Early Facebook Leader: Forget Job Titles, Do This Instead | Julie Zhuo

0:00Julie Zhuo: Everything was great and then one day it was like whoa whoa whoa.

0:02Dara: So you were at Facebook for 14 years, pretty much grew up there.

0:05Julie Zhuo: When we were 10-20 million users, he was like, but wouldn't it be so great if like a billion people used it?

0:11Dara: You specifically did not take on the CEO title. Your ideal world is a world where there are no titles at all.

0:17Julie Zhuo: Just think of yourself as a builder. One of the things we decided was to basically nuke all meetings, like every single recurring meeting except for one.

0:24Dara: You wrote making of a manager. What skill sets do transfer over to managing agents and what do not.

0:30Julie Zhuo: I actually think this is really hard.

0:32Dara: I'm super excited to have Julie Zhu here. Julie, you're a lot of things. She's a writer of Substack newsletter, The Looking Glass, which is, initially I thought it would just be a product newsletter, but it's also kind of like personal growth and personal development. She's the head of design of the Facebook app, author of Making of a Manager, and now the co-founder of Sundial, which is building an AI analyst. So it's great to you here yeah thank you for having me so i want to start off on a contentious topic which is the future of roles and titles so at sundial you specifically did not take on the ceo title and your your ideal world is a world where there are no titles at all so for the person out there who has kind of centered their identity around roles and the ladder and titles how do you advise them

1:19Julie Zhuo: evolving in the ai era yeah well i think about titles as look they're kind of like a shortcut right? Like if you work in a very, very massive organization and let's say I don't know you and I'm trying to figure out what you do, it gives me this kind of shorthand for being like, oh, you have this type of skill or this is kind of like what your job function or responsibilities are. But I also think that they can be very limiting. Like I think all shortcuts to some extent can be very limiting. And I think especially when we were starting the company, I mean, we were like, you know, first there was the two of us and then it was like four or five people.

1:53Julie Zhuo: And like at that moment just being like, oh wait, we need to know exactly what you do, so let's give you like a shorthand and like that'll sort of, you know, be your box felt like not the right way to approach the kind of company that we wanted to build. Now I get it, like eventually, you know, you become two people, it doesn't like titles make no sense, like I mean one person, the title makes no sense, you kind of have to do everything, right? And then with five people, okay, like probably still doesn't make any sense because we can talk to each other, we can kind of know what each other does.

2:20Julie Zhuo: Sure, by the time you get to like 10 ,000 people, I get the practicality. But for as long as possible, I wanted us to all think of ourselves as like, wait, what is it that our company needs us to do? What is it that I have the skills or the interest to be able to do? How can I help out? And like, let me go and do that and not start having people be like, well, because my role is this, that means I should do X and not Y. Like, I think that is extremely limiting. And I think especially now with AI, I mean, the roles are changing. What each person is capable of and what a person who maybe was trained classically in one set of skills can now be able to do with these tools at their disposal is so much more.

3:05Julie Zhuo: And I think more importantly than ever, I'm trying to tell everyone in my company, just think of yourself as a builder, something very, very general purpose. And sure, yes, maybe this builder, you have a little bit more of an understanding of aesthetics and customers, and this one has a little bit more of an understanding of technical architecture. But ultimately, we're all just trying to build really great things, and let's go out there and explore the edges of what we can do and see if we can just become better at building.

3:37Dara: Enjoying the conversation? Got a question? You can ask Julie's Digital Mind. Just check the link in the comments or description. Yeah, is there a reverse to the argument where, okay, there are no boundaries, but just because you can doesn't mean you necessarily should? Like, when should an engineer not design? When should a product person not code?

3:55Julie Zhuo: Yeah, and I think that, you know, every time we're going to build something, right, every time we engage in, you know, let's say anything, I think in some ways we're playing a game. And there are rules to the game to some extent. There is a sense of, like, what does it mean to play this game well? So if we're all builders, ultimately to play this game well means we create something that other people find really valuable to use. And then, of course, you dissect that a little bit more. You're like, OK, so that means, you know, what does it mean that it's good to use? Well, like a lot of things, right?

4:24Julie Zhuo: It's got a great user experience. It's fast. It's performant. I don't know. It has the right functionality. And then if you dissect those into, you know, kind of the underlying skill, then there is such a thing as like, yes, certain people are better at this particular skill which is going to lead to this thing you're building being more successful in the marketplace and so i do think that there is um it's important that we over time can recognize that but the best way to recognize that i think is often to test it against the market or against again like the ultimate judge of whether or not let the data speak for itself exactly whether or not we're like winning the game and so i think it's certainly cutting things short to say like, well, because you have this role, you shouldn't do this.

5:08Julie Zhuo: And I have this role, I should do Y, right? Ultimately, if we both do it, and like one of our, I don't know, apps works better, or one of our designs is like more successful with customers, then let that be the proof point that like, okay, I should do that more of that because I just happen to be better at it. And sure, you can learn. And your skills are also going to grow over time. But it shouldn't be dictated by role, it should be dictated by the actual quality or execution of of what that skill is and whether or not it ends up succeeding in the ultimate game.

5:39Dara: Do you think there's an opportunity cost to maybe not dialing in your focus on one craft and depth into design versus trying to pull in multiple skills together?

5:49Julie Zhuo: I certainly think so, but I think that goes to kind of what is each person's nature. I think there are some people, and I work with many of them, for whom what they want is depth, and what they're better at and what they align with is to go deeper and deeper and deeper. And that's what they're passionate about. And I think that's amazing. And then I think there's other people, probably one of them, I find myself curious about a lot of things. And so, I don't know, maybe it's ADHD brain or whatever you want to call it. But I find that it's very hard for me sometimes to just only focus on one thing.

6:23Dara: Same.

6:24Julie Zhuo: And whereas there's a lot in the world that's interesting. And so, I don't know, I think sometimes if we have enough luck, we can self-select into careers that allow us to like be true to like our nature but I think everyone is different and so part of like whether somebody wants to go and finds it you know exciting to go deeper is like it's actually becomes fairly obvious after a while yeah I recently read your

6:50Dara: essay I'm an idiot which I thought was very present tell me about what led you to writing that essay?

6:57Julie Zhuo: Yeah, this is interesting. So we had made a bunch of changes at Sundial with the way that we were working. Part of it was that we wanted, we felt like with all of AI, like we just needed a completely like reset in terms of like what the cadence of work was. And one of the things we decided was to basically nuke all meetings, like every single recurring meeting except for one, which was like a demo day on like, or demo hour, like two hours on Thursday. So I was like the only meeting. And as somebody who, you know, like grew the company and in some ways is a bit of a control freak and had operated with like a sense of like, I need to know everything that's going on or want to know everything that's going on because, you know, I believe that me knowing was making me more like, I guess, you know, kind of, it made me more confident in some ways, or it made me believe that, like, if I could then direct something or I could, like, catch a mistake here or there, that that would be better for the company somehow.

8:01Julie Zhuo: And I really had to, like, sit with whether that was a true belief, you know, like whether it was actually the case that I somehow had to be involved in everything in order for it to go well. I started to get the sense that, no, maybe this is actually just like anxiety speaking, not actually a thing that was better for the company. And so that was what I think led us to trying this experiment. And I think the first couple of weeks were tough because I felt like I no longer really understood at the detailed level that I was used to what was going on. But in other ways, I was like, actually, this could be what allows everyone else to be able to step up, take more ownership, like grow in their own path as leaders.

8:45Julie Zhuo: Now that we've run the experiment for like a couple of weeks, like, or months now, like, it's been great.

8:50Dara: You're not going back to meetings? I'm not going, yes. Really? Yes. Okay.

8:53Julie Zhuo: Or, I mean, there's, we would do meetings if it's like, you know, like two or three people, because they're actually talking about a very specific task, but not in the way that we had before, where again, like it felt like, you know, people who were managers were have to like, look at everything that was going on. All of us are still playing some role as ICs. And so us on the leadership team, we have certain projects that we're kind of ICing or we're very involved in. And obviously we'll have small meetings just to coordinate work or to like with a very small number of people. It's very different than kind of like the typical like larger company review script structure that I was used to.

9:34Julie Zhuo: And I think I wrote that to try to remind myself that like, you know what, but actually trying to be in control of everything is not always the best thing for a company.

9:44Dara: Yeah, you should question why you even feel that way.

9:45Julie Zhuo: Exactly, and maybe what I felt, what I want, I mean, speaking very honestly, maybe I wanted to feel important. Maybe I wanted to feel like I had a lot to contribute, and maybe the reality is sometimes I could also be getting in the way.

9:58Dara: Yeah, I mean, I'd imagine if people were to hear this, the fear of having no meetings would be intuition and taste diverging, and things not necessarily going the way that you imagine, is it a function of you trusting the taste of everyone in your organization or it's just a very iterative process?

10:16Julie Zhuo: So it's both, right? So I think that first, what's important is that we have to understand what the vision is and we have to understand to some extent what good looks like or what it feels like. How do we know if something is good or not? I think everybody in the company needs to be very aligned on those things.

10:34Dara: What great looks like.

10:35Julie Zhuo: Yeah, on how will we know if we've succeeded and like kind of where are we marching towards or where are we trying to move towards? So that's kind of like alignment on the big picture. I think the second piece is you gotta, yes, you do have to trust people. And I think that was another litmus test, right? It got to the point where I'm like, look, if I didn't feel like I could be gone for a week or two and like positive things would happen with this person, then like maybe I shouldn't be working with this person. And that was another litmus test for us to try and move our way towards what would look like for a really great AI native type of startup in my conception.

11:13Julie Zhuo: So I think that's a second piece. And then I think the third piece is going back to if there is a way that we understand, it goes back to ultimately whether or not I think this feature is amazing or not is not as important as whether the customer thinks so. Yeah, it just shows up in the data. Exactly. And again, when you are a startup like we are, there's a couple benefits. One of them is that it's okay to be more experimental. We often have a dialogue with our customers. They'll tell us if they think something isn't great or not. We can tell them, hey, you know what, this is experimental. We're not sure.

11:50Julie Zhuo: And we can have that dialogue. So that makes it so that it's easier for us to just try things and then see what they feel. or what their response is to that. And I think that's a lot of... And if I'm, again, gatekeeping everything and trying to be like, well, I don't think it's good, so we shouldn't even ship it. We're not actually getting that signal from the market on how to help steer our direction.

12:13Dara: Yeah. In you recognizing what great looks like and your team recognizing what great looks like, sometimes new great violates what old great used to be. So I'm curious how you recognize new kinds of excellence?

12:29Julie Zhuo: Yeah, I mean, I think a lot of the, it's, it's, so the way I think about it is like, there's things that are like really, really old. And I think actually those are much less likely to change.

12:39Dara: And then there's like, things, exactly.

12:42Julie Zhuo: And then there's things that are like, kind of old, like, so for example, right? Like, you know, we can look at a design pattern, like tabs on mobile. And, you know, tabs on mobile are like, kind of old. And like, if you're going to go create a mobile app, like, okay, and, you know, maybe you We shouldn't try to reinvent what the tabs look like because people are very used to this. But that's still like kind of, that's not like old, old. I mean, that's like still relatively, I don't know, like 20 years old, right? This paradigm. But if we go to like old, old, like, you know, and I'll take the example of like where we are now with LLMs.

13:12Julie Zhuo: Like speech is old, old. You know, like talking to people is old, old. Like having communication through a language is like much more Lindy than tabs, right? And so, sure, in some version of the world, people are like, oh, well, like the, whatever, the chat interface is like so much newer than like the tab menu that we used to have. But in another sense, it's like actually more true to like how humans have communicated for a very long time. And it's actually more, like it's sort of more intuitive. And I think whenever we talk about like UI, we're like, the best UI is one where like it just nobody has to get a lesson on how to use it because it's just so natural to the way that people already behave right it's already like and so that's why I've even looked at like like the iPhone when it came out everything was very skeuomorphic because people kind of understood like oh like if I tap on this or hold that it's got a certain physics and like the more you kind of make it what people are used to the kind of the the better the user experience because they don't have to be confused about like learning something new so I think we have to look at like what dimension and like the core principles of like how humans behave is not really changing yeah and sure yes if everyone in the world learns something it's that's like it's hard for them to change that because you know they have a behavioral pattern but if you go back and find something that's even more fundamentally you know intuitive then that will be a better experience even if it feels quote quote new in our technological age so are you of the camp that everything should be a chat interface I would look at like I mean I mean, a chat interface is great for certain things, but it sucks for other things.

14:53Julie Zhuo: Like right now, for example, like speaking of design, right? Like nobody wants to talk out their design. Like if you think about like, how do we understand if a design is good? Like we look at things, like we look at how fashion designers work. Like they've got mood boards, you know? Like they have very visual or very tangible representations that like speak to the inspiration. And here we are being like, well, how should I create a website? Hmm, let me like describe it. Like nobody does that. That's not actually natural in how people, you know, who.

15:23Dara: Like go to the core craft. Exactly. Of the pre-internet.

15:26Julie Zhuo: And so I often look at like, well, what's the most natural way that people have been doing it? Like put aside whether they were doing it on the web or in digital. Like what's the natural way that this would be done? And then how do we get closer and closer to that?

15:37Dara: I like that. So you were at Facebook for 14 years. Pretty much grew up there. Yes. I went through so many cycles of products and apps and experiments. I'm curious what from that era did you learn about how companies win that still applies? And what was a tactic of winning that no longer applies?

15:58Julie Zhuo: Yeah, I think there's so many great gifts that I got from being at Facebook meta now. I think the first one is just kind of boldness of vision. And I think that this is hard, right? Because I think often, you know, and I experienced this as kind of like an immigrant growing up. like, you know, your dreams or how much you believe you can do is kind of limited by maybe what you've seen or what other people expect of you. And so I think it's hard to sometimes like kind of keep pushing and reaching for that. And I think one of the things that was really incredible about working with Mark is that from a, you know, very early point in time, right, he always had this remarkable ability to be like, how do I set the vision just like so much more beyond like what kind of everyone else thinks is like possible or like you know even like aspirational and so when we were I want to say like you know 10 20 million users he was like but wouldn't it be so great if like a billion people used it right I mean it was just like so you know and then when we got closer and it became clear that like a billion was inevitable then like it became like well what about if we like connected and then it became you know AR And there's just like always something new that was like very, very bold and ambitious.

17:15Julie Zhuo: And I kind of feel like that is one of the key things that distinguishes in my mind, like great startups or great companies. It's like you kind of have to go for something really big because you probably will get a lot of what you want. And you have to kind of keep reinventing to make things, you know, because that's where you can attract really great people to work on something. That's where you can actually have the kind of, I want to say, personal growth, right? Like for me or for everyone else who was there on the team to be able to achieve that. So I think that's probably still a very durable principle that I think about.

17:46Julie Zhuo: I think the second thing is that I believe that, you know, just like all of us, we have a very unique personality. I feel like that's true for companies as well. Every company has a personality. And in some ways, I think getting to like top excellence is about like how do you really, you know, you can't really go against that, right? And often, again, it's a personality of the founders or the early set of employees. But it's very hard to win going against what is, I think, a person or a company's true nature. So part of, I think, becoming excellent is you have to know what is the unique nature.

18:27Dara: Leaning into the most authentic version. Exactly, right.

18:29Julie Zhuo: And to the strengths. And, of course, we have to, I think, dole out some of the weaknesses. but often the strengths and the weaknesses are one and the same.

18:38Dara: Yeah. You know what I mean?

18:40Julie Zhuo: And so it's just about how do we then create a structure that allows the company or the group to be able to bet on the things that they're really good at.

18:51Dara: Yeah, it's a parallel outcome. What is the thing that no one else can do?

18:53Julie Zhuo: Exactly. And I think that often it's tough because you're starting a company and you're like, oh, but X is doing this. That's so awesome. We should be more like X. We should be more like X. And that's great. I'm all about like we can all learn. We can all improve. But there has to be, I think, a personal recognition of what is true for the company and what are the things that have taken it to its successes. So I think that's an element of...

19:15Dara: That's a great one.

19:16Julie Zhuo: Yeah. And then I think the third one is just like, look, we tried a lot of things at Facebook, right? And of course, what people remember are the successes. Like, that's the ones that dictate the history. But like, I was there. Like, I saw like hundreds and hundreds of things that like we tried and didn't work. and nobody talks about those because like you know they're like lost in the eons of history but that's a lot of like what it takes like you don't succeed all the time you just kind of have to like fail a lot and be like okay with that and then get back on and then when you've succeeded enough and you look back you're like well yeah like failure so you know like that's a part of the story yeah um but I think it's hard because from the outside like we we just only get exposed to like very successful stories.

19:57Julie Zhuo: And so we kind of have this, like, I guess, impression that like, okay, to be successful, we should just like constantly have wins. But I think the real thing that probably not enough people talks about is like, it's actually the companies that are able to then turn the failures into something bigger that allows them to then like, you know, go back onto their path of winning that actually have a story, right? And then like too often, it's just easy to give up because like you know we're kind of like stuck in a rut rather than like using that as kind of like a stepping stone for it.

20:29Dara: To keep it ready.

20:30Julie Zhuo: Yeah.

20:30Dara: Is there anything you've had to unlearn from that era that you think no longer applies?

20:37Julie Zhuo: Absolutely. I mean I think very a lot of things too like when I started my own company because it goes back to like look I'm not Mark like I was you know I grew up in that environment so clearly a lot of principles and values like I take away but it also required me to kind of really deeply examine what are the things that are different yeah about like who you are how you manage yeah and it wasn't just like we can straight up take these playbooks or tactics um and and that discovery process i think has been like a really awesome part of you know being a like being a startup and being a founder seems like

21:11Dara: that's why you're right to begin with uh you wrote making of a manager while you were feeling imposter syndrome if you're a manager. What was that journey like?

21:21Julie Zhuo: Well, I think that writing for me has always been a kind of therapy, right? It's like a way of trying to make sense of things I find confusing or when there's a lot of noise. And we all know there's a lot of noise. You get on Twitter, you get on wherever, and everyone just has an opinion. And being able to then be like, well, what do I really think about this without just reacting to the last thing I heard was something that I wanted to cultivate. And so that was why I started writing. It's like I wanted to get more comfortable just sitting with different things and then trying to figure out like, what do I really think about all of this?

21:55Julie Zhuo: And I felt that a lot because like, you know, I was surrounded by very, like a lot of high achievement, very excellent people, right? And again, this is like a, and meta culture is, you know, people are very direct. people are like not afraid to like, you know, and that's kind of like the idea that like you put a lot of ideas out there and you kind of, you know, in the marketplace of ideas, like figure out which ones will work or not. And again, a lot of stuff was very experimental as well. Like we did a ton of AB testing, a ton of like, you know, trying to get things in front of actual users.

22:31Julie Zhuo: So there There wasn't a lot of coddling. And so as a result, I was like, and also I started at Meta very young, right? I was like 22 years old, I think at 25. They're like, hey, we're scaling the design team. Go and manage the team. And I barely even knew what design was like three years ago. And so now three years later, I'm being asked to manage and grow designers. And there are people on my team who had been doing design many more years than I had, right? Who are honestly much better designers than me. And so that was where a lot of that imposter syndrome came from. Was like, I had never done this.

23:05Julie Zhuo: I don't feel very prepared. But part of, when I look back now, I'm like, part of what I'm so grateful for is like, I had a chance to kind of do, like, you know, kind of trial by fire. Like, learn by doing. And that, like, and I look back, I'm like, I like, you know, couldn't have paid for like that kind of education, right? Like, and I think I decided to write the book. You know, I, for years, like people were like, oh, you're right. Like, publishers would reach out and they, you know, they like have you thought about writing a book and I was like like I don't know any of this like I'm not like an expert like shouldn't you be like an expert it's kind of like the I am an idiot exactly like you know like a 65 year old like retired person to like have the accumulated wisdom to like write a book but then when I actually spoke to a publisher and you know they were reacting to one of the pieces I wrote for new managers I realized then like that was like kind of my why now moment.

23:56Julie Zhuo: You know, I was like, oh, like, if I wait until I'm 65, I probably will have forgotten what it's like to be a new manager. Because so much of being new is the dealing with your, like, emotions of, like, the imposter syndrome, the, like, the fact that, like, you're doing, you know, this hard thing that's, like, very human for the very first time. Like, giving hard feedback or, like, you know, telling someone that they have to be let go, right? Like, these are very, very difficult things. and they're not going to get easier for anyone if that's your first time doing it. And so the why now for me was like, oh, if I write, yeah, I can probably write a better management book in some respects when I'm older and I have more experience, but I may be able to write the best one now for new managers because I'm not so far removed from that experience.

24:42Dara: Yeah, you can just do things.

24:44Julie Zhuo: Yeah.

24:45Dara: So you've talked about how you're not sure whether the manager role will remain in the way it is today and we'll be managing agents. And I'm curious, like, what skill sets do transfer over to managing agents and what do not?

25:00Julie Zhuo: I think the most important one is what we were talking about earlier, right, which is, like, having a sense of what's important and what success looks like. I actually think this is really hard. And even to this day, like, there are days where I'm like, wait, I'm going, I'm going, I'm doing stuff. But, like, hold on. Like, why am I doing this? Like, what's the kind of greater goal at the end? And even in my work with agents, sometimes I have, like even yesterday, I had the session, like we need a backup. Like I think we've drifted. Like we've kind of like followed, you know, the trail. And then there was always like a breadcrumb that was like exciting.

25:29Julie Zhuo: So like we did that. We did that. But like, wait a second. We've like lost.

25:33Dara: Too many breadcrumbs. Like we have.

25:34Julie Zhuo: We've like we don't know where exactly where like the destination is anymore. And I think that's really important. And it's actually, I think, hard for a lot of people to because often to know what really matters, like for the company or for the team, you kind of have to like have enough time and space to to constantly thinking yeah to kind of think about that right and to kind of keep that bigger picture in mind and because so much is happening you know in the world it like those things often like they're the breadcrumbs you know like they're always kind of pulling our attention so that we're reacting to like this task or this ask or like this email came in and to really have clarity on like this is what is important.

26:12Julie Zhuo: This is what success looks like. And to be able to break it down enough levels, right? Because as a founder, we're like, okay, well, ultimately, we have all these customers, we want to serve them well, we want to make a lot of money. That's the highest order. That's easy. But then the next order, which is like, okay, so what is it that is our unique strength? What can we do better than anyone against a combination of, I guess, structural? What does the company have or the resources, but also like, as I mentioned, it's like related to like the DNA or the personality of the company. And so then you have to kind of come up with like a strategy and then the strategy kind of like has to fit in or like make sense.

26:52Julie Zhuo: And then from there, everything just keeps getting broken down until ultimately you have like tasks. But like I said, it's very easy to get lost in the tasks and forget kind of what is the ladder or like, it's called the tree back up to the higher level goals. I think that's a lot of the reason why often at larger and larger companies or even not that big of ones, it can be hard, right? When you have a team of people because communication is also imperfect, things change. The organization's often not in alignment on what is most important. And especially, I mean, this is very symptomatic of large companies just by nature of the beast.

27:32Julie Zhuo: but like often what like a VP will believe matters and like what a person on the ground understands to matter is like totally different. And so of course there will be tensions and other, you know, inefficiencies that happen. And it doesn't get that much easier with agents unless, you know, like if you're managing a bunch of them because they also will drift and they'll like, you know, they'll like pick up on a thing you said and then be like, oh, well, that's the thing that's important. But like, oh my gosh, like let's go back and like realize this is what matters. And so this goes back to like, you know, people talk a lot about like writing good prompts or like I think a lot of that is just clarity of thought on like how we think about what really matters and how we define what success is.

28:13Julie Zhuo: And like success often is not so measurable, right? Like I, you know, spent many years at Facebook and it was one of the most data oriented cultures. I think one of the great strengths of Facebook is its operational rigor. and a lot of that operational rigor came down to like a very meticulous understanding of like what how to translate things into goals but even that wasn't perfect because not everything can be quantified you know and so I think as a leader you kind of have to hold a lot of these truths in the same time like you have to try to figure out how to make it as tangible as possible because people do work with the nice thing about a tangible goal is like if I say look if we hit this number we're good.

28:54Julie Zhuo: At least that's clear, right? We can all get on the same page about that. So it has the benefit of like, cool, like it's very good for alignment, but it has the, the, the like weakness of being like, well, that's probably still a proxy goal to the true goal. And the true goal is something that is very, very difficult to capture.

29:11Dara: More like a statement or a vibe.

29:13Julie Zhuo: Yeah. And, I mean, there's a lot of, you know, I mean, I think Jeff Bezos used to say this all the time right like when and I see a number and I hear a story or I have an intuition about a customer experience and the customer experience is bad like I tend to trust that because again the number often is a proxy you know for and every number can be gamed like that's the other thing like humans are very smart so tell them this is the number like we can make the number go up but then we may have lost track of like what is that needs to be constraints exactly so I often I like we can we get talking to about data I'd never believe if it's like one match or two of them all, even though I, again, love the practicality of the alignment, I think often we have to look at like a basket of things.

29:54Julie Zhuo: We have to have ideally like one North Star that's like the shorthand. When you have to align hundreds or thousands of people, like it serves its purpose. But often it's not enough. Like you kind of, the people who work closest, you have to kind of look at the basket. And then you also have to have very specific, I think guardrails, right? Things that cannot regress or things that are like explicitly trying to measure the ways that won my game. And even within that, people are creative and clever and there will be ways. There's always ways to game. And agents too, right? Agents will find a way to game.

30:23Dara: So to talk about Sundial, you guys work with companies like OpenAI, Gamma. What are some of the best AI companies doing with data that maybe everyone else is not?

30:33Julie Zhuo: Yeah, it's kind of interesting because I feel like this generation of AI companies, a lot of them, part of it, I think the reason is that we're able to get to scale much faster, right? Like it used to take a lot more time to build everything. And, you know, by the time you got to like a million or two million users, you've already like been at it for a while. And that means you've had enough time to like, I don't know, ask and get those rounds of feedback. I think in some ways, like there's just like right now there's such a pull. So like a lot of companies who provide like, let's say a great service, just instantly see, you know, a ton of growth.

31:10Julie Zhuo: And I always say about data, like if things are growing, like everyone's like, let's pat ourselves on the back. Like we got, you know, like what do we have to learn? Like let's just keep doing what we're doing, right? Even if like, you know, and you can usually what we're doing is just like going by vibes or going by like what feels good. But I do think that at every single point in a company's career, like again, it may not be in the first year, it may not be the second year, it could be many years down the line. There will be a point in which like growth becomes harder. And for Facebook, you know, I think that was like, I want to say 2008 or something, right?

31:42Julie Zhuo: Where like everything was great. We were patting ourselves on the back, you know, Everyone loved our thing, and then one day it was like, whoa, whoa, whoa. It started to flatten. And that's when everyone starts freaking out because they're like, wait, what happened? And then at that point, if you don't have the discipline, you haven't tracked things, then you're just starting to... The first... I mean, everyone always... When things start to not work, the question is why? And then you realize if you haven't actually tracked things and you haven't rigorously cataloged what might be all the reasons, then you don't actually have an answer.

32:12Julie Zhuo: So usually what happens is when things don't go as well as they were, then everybody starts to be like, OK, we got to take this seriously. Then everything gets like logged, everything gets tracked. And then and then from there you can start to, you know, have like, I guess, more discipline, more operational discipline. Now, what I notice is that companies by second time founders who have kind of gone through that, they realize that this is going to happen. So then from day one, they're like a lot more rigorous and disciplined, right? So they make sure that like the logging is good from day one.

32:47Julie Zhuo: They make sure that they have the like telemetry and the observability to sort of understand what's going on. Maybe they don't have to like optimize. They're not in that mode yet, but they try to have all of that ready and they try to keep like a greater pulse. The thing that I think data and all of that gets you is it's just to me, it's sort of like it's more information, you know. And of course, I think about data very loosely. I think of it as anything where you're getting a, you're kind of coming into contact with reality.

33:19Dara: Yeah, maybe social media. Exactly.

33:21Julie Zhuo: And talking to customers is amazing data, right? Because they'll very much tell you what matters, what doesn't. That's an ability. And so if you're a SaaS company, probably talking to your customers gives you even more data than what exactly are they doing and tracking their every clicks in the product. But all of that I think of as data. Specifically when we come to behavioral data on what people are doing in the product, I think obviously consumer apps or ones where you have a larger user base, it kind of matters more because you can't talk to every single one of those people. So from an early period, starting to get a sense of what are different cohorts doing or what are the clusters of what do behaviors look like from power users versus marginal users.

34:11Julie Zhuo: All of that is just helping us as a company learn more and get better contact with reality for what our customers are truly doing or feeling or wanting. And so that's what it is. I think the best companies, it's not even so much like, oh, they're using this tool or they have this. It's more like it's a mindset. It's a mindset of being constantly just obsessed with contact with reality for like what really matters for the customer.

34:39Dara: Yeah, what you described literally happened to us last year. We were Vibes based for the first couple of years and they were like, wait, we need to actually understand our data. So you wrote your daughter about the eye and the hand, the gap between recognizing greatness and being able to produce it. And so if AI is going to get a lot better at being the hand, how are you thinking about, you know, raising your kids and whether they need to suffer through the process of learning how to do it versus just teaching them how to recognize it?

35:11Julie Zhuo: I so I think that this is sort of where one of those chicken and egg problems. I think my personal belief, it's very hard for your eye to be trained without you doing a lot of this. Because so much of how we get better, how we learn naturally is like, you know, I think our brains are wired to like learn more when we make mistakes. So, you know, like if things are going well, okay, sure, we'll like learn some things, but we probably aren't as reflective. When we have like a bad feeling, like emotionally, like we let someone down, we failed, we like feel miserable. well, there's a lot more learning that happens.

35:46Julie Zhuo: And so again, like it or not, like that's, I think, just like, again, how biology is. And so I do feel like such a huge part of like learning is having to do something and having like, have it maybe disappoint you or like not be great or like you didn't win first prize or whatever it is. And then from there, yeah, sure, you might be like, oh, I mean, a little or maybe a lot unhappy, right? But that also is probably what makes you pay more attention. It involves your gut. Yeah, exactly. It's like what happened and the paying attention is like what I think often leads to growth, you know. And I think, again, as much as I would have liked it otherwise, I think when we're too comfortable, we just don't pay as much attention to everything.

36:29Julie Zhuo: Or it takes like a lot, like it's a lot harder to, I don't know, it takes like an amount of willpower, which like I don't think most of us, like it's possible.

36:37Dara: So then how do you think about what to outsource versus what to go through so that you can build that gut?

36:41Julie Zhuo: So I think that I'm also a believer that like even as technologies change, there will always, like, you know, because in some ways like the greater game that we're part of is like a human game, right? And the human game has a lot of elements where there are things are very like zero sum in some ways, right? Like there can only be one goat of basketball or soccer or whatever it is. Like, I mean, yes, like, you know, we can all have like bionic limbs or whatever, but they'll probably still only be one goat, you know? And so I think that there always will be kind of like what is the edge and then the edge will have like you know what does it mean to do that well or better or not right and it just means that the you know it used maybe like people aren't so focused on like most of us are not playing the game of like who can draw the most photo realistic you know person although if that's like the game you want to play by all means like you know and i do think that like again we all have hobbies we can all get better at whatever um we choose to get better at but at least in the economic marketplace like maybe that's not as important of a game with the invention of photography, but there will be the next thing at the edge.

37:43Julie Zhuo: And so maybe if, let's talk about software, maybe if software starts to become a much more competitive field, it's harder. It's not some, we can all build amazing things that probably would've been really hard like 10 or 15 years ago, but now that everyone can do it, standing out still is somehow very hard. Like it doesn't get easier to command attention because that is like a fixed thing that we have. Not to mention, you know, nowadays I feel like all the investors are talking about like hardware and like energy and like all this other massive, you know, and that's cool. Like it's cool that we can now extend our aspirations towards things that maybe 50 years ago had been like way too hard to attempt, right?

38:27Julie Zhuo: And I'm excited about that level of innovation. So I think for my kids, again, I want them to continue to have a thing where they're like, you could say they're suffering or they're like, they're not, you know, they continue to feel disappointed by maybe a little by little, but that pushes them to get better. And I actually think can be anything, you know, in some respects, right? I think if kids play sports and, you know, and they kind of learn like what it means to be very disciplined and how to like get better and better, you know, relatively, maybe they're never going to be like the number one, but I think that that process and the set of learnings and like the pride, not to mention, because so much of also you doing something that's hard and then making some progress confidence yeah it gives us a tremendous amount of pride like and that's where like self-esteem is built that's where confidence is built so i do want my children to find the thing i'm sure it will exist right again it'll be the new edge of where they are and i hope that they will try a thing over and over you know my daughter loves to draw like i don't think again i probably tell her like it's probably not going to be like a viable career to be like a hand animator but i love that she continues to want to get better at the thing and then she can be proud when she gets better at the thing and maybe that can transfer that like general principle can transfer to whatever then she can do anything she

39:44Dara: can get better anything yeah yeah speaking of games uh you've written about the difference between status games and then games that you play for yourself are there any status games or games that you've let go of recently that you feel like are no longer worth playing yeah i think what it's

39:58Julie Zhuo: It's a really interesting question because I think a lot about, like I think a lot about like why do I care about excellence? Like why do I, and you know, maybe there's no answer. But I find myself, like one of the ways I find that, you know, like as maybe a leading indicator to like, who are you really is like, what do you find yourself inspired by? And like sometimes I will read a bunch of stories or like hear about somebody and I'll be like really inspired. And I think that says something, right, about maybe like what I'm drawn to. But I often wonder like why do I care? Like, is it because I want everyone to think I'm awesome?

40:29Julie Zhuo: Like, is it like, you know, like, I mean, I want more love because I can get enough from my parents. Like, you know, whatever it is. I don't know. I don't know that there's an answer to these things. But I do think that one of the things that I said as like an intention for myself is like, I do remember, like, when I was in my teens and I was in high school or middle school, actually, was when I first discovered the internet and we had a science teacher, Ms. Brown, thank you, Ms. Brown, who introduced my best friend and I to building websites in HTML. And I just remember how fun it was, like how excited we were at that moment in time, like just to build stuff.

41:11Julie Zhuo: Like we like our first, the first time we made a link up here, right, it was like, oh my god, you know, it was like so it was so cool. And then I spent most of my, like all the free time I had in high school, just like doing digital illustration and creating websites and redesigning my website and redesigning it. And I feel very nostalgic for those times. I miss those times because I've realized like there was a lot of fun. It's pure joy. It was joy, right? It was like the discovery. And I think a lot of when we were kids, like that happened a lot. You know, we did things because they were fun and not because it was like suddenly a thing, you know, for the resume so that we could get a job so that we could like earn enough money so that people like think we're cool you know there's and then a lot of that you know gets like you say other influences come and i don't want to say that they're like good or bad i think that they're like a fact of life um but i i did find myself kind of maybe wanting to go back a little bit more to that energy and that ethos like yeah how do we think about ai for fun in fact right now i feel like so much of the story about AI is very dismal.

42:18Julie Zhuo: It's a very subtractive story. It's like, hey guys, we can do the same thing that we were doing, but with less people. Efficiency. Yeah, with less cost. Get rid of humans. It's such a sad story if you think about it. That's not at all inspiring. No. I don't think there's enough. I mean, occasionally you'll see people on Twitter, and I love that these things get viral, but I don't think there's enough talk about how can AI make our lives fun? Like, how can we, like, have, like, be playful with it? How can we, like, do things just for, like, me? Like, not because I'm going to go put it anywhere or, like, it's going to be a cool story, but it's, like, or it's going to help my company.

42:55Julie Zhuo: But just because, like, it's fun and I would, like, want to do it for, with my kids or my family and, like, no one else has to see it. And I'm all about, like, AI for joy right now. Like, as a principal.

43:06Dara: Me too. Last question. After this episode ends, people are going to continue the conversation with your Delphi. What's one question you hope they ask?

43:17Julie Zhuo: I hope they ask questions about like, I don't know, something along the lines of like, how can I get to better understanding myself?

43:25Dara: Yeah, that seems to be a big theme in your life.

43:27Julie Zhuo: Yeah. I mean, I ask, honestly, like I talk to Claude all the time. I think it's very interesting.

43:34Dara: Amazing. Thank you so much. This is great.

43:35Julie Zhuo: Thank you. Thanks for having me. Thanks so much.

43:40Dara: Thanks for tuning in. If you enjoyed today's episode and want personalized advice from Julie, head to the link in the comments or description to ask her digital mind on Delphi.