Stanford CS153 Frontier Systems | Nikhyl Singhal from Skip on Product Management in the AI Era

1. MAMike Abbott0:08

   So we're gonna have a really different type of session today. Uh, the last, you know, five speakers we've, we've had technical founders. We've talked about AI, we've talked about context, we've talked about the loop. We're gonna switch gears today, and we're gonna talk about product. And it's, this is a topic that's changed a lot over the last twenty years in software, I'd say. And, you know, the history is like twenty years ago, you'd have a, a project manager that would build a PRD, which stands for product requirements document, very structured, and then that person would give it to a set of engineers. And this is kind of the classic playbook for the IBMs, the, the Microsofts of the world. On the consumer side, what was interesting, it's different. There's not as much of a playbook. Typically, consumer companies are founded by a pr- person that is that product person. And I saw this at Twitter, that we had some product managers in the, in the company, but they were never really effective, um, just because it was always the founder who had actually drive the decision. And then you look at companies like Apple, and Apple does not have any product managers. They just ha- basically build all their products between having a designer and having an engineer. But what's interesting is right now the world's changing radically, where a designer can now go vibe code a lot, uh, and really these roles of design, engineering, and product are merging. We're gonna talk about that today, um, with Nikhyl Singhal. So I'll let you introduce yourself. Actually, before you introduce yourself, I wanted to tell you a fun story of how we met. So twenty years ago, we were both starting companies, and we were both fundraising, and he was talking to Sequoia, I was talking to Sequoia, and they wanted us to combine companies. And so we went and had coffee, and we're like, "We're not gonna combine companies." So we ended up going and actually building our own companies on our own, but stayed in good touch and-

2. NSNikhyl Singhal1:51

   Yeah

3. MAMike Abbott1:51

   ... became great friends.

4. NSNikhyl Singhal1:52

   Yeah. If it wasn't for Mike, I think Mike, uh, Mike and I one day was like, "Hey, let's go to the city. We should go to this event." And, uh, that event ended up being the event that, uh, I met my wife at. So every time I think of Mike, I think of, if it wasn't for Mike, I wouldn't be married. So, uh, anyway, that, that's a, that's a different story at a different session. But, uh, thanks everyone for having me here. Uh, I'm Nikhyl Singhal. Uh, let's see. I went to Stanford '90 to '96, which probably seems like ancient history for folks here. Uh, I think you can still co-term. I co-termed in computer science. I was a course advisor, um, which was the person that, you know, took your Polaroid at the time and, and allowed you to declare computer science, although it wasn't much of a gate at the time. And, uh, and since then I've been, like helping people in two ways. One, I, I've, I've built a bunch of companies through founding. Uh, and then I was, uh, I was an executive at, uh, at Google and at Meta and at Credit Karma, kinda helping build product, uh, organizations. Uh, and then for the last fifteen years, I probably had, m- I mean, without probably exaggeration, over 1,000 career conversations with executives, people that are early in career, people that are looking to end their careers, uh, and just trying to helping them navigate kind of the choppy waters around careers. And so, uh, the combination of careers and how to get your most out of the forty, fifty years that you all are gonna work, as well as what product and product management is, is sort of what occupies my time these days. So, um, maybe just a show of hands, you know, how many, uh, folks here in the room, uh, plan on being a product manager in the first five years after they graduate? I'd say maybe 5% of the room, I would say give or take. How many of you would've answered, uh, would've raised your hands two years ago? Yeah. So about twice as many. And so I think that these days we hear this term product management, which we'll spend a little time defining, and what we hear is, uh, that function seems to be dead with AI. It's kind of the roadkill that AI is kind of, uh, uh, kind of providing here. Um, how many of you feel like you don't understand what a product manager does or is? Okay. So a big chunk of you. Okay. So maybe I'll spend just a few minutes kinda going through kinda what is this function called product management. So basically, in every technology company, there's a bunch of people that build stuff, and then there are a bunch of people that sell stuff, and then there's the people that are in between. The people that are in between glues what you're trying to build with how to build it, and that's what a product manager is. And, and for a variety of reasons, um, it's hard to kinda connect the dots between one side and another. Um, now, for the last ten years, product management was largely f- kind of defined in the following way. Um, companies essentially go through what I call the sort of S curve of growth. So at the beginning of every company, your goal, if you... and many of you, I think, are gonna found, your goal is to find product-market fit. Product-market fit is the equivalent of like, you know, rubbing two sticks together hoping that you get some smoke. Product-market fit means that I'm building a product, and I'm hoping that people want it, and then enough people want it that I can kinda keep doing it going forward. So at the beginning, your founders are driving product-market fit. They're doing this rapid experimentation. The goal of the organization is to tr- is to have as many shots on goal as possible. You know, when Mike and I met, that's what we were doing. We were essentially trying to build something and trying to get resonance with customers. Okay. Product management doesn't exist at this phase. So for a lot of people that are like, "Oh, I wanna go do an early stage company, and I wanna help kinda build that company out," you need to found. There is no product manager that makes sense when the company is essentially throwing whatever they're doing away, trying to find some resonance. OkaySome very small percentage of those companies, let's say one, two, three, four percent of those companies will actually find what we call product-market fit. Product-market fit means that you've got a sucking sound. Sucking sound means that you've built something, all of a sudden there's a natural pull. People really want what you want. And for the first time, and, and this is kind of the irony here is, the exact thing that got you to where you are, which is experimentation, rubbing two sticks together, finally smoke is coming, you start to feel heat, you s- need to stop experimenting and throwing things away. What you actually need to do is you need to take a moment and build some resilience, some consistency, because the next customer that you're gonna bring in can't have a completely different product, and the skills of the founder are largely d- designed around, how do I experiment as rapidly as possible? At that moment, they're like, "Oh, you know what I need to do is I need to calm down. I gotta-- now maybe the company's growing up, I gotta go fundraise. I gotta go figure out this go-to-market thing. I gotta figure out the business side. I don't have as much time to change the product." Now product management comes in, and product management at this stage is a much quieter function, more process-oriented. And this is historic, so I'll, I'll talk about how it's changing in a minute. But historically, you're gonna put in predictability, process. You're gonna essentially for the first time have multiple teams, and they don't talk to each other. So you have to have a function which is gluing not only what the customer needs and wants, but also to get multiple teams on the same page. Okay, now of those companies, a very, very, very small percentage of those companies, again, maybe one or two percent, go through what we call hypergrowth. Hypergrowth didn't really exist when I was in school. When I was in school, it took ten, fifteen years for LinkedIn to get its first billion users. It took like eighteen months for Uber, and now it even happens even faster. And the reason why is the App Store, Facebook ads, you know, the distribution of the internet allows companies to grow very rapidly. Well, so there's a point in time when the company is getting so much pull, so much of that sucking sound is happening, when your product is flying off the shelf. Again, this is a one in a hundred, one in a thousand founder situation. At that stage, when you're getting so much pull, you need to not only scale the business rapidly, but you need to expand the business. Hey, we have a hit product, but we have so much opportunity. How do we get into an adjacent product line? Well, again, as a founder that's going through all of this rapid growth, it's extraordinarily hard to do both scaling and expanding. And at the same time, for the last ten years, to do both of those, you need to hire loads of people. So that's where product management and the chief product officer really came in. And so what you would see is companies would bring in s- large product management teams to essentially manage the scale of the organization to achieve two things: scale our existing product so it can get bigger and expand the product. And so that's kind of the job that I had when I was at Google, where we were trying new products because search and ads were doing so well. And when I was at Facebook, I was trying to scale feed but then expand into things like short-form video, which is where Reels came from. And when I was at Credit Karma, we were a credit score company, and we wanted to be the money button on the phone. So you had to build a team to se- essentially achieve all those. And then finally, at the very, very end, you get to big tech, the late stage, and you have to sort of start over again because now you have so much success, whether it was Twitter or whether it's Facebook or whatever these companies are, your job is to sort of combat innovator's dilemma. You need to go create something new. You need to go do zero to one when there's so many reasons not to do it when you have small, you know, small businesses don't compare to the large, huge businesses that exist. Okay, so these four phases require totally different types of product managers, but they're all product management in general. So I'm gonna pause for a second to see if this resonates. This sort of helps people understand why this function exists, and then we'll talk a little bit about what we're seeing with AI and how it might be impacting all of these functions. But pause for a second. Questions on that around the sort of phases of product management. So the question-- Thank you for the question. The question was that I worked on Hangouts, and Hangouts essentially, uh, didn't get off the ground. It was a huge priority and a huge initiative from the company. Uh, just for those of you that may not remember, so this is before WhatsApp was acquired by Meta. Um, Hangouts was essentially a combined effort of taking what existed in Gmail and Android and all the communication products that exist on the device and trying to build one single app, right? So single app would be it would both text, it would do voice, and it would have video calling all in one, and it was built on a brand-new stack. So the question that was asked was, what did it teach you about founding in the sense that that company was incapable of delivering something long term? Um, I think that it's, it's sort of an opposite problem of founding in some ways. I think it's hard for large or-- Well, maybe there's two things. One is that I think that Google Hangouts was a problem that, that users did not have, but the inside the building company had. Meaning, like, when you looked at Hangouts, not a lot of you in this room are sitting there and saying, "Man, do I wish I had one app on my phone that I could use for every person I wanna communicate with, whether it's a business, an individual, whether it's voice, whether it's text, whether it's video, whether it's group." I bet you all of you in a given day spend time in ZoomSpend time in iMessage, spend time perhaps in WhatsApp, spend time on your phone making a phone call. I know some people haven't done that in a long time, but that still happens. You get spam calls. Like, those are all different apps, and it turns out that our belief was we have like seven different code bases in seven different products with seven different registrations and identities. It would make sense if this was all consolidated. Well, it turns out customers didn't just-- didn't care. So maybe my first point is, sometimes inside the building drama and trauma and challenge doesn't translate to outside the building usage. And so what WhatsApp did is WhatsApp essentially had an extraordinarily well-executed text-only play, and their focus was just to make that work in India and make that work across all rural regions and make it so that it was the most reliable text message product out in the world. And it turned out that once they got the network, it was possible to layer on things like voice, things like video. But it turns out that in and of itself, if you're using video through FaceTime and WhatsApp through messaging, that's still a fine usage. And so in many ways, solving the actual customer problem versus conflating what's outside versus inside the building, I think is probably the first thing. And the second thing is you need to kind of stick with it, and I think large companies have trouble sticking with things that don't look like they're winning from day one. And maybe the third thing I would say, and this is something that Google did teach me, is that the best products at Google always start out looking very poor. They don't look good. They don't work well. If you ever use the Android One phone, it looked like a doorstop. It was fricking horrible. Turns out what they realized is it doesn't matter how you start, it's how fast you improve it, and this will definitely come up with AI. Um, it's the speed of iteration. Their innovation with Chrome, when they delivered Chrome and they beat Firefox, which was the incumbent and eventually Internet Explorer, it was because their greatest innovation was they shipped every six weeks, while Firefox shipped every quarter and Internet Explorer shipped every year. Android shipped every quarter while iOS shipped every year. It turned out that they built an organization based on iteration speed, and the faster you can improve and hit the objective, the better your product is. That's something that I think startups have a huge advantage of. So my note was solve a real customer problem, not one you imagine. Iterate very, very quickly, and recognize that large incumbents need to see success much quicker than a small startup. Those are the three lessons I take. So the question is around forward deploys engineers and perhaps how it juxtaposes against product management and where it might end up changing with inter-- uh, with AI. So I think the-- First of all, I'll de- I'll define what I understand about forward deployed engineers, and Mike, you probably know this better than me. But, um, a forward deployed engineer is essentially this new kind of s- uh, style of engineering who is technical, and yet they, uh, go in to solve an actual customer's problem with the product, and then from that, they pull, uh, expertise, standards, learnings into the core product. And I think that there's an insightful question here, which is like, well, that sounds on the surface a lot like what product managers do. Product managers spend a lot of time trying to understand, like, hey, what the heck is going on with customer use case, and then how do I turn that into a core product that many, many customers can use? So I think that the concept of actually going into the customer and delivering things is absolutely, has been very beneficial to certain, you know, somewhat complex products, uh, like the Palantirs of the world, where there's perhaps fewer enterprises, but very, very deep, complex enterprises, and Mike can speak to that at General Motors. But I think that in the world of AI, what we're surprised to see is how effective AI is to be able to understand nuance and detail on what's happening at the customer level. You know, so if you're a product person today, the way product management has now shifted, and we'll get into a little bit of this here in a second, but, you know, you oftentimes have agents that in the morning give you a summary report of every single customer use case that has been presented in, uh, a chat session with your customer service reps. You have a summary of every sales call that took place in the past week, okay? In the past, like if I was to come a year ago and have this conversation, that would sound like science fiction. What you're telling me that every day I can wake up and whether I have two salespeople or a hundred salespeople, whether I have fifty people that have filled out the survey on the website, whether I have a thousand people that have complained about this feature or this product, I can get a clear summary. Not only can I get a clear summary, I can see that in prioritized order. I can see a prioritization order based on how much money that might generate for my business, how complicated this might be to implement, how inconsistent this is from the system that we are trying to build or the brand that we're trying to create, and I can, based on that, have judgment. So what I'm saying is that a forward deploy engineer is extremely helpful to pull out that insight. But the reality is now we're able to even move a step beyond that because there's so many signals that contribute to what should we build and why should we build it. Does that make sense?

5. MAMike Abbott17:42

   That's a good question.

6. NSNikhyl Singhal17:43

   You have thoughts on that?

7. MAMike Abbott17:44

   Um, well, no, I was just gonna say, I was thinking back in the day, a forward deployed engineer was really a services group.

8. NSNikhyl Singhal17:50

   Mm-hmm.

9. MAMike Abbott17:50

   Like professional services. So I think Palantir's did a brilliant job of rebranding that role. But I always saw in that role really close to product management, and a lot of times at the beginning of a company, that was our product manager.

10. NSNikhyl Singhal18:01

    That's right.

11. MAMike Abbott18:01

    Because they're so close to the customer.

12. NSNikhyl Singhal18:03

    Yeah.

13. MAMike Abbott18:04

    Right. So let's get into the AI, the, like the transition. So you, you kinda hinted at that. Um, why don't you mention about Skip first, so you know, they, people know where you're getting this data from?

14. NSNikhyl Singhal18:13

    Yeah. Yeah. So, um, so as I mentioned, you know, I-- so I, I appreciate the questions on some of my operating experiences. So after, uh, I left Meta, you know, I, I liked thinking through... I mean, b-basically what's gonna happen is, uh, the rough, the rough tenure of people that tend to work tend to be about 40 to 50 years, right? And, and that, that number theoretically might grow. So when you are 25 years old, you're probably not gonna retire age 65, 'cause unless I'm mistaken, no one in this room is gonna physically labor in their jobs. They're gonna sit at a desk, and though that does have, you know, issues, it's not going to break your body down. So you could choose to work for 50 years. The average number of years that people typically work in a tech company tends to be about two to three years, which is a shocking small number. Most people think that they'll work, uh, maybe five, eight, 10, 12 years. Historically, and, you know, looking at my career and Mike's career, we basically spend, you know, between two to four years in most of our endeavors, including the companies that we founded. So if you do the math, if you're gonna work for 50 years, and you're gonna, you know, on average, your jobs are gonna be between two and three years, uh, math tells me that you're gonna have between 15 and 18 jobs. Okay, 15 and 18 jobs means that your career is not like periods in a hockey game. It's like chapters in a book. So what I've realized is that the most important thing that you need to think about is how do you get the maximum value out of those 15 jobs, and how do each chapter of the book build on itself? The reason why I called my property The Skip is that I think that the best career advice is to think about the chapter after the one you're thinking about now. So it's less about if all of you are thinking about, "Well, where's my first job gonna be out of college?" My ask would be, what's your second job gonna be out of college, and how do you make sure your first job sets you up to get the maximum amount of opportunity in your second job? So I think so highly about this career calculus because, you know, the vast majority of people I talk to, um, do a fairly poor job of managing their careers, and that is very, very expensive. Um, I mean, if you go, go and have conversations with people that are, that, that are adults, very, very few of them will say that I was very intentional about the career decisions that worked out, and many of them will have five, six, seven of these jobs to be very suboptimal. So it is my opinion that one of the most important things we need to work on is to get you to get the right jobs and the right sequence of jobs to maximize your career, maximize your impact, maximize your earning, ma-maximize your, your happiness. So r- so what Skip was designed to do was almost be like a talent agency for product people, and not necessarily product managers. You know, we have 12 in my group that are actually founding. I think product builders is kind of the direction that we're going in. But my belief is that the top people that exist in our industry should be represented, much like Hollywood is represented, much like athletes are represented, and in some ways, that will get people to make better decisions in this very important endeavor. So, um, to build a talent agency will probably require about 30 years' worth of work, 'cause that's not the way our industry is constituted today. And so when I was about 52, I said, "I wanna work until about 82," so I spent a 30-year career. Uh, my last, uh, turn was to work on this property, which includes some content. So you can go to skip.show, and you get some of my Substack and some of the content that I publish around careers, and it's mostly around product leadership because that's the area that I spend a lot of time with. We have about 125 heads of product for most of the biggest companies out there, from Anthropic and OpenAI, all the way to Meta and others. So, um, our goal is to sort of take a community of the top folks, not only get them to connect with each other and advance their career, but then build a set of properties, in this case, the thing that we called Skip Coach, which helps everyone trying to advance in the sort of tech ind-industry to get the latest learnings and the findings. And, you know, AI has had a huge impact on everything that we do in terms of building companies and building products are changing. How do you stay current? Well, that's a lot of what our mission is, is to help people understand what's actually happening versus what's viral on Twitter, which, you know, is accurate in some cases, but oftentimes is extreme. So anyway, that's the, the nature of what Skip is, and feel free to check it out, and drop me a note if you have questions.

15. MAMike Abbott22:38

    So, so with that group, what are the trends that you're seeing today with the impact of AI?

16. NSNikhyl Singhal22:44

    Yeah. So I think that, um, uh, maybe a few things. I think that... Well, I mean, maybe I'll ask this question by show of hands, and then I'll answer mine. It's like, how many of you are anxious, uh, about, um, getting a job or, you know, as you leave college and as you graduate? How much w- I'm just trying to get an anxiety meter right now. Yeah, so basically the entire audience. Um, that is very much in contrast, I would say, to when I graduated. I don't think anyone was anxi- Maybe we weren't, we weren't as smart as you folks are, but we definitely were not anxious, per se. Um, and I think that, um, how-- uh, and maybe I'll ask the second question is, how many of you are having a lot of fun building things using AI? Okay, same group. Okay. So this is essentially the exact show of hands that I have with all of the executives I work with. Um, if I asked these two questions two years ago, the anxiety in the room was very low, 20, 30%, maybe two years ago is a little bit, three years ago. Before we started to see the mass layoffs when z- interest rates were 0%, it was, like, fricking great to be in an industry. There, you know, every person, even if they were, you know, average skilled, would have six job offers and, uh, and, and, and there was low anxiety. HoweverEveryone disliked their jobs. I could not find anyone three years ago that literally loved their roles. And the reason why is it was all filled with responsibility without authority. Product in general was essentially a movement of information. No matter how big or small the company was, your job was to package information for some other decider. And that is a horrid job if you are a-- I mean, being a bureaucrat sucks if you're a builder. And, um, the vast majority of product people get into product 'cause they like building stuff, and then very quickly they realize, like, hey, once you know how to build stuff, you ought to organize others that build stuff, and that's when things kind of go dark. So what I'm noticing, the number one thing I'm noticing is there is more joy for the leaders and people that are in tech than ever before because they don't have to depend on an engineer, on a designer, on a founder, on their boss in order to get something done. They can all build. And there's nothing more empowering than being able to whip out Claude Code and build something, or more specifically, to obsolete yourself from some status report that you hate having to file. If any of you have done internships and all that, the concept of a status report and writing, I mean, it just feels mind-numbing. But if your job is mostly moving information, and now for the first time you're like, "Let me get this straight. All the parts of my job that I dislike and hate, I can essentially obsolete myself, engineer myself out of, and then the parts that I love, judgment, decisioning, uh, being courageous, testing things in the wild, talking to customers, working with engineers on a really hard problem, partnering with another company on expanding the pie, those are the parts of my job that exist," that is amazing. So the first thing I would just say is the industry that you enter into if you are joining tech is more interesting and more fun than in any other time that I've ever experienced. And so it is going to be an incredibly interesting job. The second thing is everybody at leadership is anxious. If you're a big tech, you're looking at somewhere between thirty and seventy percent layoffs this calendar year. It is rough. But you're also looking at skyrocketing salaries and massive hiring, which I'm like, how are these two things possible to exist in the same sentence? But it does, because the idea of someone coming in and being a movement of information, that is a dinosaur. I can hire an AI to do that. But I need a lot more people that have a lot more judgment, and they have to be paid a lot more money because we can build so much more. We just need to know, is it the right thing to build? Is it working? And does it fit into our system design? And so we have the most unusual time period that we've ever seen. One side is we have an am-amazing opportunity to build more. We have a lot more of an opportunity to be hands-on. Companies are a lot denser, meaning that there's not twelve layers between yourself and, say, the CEO, and so there's a lot more connection, there's a lot more camaraderie, but the pace is extraordinary. Everyone in my group is working twice as hard as they ever have worked in the past, and they're building, and they're having a ton of fun. So it is the most unusual time, and frankly, I think the, the part that's most difficult is not represented in this room today. It's the middle manager who was an okay builder, but during zero interest rates were promoted into like, "You can talk. Why don't you manage other people? You don't really have much of an opinion on what we should build, but you have a lot of opinion on how to move information." These are the people that, you know, Mike referred to at Twitter and others. There are a lot of people that were moving information. They then see the Claude Codes of the work. They have kids. They have, you know, aging parents. They are in their mid-thirties. They're sitting there like, "I don't have time to spend reinventing myself." Those are the people that are gonna get laid off. Those are the people that are very difficult to find new roles. While folks like yourself, who, like, start their journey living in these AI tools with opinion, perhaps experiencing founding and product development from the beginning, those people are extremely well-positioned. And a big change that's happening right now is when I see high-quality employers like the Anthropics, the OpenAIs, and others, when they interview, they don't actually care what brand of company you've been at. They wanna know how modern you are in how you think and how you use your tools. They can tell when people are learning the tools at the same time that they're interviewing. They can tell when people are like, "Look, I, I'm here 'cause I wanna manage." It's like, well, there's no more management. It's all about building. It's all about being hands-on, and the folks that are here are actually in the better position. So I actually think that, you know, if I was to predict, the room that is the most anxious doesn't exist here. The room that's most anxious are the middle managers that I think are eight to fifteen years into career. The executives are the one that are driving a lot of the change, and then the folks that are starting out are getting essentially dislodging what we see in, in, in historics.

17. MAMike Abbott29:45

    I think that's such an interesting change. I mean, one of the things that, you know, I think is so important for everyone in their careers is to stay curious. Because if you're curious, you will stay modern. And I think it's these folks, to your point, that are not, that will go, and they won't find a-- they won't be able to-

18. NSNikhyl Singhal29:58

    That's right

19. MAMike Abbott29:58

    ... find another job.

20. NSNikhyl Singhal29:59

    That's right, and that's why I think that people in this room have an edge up. I mean, to get into a school of this caliber and then to stay current is kind of a requirement.

21. MAMike Abbott30:07

    Yeah.

22. NSNikhyl Singhal30:07

    I think that being in a class like this and trying to understand what's the latest, but I think you have to translate that into being hands-on and being gritty. And I mean, that's, that's, that's literally the skill that matters. It's being current as well as, you know, just having that opinion on what needs to be built and why you wanna build it, and then being able to quickly validate it. I mean, that's kind of the founder skill that's essentially carrying across all these organizations.

23. MAMike Abbott30:32

    Let's go to a question

24. NSNikhyl Singhal30:33

    Probably the most common question we hear nowadays, and so it's very valid. Yeah, I'll repeat it. So the question is, you know, in light of layoffs like we hear from Salesforce and from Block, I mean, obviously Snap did about 10% yesterday, you know, what's the future of s- a, a function like product management when many of them are being laid off? It, you know, it seems like that era perhaps is over, uh, and, you know, is, is that, you know, I think the implication is, is that a function to avoid, especially when engineers are sort of empowered to do everything? Um, my, my sense is that, uh, I think that this is where the data doesn't actually match the, the historic. So, um, right now, there are more open product management roles than, uh, right now in industry than in ever, uh, before in history, um, which is completely unobvious to folks. Um, but getting hired is actually the number one-- Like it's, it's, there's, there's, there's more positions that are available. Um, salaries for the top 1% of product people have more than doubled in the last 18 months. So it is, um... There are four, four contracts that I have been part of helping negotiate that crossed eight figures annual, um, in terms of compensation for product leaders. So the idea that companies will no longer need product, I have not found, uh, many successful companies that are actually suggesting that. I think that what is actually happening is most companies exploded the number of PMs in the last five years, particularly during COVID. Um, what ended up happening is there was free money, so it was very easy to expand the business, and there was a connection between if I want to expand the business, I need people that can help organize that team, and that is the title of product managers. So those people came in to essentially not build, but to organize. And there was this conflation. I mean, there are sort of two ti-- two words in the term, product and manager, and the manager got boldfaced. And so if you look to most companies that are doing layoffs, what they're essentially doing across the board, they're not just targeting product, but product does get hit. They're essentially pulling themselves back to where they were five years back. And I think a lot of it is because in the last five years, they've expanded their staff, and they haven't gotten a lot for it. And so I think that that's somewhat different than AI is taking their jobs. But I would say that they aren't going to hire that type of PM anymore because they get a much better result from AI presenting the information. But they need a different class of skill, a product builder skill, which is a little bit broader than just an information mover. And then we're seeing designers who stand up in an organization, and they say, "Well, if the idea is for me to have an opinion on what we should build, well, that's kind of me." I mean, designers also subclass into people that build pixels, and those people are struggling, and then people that actually decide on what the product should do. Engineers, you can argue, everyone says, "Hey, engineers are essentially, you know, in the best position." Well, I mean, maybe. If the models continue to improve the quality of code coming out, it's hard to stay above that. Yes, you can leverage that, and those people will be paid well, but often it's the engineers that have an opinion on what products should do that are actually... So what we're seeing is a merger between a different class of role that's not so siloed as this is a PM, and they only do this. This is a designer, this is an engineer, this is a data scientist. We're seeing a lot more of a merger, but we're seeing a huge spike because companies can build way more, way faster, and they desperately need people to decide on what to build. At Meta, the metaverse appears on the outside to be a failure. How did the company let this happen? And in particular, uh, is the culture so yes man that it has no ability to prevent such a, uh, challenge from taking place? Is that a fair characterization of the question? Horizon Worlds was shut down as evidence that the metaverse is essentially pulling out. Yeah. Yeah, and I think that, I think that, you know, I personally can give you my opinion on how that happened. I, you know, and many others were not part of that decision, and companies like Meta are quite large. And, um, I would say that, uh, I have a lot of, um, admiration for Mark and his ability to drive belief and his balance sheet towards a degree of belief. Um, I don't think that Mark always gets it right, and I think that he would argue that he doesn't get it right either. But I think that Mark's belief, and I think that there's merit to this, is that he wants to essentially be the innovator of the next platform of computing. And I think that, uh, Meta as a company was not an innovator on mobile. It was a, uh, it leveraged mobile. It leveraged the web. And for Meta to get to the next level as a company, I think it needed to be the innovator of that platform as Apple and Google had historically been the innovator. It was also not the innovator of the cloud, which was another big innovation. I believe that Mark felt that in order for the company to get to the next level, he wanted to be the innovator of that platform and believed very strongly that that was the metaverse, but believed he couldn't iterate his way thereYou know, unlike what I described happened at Google, where they tested the market with Hangouts and then they pulled back, Mark was essentially like, "I'll, I'll spend 10 years trying to get there because I believe that I can drive that next platform." Um, I think that now that we're, what, five years in and we're not seeing that, I think that there's an open question as to how much does he continue to invest in his mind, and how much does he look at the next platform. The next platform, we could argue, and the one that is kinda coming before the metaverse is AI. And so what you see is Mark's like, "Well, I wanna be a leader in that next platform. I'm not in a position to perhaps drive that for the first, but I am in a position to have a strong play." So he's again putting a ton of his time, energy, and investment into that. And again, I have tremendous admiration for that. Now, why does a company not have, uh, you know, a ability to push back on that? Um, I think that Mark is very founder-led, and his, uh, the company is run because of the area and direction that he thinks is the correct way. That was not the culture I saw at Google. Google, as example, a lot of the discussion internally was you have to first solve the inside the building problem, make sure everyone's aligned, and then get things out. But I think at Apple, I think Mark-

25. MAMike Abbott38:08

    Very different. Yeah

26. NSNikhyl Singhal38:08

    ... Mike can make this point, and I think at Apple and, and to some extent, uh, Meta run differently. And I think that, um, to some extent, um, there's not a right or a wrong version of this, but I think to do big innovation, you can't do it by consensus. And so what Mark essentially did is he said, "I need to do a big unobvious innovation that is essentially gonna take a dramatic amount of capital. I'm going to deliver on that, and if it doesn't work, I'll go find the next one," which is kinda how it's played out. And I think Apple probably runs closer to that, but I think Google doesn't do that. They're much more of a reactive organization.

27. MAMike Abbott38:49

    There's also this term sunk cost fallacy.

28. NSNikhyl Singhal38:51

    Mm-hmm.

29. MAMike Abbott38:51

    Where, and especially in big companies, but this can happen in a startup where you've invested the five years, and you kinda have to ri- you rationalize in your mind, "Well, I have to keep going 'cause I'd put five years into it," where the better answer would've been just to kill it. Um, Apple tends to do a really good job at killing things, but not always. The car, when I was at Apple, like it was obvious, like that project was never gonna happen, and we kept throwing billions of dollars at it, and everyone outside of the group was like, "Okay, we can't hire two people, but they can hire a thousand." Um, and it, it-- that took too long, I think as an example of saying, "Hey, we can't do this."

30. NSNikhyl Singhal39:24

    I, I think that it's also what, what you don't realize until you work at these larger companies is when your business is, like, putting out, you know, tens to hundreds of billions of dollars per quarter, um, if you come up with a new business that makes a billion dollars, like if you, if any of you invented a business that made a billion dollars, that would be a, I mean, that'd be a life-changing moment and a career-making moment and an, an amazing, impossible act, right? To make an extra billion dollars at a company like Meta, it was like a four-line change into a ranking algorithm. You know, that, the amount of scale that these companies have and are so large that for them to be able to say, "Well, how do I continue to grow 20, 30, 40% in a new and adjacent field?" You need to build a car.

Episode duration: 1:03:14