. It's like a little bit of philosophy, a little bit thanquist. did computer science. I learned how to code when I was a kid, and I did a lot of art,

and also in school. So, like, try not to, like, there's so many things you can steal from all of different places, right? And it's like, the boundary are sort of man-made, and then in the notion was most of our designer can call. But Georgia, 80% of our designer can call. Because the moment you can, as a designer, or as engineer, you can call, or you can design, you can make really interesting trade-offs, right? At the end-of-day technologies, at least in my opinion, is about trade-offs. What kind of trade-offs you can make that unlock new user behaviors that's valuable. But if you can do more things, you can make more interesting trade-off that other people cannot make.

As a designer, if you can code, you know how to change your design to make easier to build as engineer. If you can design, you can do the same thing, almost like, squeeze the air bubble to whichever direction is easy. to squeeze the threat. And therefore, I think being a more holistic help, at least notion as a company energy, we're trying to be holistic. It also helps keep our company team very small. Like we're usually one of the smallest

relative to our business scale because people can think, can do more, can be more holistic. And people enjoy that too because they can do more things. It doesn't feel like they have one role, they have to be doing that repeatedly. That's a lot of different benefits. But it's much harder to find such people.

An important question here. Please. So if I think about the first office, and this may or may not be true still, notion is a no shoes, was a no shoes place? Did this contribute to the company energy? I'm Asian. So when you go home, you take off your shoes. Our first office, no shoes.

It lasts us to 10-ish people. Second office, 20-ish people, no shoes. Third office, no shoes. no shoes. It actually has heated floors or even better. It was all in the mission. The fourth office, we try to do no shoes still in the mission. I think I made a wrong choice in the rug. The rock kind of hard. It's a hand-based rock. So when you step on it without shoes with socks only, it's so it hurts. So we decided not to do no shoes at fourth office and so far has been stuck that way. Yeah, applied intuition has socks and slippers at the front.

So that way if you need the padding... I know, but the question is like, where do you store the slippers, it becomes stinky? It's like... And then you just Jay a Sutter. Yes. If you come to our office, it's like we're still trying to be not corporate. It's like they were trying to use the furniture that people use for homes. I'm pretty picky about what kind of furniture is in the office.

Like ideally design classic the last 50 plus years, so inspire us to build software that way. Right. So they made... trade-offs. People who design a chair, make data, they made a really interesting trade-off force to solve certain problems. If you know the history of it, so we try to in the office use good software, good chairs, good lighting. There's back to the aesthetic point that I made earlier. I actually felt that in the offices as well. There's that ongoing cohesion. Even the music, I remember, I think it was in the second offices, but it was always jazz in the background.

And I just felt like it all kind of was this consistent vibe, you know, so it's pretty cool. I think Notion, are you using a singular underlying LLM or are you at this point using multiple different things for different use cases? You mentioned sort of the high-level reasoning versus the fast-cheap sort of synthesis. We try everything. Open-Aid and Tropic are the high-end model. We want reasoning, which is we work with a high-end model, right? Yeah.

It's kind of like everybody building different flavors of this. Yeah, makes sense. And then as you look at, it feels like with Notion, there's a set of core sort of templates or use cases. You know, there's things around project management. There's other types of almost like applications that people have built to use. There's knowledge-based related stuff. There's the things that you mentioned.

Are any of those you feel differentially impacted in terms of how you think about future AI roadmap or things that, you know, will really change the game dramatically in terms of some of these areas? Yeah, I would say rack changes all the knowledge, sorry? fundamentally, you no longer need to organize. So the notion one of the things people love is the life sidebar, right? The life sidebar, you can organize your knowledge-based, organize your personal workspace.

Maybe the future doesn't have to have that. Like what it's like to, you know, not far into your own innovative dilemma to double down that U.S. paradigm, but just having a notion that you can just dump things and retrieve, That's knowledge style. That's actually really interesting at a high level to think that everything sort of moves to a form of search over time. Move over to search over time?

Like you're kind of losing. You don't need to self-organize information anymore in this new world. You can just create a mechanism to interrogate it. Yeah, at least you don't organize your brain. You just dump it into it and you wake up, oh, you remember that thing. So like some people do. Like the art of low-key, the art of memory, you actually visualize your brain.

but for most people, it just works without any organization, magically, right? What it's like for software? We're getting there. Yeah, it's kind of fun. Are there areas of software more broadly that you think are outside of notion and scope that you think are going to change a great deal from AI?

Wow. In some sense, it's kind of a race. There's like the, we're in the, notion is in the bundling business, where are, we are, we're in the bundling and front office business. Front office, my definition, our definition is, what's happening in, like, imagine a 1960's office, right? What's in 1960 office? On your desk, you have a notepad, you write on something, maybe have a typewriter.

Then you have your binders on the left and right. That's essentially, the notepad is your documents, your notes, and notion. your binders of things are like your wiki knowledge-based notion. And behind you will be the file cabinets. That's your relational database in notion. And you have a little push card to put things into there. Then there's a back office where it's like the librarians organize all the things. That's snowflake, right?

That's the back in the day's IBM. We don't touch that. We largely touch our strengths, like I mentioned, is software interface, UI, UX, which is largely what's in front of the human. And we're trying to bundling this in one space. At the same time, there's also largely back office power use cases. They tend to be verticalized, specific to health care, specific to some kind of workflows. Manufacturing.

It's very specific, but it's very essential to store somewhere and the vertical integrate that use cases. That could be AI-Fi, too. And people, in fact, we see this in law, we see this in a bunch, very specialized thing, that people have that domain knowledge and trying to figure out how do you, instead of human shuffling this, the language model help a lot of that, right? The front office type of things is kind of open-ended. The back office power things tend to be specific. So I think it will be a race, but the market is just so large and it doesn't, it's not zero sum necessarily. Maybe you can talk about the market as you see it for notion. So, you know, when you

you guys began, I think early adopters, startups were the first to get onto Notion for knowledge base. You're a much bigger company now. We're also in a different macro where you know startup budgets are less robust. Like how do you think about the enterprise and helping the enterprise adopt AI or you know do knowledge management? Yeah, we're still early, not at scale yet. I would say bundling in Transit League do a lot of good things. One is you don't have to jump between different tabs to do things. Second is save your costs, right?

Like we save a lot of customers built for their project management tool around their issue tracking tool. And that's... Enterprise really care about that. It's very CFO-friendly today with this macro. So, yeah, Spondonly has many good benefits besides money, besides information they're supposed to money. Ivan, I mean, this conversation has been so many interesting topics. Thanks so much for joining us today.

Thank you. Yeah. Great to see it. Good to see you. Find us on Twitter at No Pryors Pod. Subscribe to our YouTube channel. If you want to see our faces, follow the show on Apple Podcasts, Spotify, or wherever you listen. That way you get a new episode every week. And sign up for emails or find transcripts for every episode at no dash priors.com.