Overview
Nikhil Singhal hosts three product leaders from his Skip Community: Dana, who runs core product at Harvey; JZ, chief product officer (CPO) at Laurel; and Hemal, who recently moved from OpenAI to Google. The first half covers how they handle young kids, demanding jobs and side projects. The second half covers how product managers (PMs) without AI backgrounds landed top AI roles, and how to pick a company now.
Key Takeaways
You don't need AI credentials to get an AI job. Dana pushes back on the idea that PMs must be steeped in AI to get hired. She sees junior PMs use prototyping tools to fiddle with buttons, colors and fonts until they've lost track of the problem they set out to solve. She hires for first-principles thinking and closeness to users, because anyone smart can pick up the tools. Hemal adds that the PM rosters at OpenAI, Anthropic and Google are full of people who were strong craftspeople at top companies years ago.
Hemal's three core skills. A great PM right now should be a historian (why things are the way they are), a technologist (understand complex systems and explain them simply to researchers and customers alike), and an anthropologist (study a domain like law deeply enough to describe it to whoever codifies it in a model). JZ adds curiosity and comfort with change, since a research breakthrough can wipe out a roadmap overnight.
Joy as protection against burnout. JZ argues that constantly weighing whether each activity is worth the time is what exhausts people. Her recent sprint at Laurel, with a major law conference, a customer summit, an offsite and a platform launch, was the hardest she's worked and the most fun she's had. All three agree you can't have it all. Hemal frames it as ruthless prioritization around what you'd regret not doing.
Pick companies on data, not product category. JZ joined a timekeeping company because large, security-conscious law firms trust Laurel with their data. She says firms can use that data to make an open-source model 20% more efficient and depend less on frontier labs. Hemal looks at four things: proprietary data, distribution, strong people, and energy. He reads junior employees complaining about shifting strategy as a good sign, and finds leadership that sounds too confident a turnoff.
Practical Steps
- Take a non-PM role at an AI-native company if that's what's open. JZ recommends forward-deployed (customer-embedded) or AI operations roles, saying two months inside an AI-native company teaches more than staying at a slow-growing SaaS business.
- For early-career candidates, Nikhil suggests skipping the famous names and joining an unknown startup that will pile work on you. The vocabulary comes with the job.
- Stay current through people rather than feeds. Hemal keeps a short list: a friend who tries every product, AI researchers who explain breakthroughs, and industry contacts who report whether new models actually matter to them.
- Keep a fixed set of test prompts so you can compare model performance over time.
- When evaluating a job, ask how customers use and feel about the product, not about valuation.
- Build a reusable system for recurring work. JZ built a "teaching brain" that pulls recent news and updates her course materials from voice notes.
Notable Quotes
"I actually totally disagree with the notion that you have to be AI-filled to get an AI job." (Dana)
"Yes, it was the hardest I ever worked, 100%, but it was also the most fun I ever had... And that's where I feel like you don't burn out. It gives you energy." (JZ)
"I'm like, Are you a VC? Have you asked, like, how do customers, like, use the product? Do they love it or not?" (a panelist, on candidates who fixate on valuations)
Full Transcript
I actually totally disagree with the notion that you have to be AI-filled to get an AI job. I run the core product at Harvey, and what I found is, like, we have all of these amazing tools where PMs can go out and prototype. What that means is, like, more junior PMs, they'll just start prototyping and adding buttons and changing colors and changing fonts, and they'll get so far away from what the problem is that they're trying to solve, they've lost, like, their first principles thinking. I always find it interesting that when people also look for jobs, they, like, don't remember the job that they're applying for. And so what I mean by that is you have, like, I have a bunch of people who tell me all these things where they're like, Yeah, I just don't know if this company at this valuation and everything makes sense, and I'm like, Are you a VC? Have you asked, like, how do customers, like, use the product? Do they love it or not? We just came from a crazy sprint at Laurel where we had, you know, the biggest law conference ever. We had our customer summit. We had our company offsite. We turned our product into an entire platform. And yes, it was the hardest I ever worked, 100%, but it was also the most fun I ever had, yes. And that's where I feel like you don't burn out. It gives you energy. Hi everyone, I'm Nikhil Singhal, and welcome to another episode of the Skip Podcast. If you don't recognize my surroundings, it's because you're a regular and you don't recognize me not being in my basement. I actually am at a studio, taking advantage of a new format that I'm experimenting with. Now, many of you know that I have a product executive community called the Skip Community, and that group of 125 heads of product and product executives help each other and help all of you try to advance in your career. So I've welcomed Dana, who's at Harvey, Jay Z, who's CPO of a company called Laurel, and Hemal, who recently joined Google Cloud for OpenAI. I've welcomed them to come to the show and talk through, you know, how do they balance home and work and staying current in the age of AI? You know, I know that that's a topic that we've talked about, but I was able to get some really key insights around the importance of joy and how to drive you forward when you feel really exhausted. And then the second thing I really wanted to push on is how to navigate the job market these days. How do they choose the roles that they landed in? How do they make the decision on going to an AI company versus a non-AI company? And ultimately, how do they go from being non-AI product managers to being AI product managers? It's a terrific discussion. I love the connection. It's a format that I'm going to try to take advantage of more and more, and I hope you enjoy it as well as I enjoyed having it come together. So without further ado, just a word from our sponsors, and we'll get right into it. This episode is brought to you by Firecrawl, which gives your AI agents data from the web and beyond. Your agents can search and scrape the web while accessing Alexandria's rapidly expanding library of over 100 million index sources and more than 200 data providers, all through one connection. Connect Firecrawl to ChatGPT or Claude to find people and their work emails, pick your fantasy football lineup. For analyze trends for your next marketing campaign. Firecrawl is building the library for superintelligence, bringing more of the world's knowledge into the tools you already use. Go to firecrawl.dev/skip to get started. Hi, everyone. Welcome to another episode of the Skip Podcast, a podcast for your careers. And, you know, three of my closest friends are here, Jay-Z, Hemal, Dana, and this is a new format for us. This is a new way of us trying to get the stories behind the stories and what it's like to work in this crazy industry that we're in. Before we get started, you know, I'm curious, Hemal, you have a newborn on the way. Yeah. The two of you have two young kids. What gift do you think new parents need that don't normally, normally show up on a registry that you're like, this actually was the thing that I needed the most as a young couple? And why is it wine? The number one thing I say, and it's been the same thing, is the snoo. So the snoo will rock your child to sleep, which is a gift. Do they make that in adult size? Would you like that? Yes. A hundred percent. You will not need rocking to sleep because— I think it's called Ambien. But I think you can't give that to the kids. Every second that you can sleep, you will fall asleep. The three of you are, like I said, close, dear friends, part of my community, and, you know, we're at the top of your careers, which makes me super, super happy to spend some time listening and learning from all of you. You know, Jay-Z, you're at Laurel now a year in. Yes. And, Hemal, you're, what, four months in? About five. Yeah. And so both of you made recent job changes, and Harvey, you're like six. Yeah, look at that. So, and, you know, the fact that you all three are in, let's call it AI roles at very AI kind of leaning companies. But, you know, in some ways, you know, what we're going to explore today is just how did you get those positions? What makes you qualified to work for these AI roles and positions? And what's your advice to folks that are kind of trying to navigate this crazy career? But I kind of want to start with the personal side because I feel like that's the story we don't hear as much on these podcasts. So, you know, you folks have all been building, and, you know, that's one of my favorite parts is all three of you are super, super hands-on and tell me all this crazy stuff you're building. What's your favorite thing that you've built at home, you know, to help you, your partner, your kids? You know, what have you been doing that's really helpful on the AI side? Maybe I'll start because my daughter—I have two kids, and mine's a senior in high school, and we're starting to apply to schools. And, you know, you hire these college counselors, and they tell you all of the schools that your kid is designed for, et cetera. And, you know, college, I don't know what it was like when you folks were seniors, but it's very, very hard to navigate. So what I did is I got a set of her criteria, and then I did a search, and then I built, like, this really comprehensive, like, spreadsheet, which essentially has a ranked list of every school she should apply to, what were the pros and cons, what were the application process. Because for high school seniors, they're like, I don't want to add that school. And I'm like, Well, why? And they're like, Well, because it's another essay I have to write. So then you write, like, a program to figure out what the school really wants, and then you combine it with what she already has to offer, so you give them prompts. So essentially, you can systematize the entire college application process so that it makes it a lot more understandable, appealing, et cetera, without, you know, cheating or any of that kind of stuff. So that was a huge help for the whole family. All the anxiety in our house is down right now as we've been able to use, like, these crazy, crazy— That sounds amazing. Yeah. Yeah, I should probably sell it. Yeah. My mom is actually a college counselor, so I feel like we— She's AI? She does, but, like, not like that. We should talk. I would love to. So what have you built that you just love that you still use? I'm going to be honest. Most of what I built were, like, mini agent watchers. Okay. So they were watching for reservations at certain restaurants I wanted to go to. They were watching for my daughter's swim lesson availability to open up. They were watching for, like, little things that I was manually checking, and all of them have been replaced by Instinct and Muse and Grokbots, which I'm playing with all of them. So I don't need to build my own custom watchers anymore. So, like, the things that I loved the most, and my mom friends were like, Can I get— can you tell me when swim lessons are open? Like, I'm like, Just download Muse. That's great. Isn't that awesome, though? Like, you build something and then, like, three months later, somebody else is maintaining it. Yeah. Do you see that as a bug or a feature? Like, I think it's actually cool when I stop having to build my custom thing, but I have an understanding. It's validating. Yeah, it's validating what I was doing. Like, if I waited a little bit longer, I wouldn't have had to go through the pain of building it myself. But yeah, I love it. There's always better form factor, or there's, like, something that somebody else has thought of to extend it. So it's one validating, but always generally pretty cool. Are you going to build a wine app, Emma? Oh, no, for many reasons, but that's a different topic. Mine was actually, like, we had a baby shower and, like, had an agent that managed the entire thing, gave it an Excel spreadsheet and gave it access to just my emails for, like, RSVPs and said, Here are the people that we want to go and hire and what I want it to have. And honestly, it was, like, very low stress to put it together. It owned the communication, everything. And, like, I just, I would get the emails for payments and I would just authorize the payments. But coordination and kind of, like, what's the run of show and scheduling, it really took care of. And we would give it the parameters of, Here's the times, here's how we kind of want the thing to work. It would follow up with different people, and it was, like, a really easy way to actually throw an event. How did your partner, how did she interact with it? Because, you know, she's— She has no idea with an agent. She thought it was all me. That's how you get bonus points. Yeah. Smart. Good PMing. Yeah. Yeah. Jay-Z, what have you been building? I had the same experience where you basically, you were struggling because, you know, 12 months ago you'd be like, Oh, I'm going to do this on OpenClaw, and you're like, Ooh, I just, like, did something in my terminal and I got confused. And all of it is just out of the box with Instinct and Muse. But I've always wanted to build in the education space and in the, you know, children toy space, mainly because, you know, I have such a passion for teaching. I build things with my son. So we go into Claude Code and we build games together, and we've evolved that. And now I've actually turned it into, like, a real thing. There are two things that I built with him that he loved: Treasure World, and it's all about math. It essentially teaches, you know, him and now his younger brother math concepts, because math is a hard thing to get someone to sit down and want to do. At the same time, they have these little figurines that they love, and we would tell stories about the figurines. And every single night, every single night, they would be like, I want a story about the grave digger. Like Monster Truck, and I want it with some random thing that does not live in the Monster Truck universe. It'd be like, I want it with a dragon, right? And so what I ended up doing is I created this thing called Story World, where it has all these little figurines, and it has memory and stories behind it. Now my older son knows how to read. My younger son doesn't, but they are able to literally talk to this app, and it will listen, and it will remember and give it stories. This is really interesting because I did build two apps for my daughter that are, like, very similar to this. One of them was a storytelling app where she has—I started when she was, like, four and has, like, various fears. One of them was of swimming. And so I wanted to build, I wanted to, like, write a book about Hannah, the girl who is afraid to get in the pool, and, like, have her picture kind of look like a Pixar-esque version of her and show her, like, overcoming her fear. So, like, definitely something in storytelling. And then the other one, like, the interactive one I built, I was like, okay, I've never built an iOS app. I should learn how to build an iOS app. And, like, kids are really challenging, and food is one of the things that's really challenging. Like, she won't eat dinner. And so I was like, if I can build an app that builds our refrigerator into, like, really cool icons, like food icons, and she can then choose one main, and you have to choose one vegetable, and you have to choose one fruit, and then when she makes her plate, like, does she then feel like she had control over creating that meal and gets to eat it? And I would say it had, like, a 20% increase in her ability, which is, like— How do you both find time for this? Like, you both have young kids. You know, Emily, you have one on the way. You all work kind of, you know, more than, let's say, 40 hours a week, for the sake of argument. You know, Jay-Z, in your case, you're teaching both Reforge as well as Yale. You've got two young kids. You've got a remodel going on at home. You've got aging parents. You know, the sandwich generation you all three of you are living. I just don't know how you have these crazy jobs where you're head of product at a really hot company and make it all work. And is there, like, a pro tip that you could share that all of us can, like, find the sort of spigot of time that none of us seem to be able to have? There's a book I really like, and we're reading it actually as a team, Angela Duckworth's book around basically grit and how very successful people are incredibly gritty. And the two big factors are basically perseverance and passion, right? And so I think it's really that passion piece. For anything that I do, I have to find an incredible amount of joy in it, especially if it's anything outside of work. So I love the physical space. So doing a remodel is basically designed of— of physical spaces and it's just a passion of mine, and something that I've held for a long time. And so remodeling doesn't feel like a chore. I mean, honestly, I used to flip houses for fun when I worked at Airbnb. And so it's like, it's truly just a passion, right? The other thing I would say is AI. AI literally makes us able to do all the things. So, yeah, this, you know, every year I teach a semester at Yale, a quarter at Stanford, and I do Reforge. And before, the idea of updating all of your materials, being able to do all those things was incredibly difficult. I've essentially built a teaching brain, and I have wonderful co-teachers, Anand, Nikki, et cetera. And so we collaborate together on our materials, and we are able to say, hey, how do we, you know, again, content updating used to take forever. The very first time that Anand and I built our curriculum for Stanford, it took us months to build our curriculum by hand. And today we can actually feed it what we know and what we do best, and we can voice note it, we can talk to it, and it can update. It can actually pull the latest news and suggest to us, would you want us to incorporate something along these lines? What's your opinion on this new development? And so when you build that operating system, I'm able to update all of our content all at once. And so this year, even though I'm teaching more than I've ever taught before, it's taking me a lot less time, and I'm able to refresh. We just released a brand new updated AI product leadership course. We completely revamped our product foundations course to actually be super AI-native with Reforge. We also released an updated version with Stanford and Yale's new. And so you're able to do all these things because of AI, which is why I find AI so exciting, because what a time to be alive, to be able to amplify your leverage. But don't you feel, and I think all of us have this sort of, if I was doing fewer things, like I sometimes when people come to me and they say, hey, I'm really struggling to keep it all together, you know, I sort of say, hey, at this generation, you know, at this age, at this stage, your goal is to equally dissatisfy everybody in your life, not one in particular. And yet, in your example, if you had spent less time in, you know, teaching, maybe you could spend more time with work or more time on the— With work or more time on the remodel, like, do you don't feel that trade-off, or do you feel like it's the combination of all these that kind of create the joy that really makes it exciting to do any of that work? I'm just trying to crack the code. The trade-off is very intentional, and there are a lot of things I say no to, to be very clear. For example, I only, you know, work with companies that I really love, for example. And right now I'm not spending a ton of time investing like I used to, all of those things, mainly because I want to protect the areas that I want to deeply invest in now. So there are a lot of trade-offs where you say, Hey, that was a previous part of my life. That is not going to be the main focus of my life now. The specific calculus here is I actually think that all of these things are very much reinforcing each other, and I'll talk about teaching and work really quickly. I've found that in today's age, I'm really looking, when it comes to building out a team, for very experienced people who are much more of the systems creators and the architects. And then I'm looking for incredibly high potential, talented young people. And so the reason I actually said yes to teaching at Yale, when I have more than enough stuff to do, is because I think that, because the access to talent is really unprecedented. And so for me to be able to say, you know, people overlook Yale because it's in New Haven, it's a liberal arts school, but they have incredibly talented people. And today, when anyone can build, right, like what I'm doing with Adventure World, I could never have done two years ago, and I went to Yale. And so again, there's also like some— Who is overlooking Yale? Yeah, there are a lot of people. No, for tech talent. I think that the point is the tech talent. Yeah, for sure. Like, you don't think of Yale as like, Oh, I'm going to go find my first engineer. I went to UConn, so screw Yale, but yeah, that's fine. I personally think that—I agree with what you were just saying—is that I don't think you can have it all. I think it's like a fallacy. Yeah, it's like you have to choose. Yeah, it's like, I want to be, like, incredible at my job. I want to make tons of money. I want to, like, work out and be, like, super ripped. I want to have, like, all these hobbies. The working out is not happening. Yeah, exactly. Like, yeah, I want to have all these friends and all these socialized, and I want to, like, help with mental—like, there's no way you can do it all. No way. And I think that sometimes I don't think that we all give ourselves the grace to remember that, like, you make intentional trade-offs. That's okay. When you say no to a friend, like, feels bad, but that's okay. And that's something that I've struggled with a lot too, because, like, I always want to be at, like, inbox zero on my personal device and respond to every message and take, like, every conversation that you can and have time for hobbies and, you know, spend my weekends entirely devoted to, like, family and all that. You can't. But then the question, I guess, is for you: how do you prioritize? I mean, you have to have a list of what are you optimizing for, and in a world of, like, what are you not going to regret? And I do think right now, like, it is the best time to build. It is a very, very good time to build and focus on work. That's not always true. In different eras and in different times, there were scaling and different types of problems. But, like, we live in a world right now where, like, the technology is the most interesting probably of our, at least, generation. And so I just think that if you, for me, that trade-off was I'm willing to sacrifice a lot to optimize for that. So, Dana, you know, when you moved this year and you had been at Airbnb for over half a dozen years, so let's—I think it's safe to say that you were very much in flow. Like, you knew that product. You were passionate about the product. You knew the team. You had a relationship with your manager, with the leadership team. I think that it was easier for you to get big things done because you had really helped establish that place in some ways from the product perspective. Then you decided, with young kids, to take on a crazy company, Harvey, growing super fast, brand new product team that was essentially being reestablished, and learn something totally different, you know, in person, you know, as a big piece of that company. And so it was really a very huge change in your life at a time when, you know, you had young kids and you have a partner who's very busy as well and working as a tech exec. How did you come to that conclusion? How much wine did you drink to get there? Like, as you're saying it to me, I'm like, well, I'm crazy. Like, actually, that's— But you're not. You're not crazy. You are working hard, but I get the sense that you're making it work and you're happy. Yeah, yeah. It just felt like time. Like it just, I felt like I had reached a ceiling at Airbnb of learning, and there was so much more to learn in the AI space. I mean, building for lawyers versus Airbnb hosts, it couldn't be more different. Going to enterprise versus consumer. Like I just felt that now is as good a time as ever. My youngest daughter had turned one, and I was like, we're out of the woods, which turns out to not be true. Terrible twos. Yeah, it's rough. But I, I don't know. It just, it felt like the right time. You know, over the course of seven years at Airbnb, I'd gone through various recruiting cycles and ended with different job offers, and it just never felt right. And I was just so excited about everything that I was building, and I wanted to learn new stuff. So I don't know. It was timing and just something in my gut. How do you kind of rationalize what I would consider to be all three of you somewhat irrational decisions to make job changes when you have so much going on in the rest of your lives? I go back to, like, whenever you want to go and study, like, any field, take athletics. It's an easy one. And you go look at, like, the best of the best who are out there. You only see one dimension of that extreme. You only see, for example, Super Bowl rings, or you only see, like, obscene amounts of money that people earn. You don't see the rest of their lives and all the consequential decisions that come out of those choices, whether it's strained relationships with families or, you know, very lonely, you know, or personality types of quirks that come from those types of things, or being kind of one-trick ponies. And then post-retirement, the best that they can hope for is a career with Fox. And ironically, we all sit up here where the skill that we're supposed to be the best at is, ironically, prioritization. And so, again, I go back to if you are not ruthlessly prioritizing for one thing, and you are not saying, what is one thing I would regret not doing with my time right now? Because that is all you have. That, to me, actually feels like the more painful decision to make, is trying to have it all and not being exceptionally intentionally clear, is like, what are you ruthlessly optimizing for right now? Now, you have a young one on the way relatively soon, the next month or so, and it's your first. Congratulations. Thank you. Have you thought about what you're ruthlessly prioritizing away? Oh, man. I mean, honestly, I think it is supporting my wife. Like, the honest truth of how her household works is I am a one-trick pony. I wake up, I walk down to the office or the car to go to the office, and that's about it. And I'm very grateful that she's actually able to keep everything else together while she's also an emergency veterinarian, like, full-time too. So it's not like, you know, she has her balancing act as well. But the thing that I need to make sure of is, like, do we have all the right support, like, infrastructure in place at home, whether it's like a night nanny or some help or things like that, and that she has what she needs to kind of keep, like, the things going overall. And that's just the honest truth around it. I do worry a lot, like— I worry a lot about things like, hey, do I have to pull back from the work hours? And if I don't, am I going to be like a present dad? Is my child actually going to get to know me in the early years? I worry about if I pull back, is that going to mean I'm going to be able to put in as many hours? And I don't usually think I'm the smartest person. I'll probably work as hard as anybody else, but I worry about those things. But I've also learned from, like, really great parents, I think, that kids really help you on two dimensions, like professionally, which is you have purpose. You, like, really know what a lot of this is for. And two, you are a lot more efficient with your time in other areas because you know the opportunity cost from it. Like, I'm curious for all three of you, but that's at least what's the blind advice I'm operating with. Yeah, I mean, I'm curious around, what do you guys, do you think you were better pros after had kids? A hundred percent. But I also think, like, when I think, you know, the act of prioritization at work is very different than at home, and that I think at work it's generally like quarterly planning or, you know, yearly planning, and you, like, have a stack rank of priorities, whereas at home I'm constantly reprioritizing. Like every single day I'm making a decision about what the priority is. I got asked to do a talk that's on, you know, my baby's first Halloween parade, and the talk would be, like, a really good thing for my career, but my baby's only going to have one first Halloween parade. So it's like you're making these decisions every single day, and I think that it's hard to, like, prioritize from, like, a planning perspective, and it's just what feels right on any given day is how I do it. Were you a better professional after you had kids? Yeah. I mean, I think, one, I realized the amount of pain that I could go through and survive. And, you know, I was at Webflow when I had both my children. I took maybe sub two months of mat leave after each of them just because the business needed it, and it was important for me to be there for my team, and honestly, I cared. And I think for me the biggest thing is joy, right? So if I'm spending all of my time thinking about, is this worth the time? Is that worth the time? That is exhausting, and that to me is what leads to burnout, when you're constantly analyzing, thinking. When you are in flow state, you love your work. You're, like, work— My CEO, Ryan and I both believe this really deeply, which is, you know, you basically want to be somewhere where work feels like play. And, like, you're having so much fun that you don't know that you are working. And that, to me, is when you can do your best work. Like, I don't know when I'm working when I'm teaching. I could spend hours coaching students, and I feel completely—like, in fact, I feel more recharged because I feel content. You know, I love what I do at work. I think there's so much interesting intellectual challenges about what we can do at Laurel. It is such a great place. Like, it doesn't feel like work to me. And so that's why I feel like you need to figure out how can you be in flow state. What are the things where you have so much joy around it that it no longer feels like a calculus in your head all the time? And then, of course, the trade-offs of, hey, I have this really important customer meeting. I can't go to this performance. And so, you know what? My husband's going to go. He's going to film it. And then I'm going to sit at home when my son is home. We're going to watch it together, and we're going to talk about it. Like, those are the things that you just have to do to make it work. You know, I was actually—I'm curious to kind of get what you guys do, but I also found that celebrating really matters. There's, like, two lessons that I learned. Like, one, you celebrate the small things, whatever they are, and you, like, really be present and live in the moment. But two, I also think that we also forget that, like, humans are meant to do really hard things. Sometimes I think we forget that, you know, hey, these choices that we make, the hours that we work, sometimes there's, like, really substantial payoffs at the end, whether that's literal or just, like, an exceptional career, like additive job that might not have paid off in that moment, but it really helps you get to that next thing or some skill that you feel like you learned. I don't know whether it's just celebrating or remember that we're supposed to go through hard things. Sometimes feel like we're, like, more instant gratification culture nowadays, especially with the next generation. I do think that joy is one of those things that when I stopped working full time, I work the same number of hours, and, you know, and I keep myself similarly busy, but there's a lighterness—if that's not really a word—but there's a lightness to the energy. And I can, my kids can tell. Like, I had this conversation this week. You know, daughter was filling out applications, and then they said, Well, Dad, you're retired, right? And I'm like—like, I was mortified, and I, like, ran around the kitchen. I'm like, I am not retired. I work harder than I ever— But they're like, Well, it just seems like it because you're just, you so much— You're enjoying yourself. You're just enjoying yourself. And I do think that there's a lot of wisdom in what you're both saying, is that, you know, we're supposed to enjoy the time we spend, even if it's hard. Yes. Yeah. In fact, I honestly think my most memorable times are when you do something incredibly hard. Yeah. You come out the other side and you're like, Yes, that launch went well. We just came from a crazy sprint at Laurel, where we had, you know, the biggest law conference ever. We had our customer summit. We had our company offsite. We shipped—we turned our product into an entire platform. And yes, it was the hardest I ever worked, 100%, but it was also the most fun I ever had. Yes. And that's where I feel like you don't burn out. It gives you energy. Stripe is built for whatever comes next. Your business is moving fast. Your banking should too. Stripe Treasury is a new financial operating system built for what's next. With Stripe Treasury, the place your business accepts payments is also where you can store balances, move funds, and pay vendors worldwide. When you're operating across markets, your money needs to move when and where your business does. Stripe Treasury gives you a finance stack that's automated, intelligent, and always on. Stripe Treasury is business banking on the same platform millions of businesses already use to get paid. Learn more at stripe.com/slash treasury. Flash Treasury. Stripe is not a bank. Banking services provided by Fifth Third Bank, National Association, member FDIC. I want to switch gears for a second because, you know, Jay-Z, you're a CPO at Laurel. Well, you went from OpenAI to Google, but prior to that was not in an AI company. And Dana, you obviously are at Harvey now in an AI role. Now, a lot of people that I talk to, they are nervous and they desperately love to be in AI companies. Like, that was what they would love to do. But the idea of being in an AI company, they're like, Oh, who has the time to be AI credentialed? And so the question, I guess, is the three of you, I know you well, you didn't grow up as AI researchers. You have learned the tech as all of the tech has advanced. I mean, five years ago, AI wasn't really a thing. So my question is, you're in three of the plummest AI roles at, you know, hot AI companies, yet, you know, somehow there was some transformation that took place in some relatively short period of time where you convinced a set of companies that, hey, this person really knows what they're talking about, even though we're at the cutting edge of AI or we want someone to transform us to be AI specific. What did you do to prep for that? How did you convince others? And the reason I'm asking is— I begged. Not just me. I said the word agents like 50 times in my interview. The thing is that somehow it worked because a ton of people are like, Oh, if I knew how to change myself, it would be a totally different way of, like, my career advancing. Like, I want to go to Silicon Valley and go to an AI company. No one will—they'll never—I don't have time to do the research. Oh, I'm going to take the MIT class and I'm— You know, what did you guys do? I actually totally disagree with the notion that you have to be AI-abled to get an AI job. So I run the core product at Harvey, and what I found is, like, we have all of these amazing tools where PMs can go out and prototype. What that means is, like, more junior PMs, they'll just start prototyping and adding buttons and changing colors and changing fonts, and they'll get so far away from what the problem is that they're trying to solve, they've lost, like, their first principles thinking. And so when I'm interviewing people, I'm actually looking for just, like, really great fundamental— Product managers who understand how to think in first principles, because anybody can learn these tools. Like, if you are smart, if you know how to get close to a user, if you understand, like, how to solve for users first, AI is not hard to learn. Like, I think that's the beauty of it, is you just go talk to a thing and it tells you how to do the thing. And so I think that, you know, having product craft is still extremely valuable, and those are the people that I look for when hiring. And I think that's a large part of why I was hired, because I had been building consumer products for a while, and Harvey was in need of, you know, building an end user product and looking for somebody who, like, had that product craft. I did say agents like 15 times in my interviews also, but I don't know. I think that, like, you can get really far away from first principles using AI too much. It turns out you do all this great research in these labs, and then you have to actually make it usable and useful to the population. And it's funny, again, I teach, right, and it's all the later stage product stuff changes all the time, like what tools you're using, you know, how quickly should we be building, et cetera. But the fundamentals are the same. Like, do not jump to the solution before you understand the problem space, right? Who your user is, what your business needs. And I think Claire said this recently, but it's like, congratulations, we can ship a lot faster. It just means we've shipped all the stuff, all the garbage that we can possibly think of. And if you don't have the principles correct, you will never find your way through the woods and actually build something valuable for users. I honestly think we have so much crappier products now, ironically, but yes. I think the three skills that you need to have, it's less of this whole, like, you have to use everything and be up to date on the latest and be able to, like, you know, build your own model on the weekend, like maybe. But I just think that the, like, three core skills of being a great PM right now are, like, be a great historian, right? Like, understand why something is the way it is from the past and where it is today, and potentially use that to help predict the future. Be a good technologist so that you can understand very complex things, but you can also explain them in very simple... Very simple ways, whether that's to the AI researcher as much as your customer, and then be a really good anthropologist so that if you actually study somebody who's never been a lawyer but can go really deep on understanding why does law work the way it does without having to go. Yeah, exactly, and be able to describe that to somebody else who actually is then going to try to codify that in a model. Those are like the fundamental three skills. And every time I see PMs who tell me that they've missed the mark, it's because they didn't have either first principles thinking, and they're just trying to show how many courses that they've taken or what they're building on the side projects for their weekend and like, look at my OpenClass setup. And so I just think that's wrong. You know, it's interesting because you folks are saying like, hey, 80% of this is the same. The outside is very different, but the inside remains the same. There's some foundational truths. When you were at OpenAI, you know, the conversations that we typically have when we talk to folks in the AI labs, in particular frontier labs, is they'll say, look, everything's super different. You know, the technology comes from the research. We don't ship products based on what people are asking for, but the technology is going to go, et cetera, et cetera. They're talking about PM being quite different in some ways. But now, as you look back of your experience and you have more experience with other frontier labs as well, do you think we're moving to a world where things are more the same, or are there foundational differences in how we see product working that will just change the way that both people are skilled as well as who will apply in the market? I think the way that you think about building product, like to that point, the inputs of what you actually weight now of like, okay, well, here's where users are today versus here's where we think model capabilities are going to be in three months. Like what you go and weight in terms of pixels versus, again, kind of thinking about what can the model and the harness go do, those I think are different. But, you know, if you just go and open up the roster for the PM team at OpenAI, at Anthropic, at Google, like— A lot of the folks that you will see there were a lot of the folks that were at top companies, you know, five years ago, ten years ago. They were very strong, skilled product engineering leaders. And so, again, it continues to reinforce what I think is if you are a strong, like, craftsperson at its foundation, you should be able to get into these types of companies. And so just working on a strong craft is, like, my always quintessential advice, even if the inputs change. I think the two traits that have become even more important is curiosity and being able, essentially, to adapt to change. So the reality is change happens all the time. So being able to be flexible, to be like, hey, this new, you know, for example, you're in a lab and the research team comes up with something new. You need to be able to respond to that, right? So, like, that change, responding to change is really important. And then that curiosity is also really important. You need to embrace—part of being able to embrace change is having a curious, like, mindset. And so you don't want to feel like, Oh, I have my plans. I have my roadmap, and this new thing came in and it destroyed my roadmap. You should be excited and curious around, like, okay, well, what can this unlock? So I think those are the two things, like curiosity leading to you being more flexible in the face of change. How do you, you know, a lot of what you're describing is, look, things are changing so quickly, you got to stay on top of it. If you don't, it's going to be hard for you to put that in your product. It's going to be hard for you to be relevant. How do you stay current, you know, when so many things are changing? Like, I feel like I can stay current more easily, but I don't have a day job. Like, literally all day, every day, I can play with all these products. But, you know, Denny, you described, you know, Instinct, Grokbot, Muse. I mean, those products are like 50 days old in terms of, like, being in the common vernacular. You're talking about models that come out. I'm sure that as soon as the models come out, you're right on top of figuring out, hey, what do we need to remove? What do we need to add? Like, it feels like almost like an impossible thing, and it feels like the pace is even picking up more and more. But yet you can't really build on the side and hack while you have the day job and the children. Like, I don't quite follow how you guys stay current. For me, it is the job. Like being at an AI native company, like, you know, we release models within a few days of them being out. We get early access to testing them. Since I own the core product, it's all agents. So whenever a new agent product comes out, I'm using it. I'm using Muse and Instinct and Grokbots and, like, trying to figure out what I like best and what doesn't work and what could work for lawyers. And so I think it just comes naturally with the job for me. I'm not building as much personal stuff on the side, but I think that staying up with the technology is the job. Yeah, I see. But what about for you folks? It's honestly a weird answer, which is it's like people curation. I have, like, a group of people that I can go and rely on when there's, like, a change, and I know, like, so-and-so is really, really strong at trying everything, and they're going to comment on the paradigms of how you actually engage with it. I have AI research friends who are really good at when I, like, what is, like, this new way of coming out? What happened to LLMs? What is, like, oh, well, explain how that is and how that could actually go and work. And then I also have, like, industry friends who I go and ask, are you using Frontier Labs still? Like, have you actually cared about the jump from, you know, five to six or whatever it is? Why? Why not? And so rather than me trying to go and consume everything and piece together the picture, my system has been very much people curation. And then so when there's a change, I try to pick up the phone, who can I talk to, and then try to learn. That's for sure. That's like a more sophisticated version of social media. I was going to say X is kind of the same. Like, if you just go to Twitter, like within a week of a model being released, like you're going to know if it's good or not. But you get a lot of people who are influencers. X is the worst because I feel like on the same timeline scroll, I can be like, okay, I finally understood it, and then you scroll the refresh, it's like 17 changes have happened. Yeah, and every half the people love it, half the people hate it. You know, everyone's got—I mean, some of the stuff you wonder about how much is, like, prompted and promoted, and some people are really good at manipulating it. But your point is, no, no, I've substituted it. For people that I like text and get their point of view. And I know that they have a unique reason why they should—like, sorry, they have the unique perspective why they should have it. Like, if they are an AI researcher, they can always take concepts and explain them into ways of why are models better at this and why do these models work this way? What was the breakthrough that nobody could previously solve before? And then I can ask them questions of like, okay, well, can I make it to a customer? Is that better on unit economics? Like, how should I think about that? Or when somebody tells me about the interaction patterns, they're like, you know, this is why Muse and Instinct feel really good, and for these types of, like, prompts that you'd go and put in, like, the browser on Muse actually feels really good even though it's, you know, it's on a mobile device. I'm like, interesting, why? It's like, well, this is the type of things I'm doing. I think that that's, like, the way that you can actually also add value to others is, like, find a thing that you uniquely test. I think Elsie was on your podcast the other day. It was like, you know, she really trusts people who have the same set of prompts that they really use, and they can kind of see the different model performance over time. And if you abstract that concept, what is that, you know, what is your thing that you are putting your stake in the ground and saying, even when everything is changing so rapidly, I have one thing that I can use as my pivot point and then add that unique perspective to others? You know, Dana, you've been—you've mentioned the sort of building for lawyers a few times. And what's funny to me is, like, half a dozen years ago when you were thinking about the Airbnb, which is probably more than half a dozen years ago, you know, you were so compelled by travel and just the concept and how powerful that business is and the culture and everything about it. And I was like, well, you know, jobs are a combination of what you work on and what it's like to work on that problem at that company. And now, you know, you work—actually, to some extent, all three of you have a touchpoint. To lawyers, which, you know, no knock on lawyers. I love lawyers. No knock on lawyers. Lawyers are not, like, the thing that people wake up and say, I got to make my lawyer more efficient, you know? We have different optimization points. And so I'm curious, how did you get comfortable with moving from what was a clear passion, which is travel, to something that I would say is useful, but more of a company decision in some ways? My passion, I think, going to Airbnb, like, started as travel. I think that the passion that grew out of Airbnb was, like, the craft of building products that people love to use. And so when I was evaluating, like, my next role, I was looking at Harvey and another company, and the other company was, like, a lot of, like, backend systems, and Harvey was really owning the core product and building the end user experience. And there was so much low-hanging fruit, so much opportunity, and that was what I ended up being passionate about at the end of Airbnb. It wasn't travel. It was getting to work with amazing designers and build beautiful products that people loved. And so the idea that I would get to go and do that at Harvey, and it was so green and new, and they were, you know, really hungry for consumer PMs who, like, were close to the craft, like, that actually was what was exciting to me. Well, and the same question applies for both of you. I mean, JZ, you've been on our show a few times, and you've mentioned, you know, timekeeping software, you know, wasn't necessarily on the short list of, like, yes, that is absolutely. But when you called me after you had met the CEO, you just couldn't be more excited. I mean, you're an excited person in general, but, like, that conversation, I still remember that day of, like, you're just like, I think there's something special here. And, you know, I think your decision to move to Google, I think, again, a lot of people have a perspective around big tech, and yet, and you had a lot of choices coming out of OpenAI and, you know, some really heady ones. So I'm curious around both of you are decisioning, like, how did you make the decision despite maybe the traditional thinking around, oh, that product area may not be the most interesting to work on? I think these days, the thing that I think is most valuable is data. So I think that technology moats are hard to maintain. Data, when you look at the great companies today, they have unique access to data, and— When I looked at Laurel, I basically saw an opportunity where, because of the workflow that Laurel does, the biggest enterprises with, you know, the almost like the strictest cares around security and governance, they trust Laurel with their data. And so our opportunity to serve these businesses with their own data was huge. And that is the opportunity. That's why I was so excited when I called you, where, you know, back in the day when Laurel was only focused on timekeeping, that's just a workflow. But when I think about the opportunity today, whether it's Laurel or even any company that, you know, I talk to or advise, it's what is the data that you have that's unique and that only you can retain? And that is why I think Laurel is such a massive opportunity, because when you look at Laurel's data, the firms themselves, the companies themselves, can take their own data back and make an open source model 20% more efficient. That is world-changing. That means that they don't have to be reliant on the frontier labs. They have control of their own destiny, and that is super exciting. And so for me, I think a very long time ago, you know, I've worked on everything from deep infrastructure, developer APIs, to consumer, you know, core workflow, lovable product. I feel like the core for me has always been, can I find something that nobody else sees, and can I solve a problem or get a company or a product over a hump that has not been there before, right? So Webflow was an enterprise business when I started, and it became one, and it's actually incredibly important because when you think about building a website, being the marketing orchestration kind of like layer, you got to be in the enterprise because otherwise everything else is getting decimated by the fact that you can just Claude code a website. And this is why I also like investing. Like, is there something there that I just, I have deep belief in that I can nurture? And again, the reason I operate and don't invest. For now is because I, I'm like. For now is because I'm like, as a product person, I can make it happen. I can make the change something. Versus I think when I was consulting in the past, or the thought of advising, or like advising or the thought of investing, you're kind of on the outside. You're not the one driving the ship. And so being able to be like, I see something here that nobody else sees. It is undervalued, and I'm going to build that value and extract it out. That to me is super satisfying. What I love about that is it's, I think that thing I see is highly ambiguous, yet I have an opportunity and perhaps a skill to be able to navigate that choppy water. I think that applies in your case too. It's like, hey, there's this thing that I see, perhaps more than others, but the thing that's really interesting is it's very ambiguous. And I think what I'm hearing from you about Airbnb is Airbnb had just become less ambiguous because you had been there long enough that you knew what was up. Google has a lot of ambiguous problems. I just wanted the little nuclear hat. I think that the playbook in today's day and age and how I was kind of evaluating decisions kind of came to four dimensions. It was very much, what data do you have, right? We haven't used any AI slop terms. Let's talk about what's the durable moat. But I do think that what data does a company actually have in the space that it wanted to play in? Distribution, I think, is like a really big thing. It's really tough to build customer relationships that have high trust when everything changes so fast. You know, the other one that is quintessential is there's great people, right? You always just want to be where there's a center of gravity of really good people that you know are going to be fun to work with, but also drive and ambition. And I think a lot of companies right now, you know, we had a talk one time, I remember I picked up the phone and I was like, everything feels so frenetic at X company. And you were like, That's great. And I was just like, I'm sorry, did you not hear me? Like, it feels frenetic. And you're like, Well, do you want the opposite, where there's no energy? And I think that when I was, like, interviewing at some different places, ironically, I found that when more of the junior employees who kind of haven't been through technology shifts complained about how leadership was constantly changing the strategy and we were throwing away things that we were building, I was like, Oh, this actually is probably pretty good. And so I think that that was kind of the dimensions I look at. And, you know, the place that I was able to end up at, at Google, had all that recipe there. So I was like, I think that there's a lot to go build here. How many people, when you joined the company you're at, did you know ahead of time? Like, at Google, obviously you knew a lot of folks, but were there people that you trusted that were in the inner circle that were like, Okay, I know it's safe over there and I can join? Or was it a fairly clean, like, I don't really know anyone in this area? Probably more of the latter, because I don't know if anybody can say confidently what is safe or not safe and what is the inner circle or not in the inner circle. Things change so rapidly now. And so when there was confidence, that was almost my—it was like almost a turnoff to me. I actually, why are you so confident? It's because you weren't willing to actually change and you felt like you had something to stand on. That actually was a scarier place for me when I was getting that signal from folks. Interesting. What about you? Did you know anyone at Laurel when you joined? I don't think I knew anyone at Laurel personally. I knew a lot of people who knew Ryan, the CEO, but I don't think I knew anyone who had worked there. Harvey, did you know anyone? I didn't know anybody. It was the place that, you know, I talked to investors, VCs, a good friend who's, like, in the people space, trying to figure out what my list of companies was, and it was the one company that came up on everybody's list. But I didn't know anybody. I knew a lot of people who knew Anik. Well, and it's funny because it's like there's a slightly known, but it wasn't like a shoe, and it wasn't like a place. I think when you went to OpenAI, you had a little bit more of an understanding, but OpenAI was also in a very different phase. Yeah, it was much more scaled at that point too, you know. But I mean, I guess whatever raise it's going to be now, it looks so much smaller, but still. All right, so like any other advice that we have for folks that are kind of thinking about making this AI plunge, you know, for people that are, you know, in transition, trying to understand how to navigate these crazy companies, number one concern I hear is burnout, pace. We talked a lot about that. The number two concern is, I don't know whether these things are going to be durable. I don't know if I will enjoy that product space. And so we've hit on these areas as to why it matters. Anything else that we haven't hit on? If you were advising someone who's sitting on the table and they're like, I don't know, I'm just starting my search, but I want to join a hot AI company. What do I do? Shut up and do it. A little bit of that. Like I want to talk about the title and the role. I think a lot of people want the PM title, and I do think that—I'll speak to Laurel—you know, we have few PMs, not a lot. And it's because every single PM can do so much more these days. We are building out, for example, our four deployment teams and, you know, our customer-facing teams at Decagon Sierra, they call them agent PMs. This idea of you're essentially doing the PM's job, but you're very go-to-market-esque. You're embedded with the customer. And my strong recommendation to a lot of people, especially if you're sitting at a non-AI-native company in a product role and you feel like you're not learning, because honestly the answer, besides that we have, you know, some kind of brain trust, is we learn on the job. That's why I love operating. I learn every single day. And so my recommendation to folks is go take that forward deploy role. Go take that AI operations role. Go take a role that isn't necessarily, you know, the one PM role that they might have open, or maybe there's, you know, there are zero PM roles that are open at that company, that AI-native company. And again, a lot of AI-native companies that have strong engineering cultures or strong research cultures, they don't need that many PMs. You know, at Midjourney, there was a very, very small team. And so take the role that might not be the traditional PM role, because what you will learn in two months of being at an AI-native company is life-changing compared to sitting at your SaaS company that is growing at only, you know, double digits. What I love about that advice is it resonates with, you know, someone who is new to career. Actually, it was a friend of a friend who was looking for help on trying to find the first job. And they were interviewing at the OpenAIs and the Harveys of the world, and they were like, Well, these are some of the best-known companies, and I want to be very AI-forward, and I'm struggling because, you know, I'm matched against people with experience. So when I'm in the interview process, they're asking for, like, a take-home, and, you know, the feedback I get is, since I've never worked and I've never been in, you know, in a room with executives, I just didn't have the vocabulary, and so I was trying to figure it out on the side. So my advice is like, look, that's your skip job. Your next job is forward deployed at some startup we've never heard of, because the startup is going to be so desperate when you join to put everything on your plate, and there is not going to be a specific title requirement. And if you can— Put a bunch of wins on the board and learn. You'll learn the vocabulary. You'll learn all these pieces, and maybe you'll find product is interesting or go-to-market's interesting, but you'll be very current. And being current at a fast-growing company, and maybe not think about this as a five-year decision. You know, I don't know how long you folks think about your current roles as, you know, are they 10-year? Are they five-year? Three-year? But these days, I think people are like, look, I'll stay, but everything's changing so rapidly, it's hard to project, right? You just want to make sure you're being relevant and working on ambiguous problems and learning, and then there may be a different day. Right now, when everything's changing, this is the fastest way to stay current. And interestingly enough, the first startup I introduced him to made an offer today. That's amazing. So I think that there's a really good chance that, you know, he ends up, like, really getting... And I think it's a very different way of thinking about career and how to get started, but I think it's very, very relevant to the point you made. So I love that. I always find it interesting that when people also look for jobs, they, like, don't remember the job that they're applying for. And so what I mean by that is you have, like, I have a bunch of people who tell me all these things where they're like, Yeah, I just don't know if this company, this valuation, and everything makes sense. And I'm like, Are you a VC? Like, have you asked, like, how do customers, like, use the product? Do they love it or not? Can you have vision in this space? And can you actually, like, understand the problem deeply enough that, you know, the customer pain point is why they got started? And can you kind of think about the next set of features? Don't talk about the valuation and the moat and, like... So that's the other thing too, is I feel like a lot of these companies and people are trying to pick their jobs. I'm like, What role are you applying for again? Keep that in mind. And what are you going to learn? Yeah. What are you going to learn? What's your day-to-day job? This was great. I hope you stick around for part two, but I appreciate all three of you and your wisdom, and I know you've helped a lot of people today. So thanks again.