The New PM, Claire Vo | Compile 26

0:00 I'm so excited to be here and I was really thinking about I've done a couple talks and I was like what is the worst slot? Is it worse to be between you and lunch or is it worse to be in like the post- lunchunch coma? So I'm hoping that being the 11 a.m. 11:30 slot you all are still I think you know the subtitles to my talk could be like stay humble, stay hungry. So like we're going to stay hungry together. We're gonna figure out what it means to be a product manager in the age where anybody can build anything. And I'm using this phrase product management loosely like product person, product agent, product autonomous loop, whatever we're calling it. You know, the decision of what to build, why to build it, and how to build it, I think has changed so much in my over 20 years in in this technology world. And I I want to talk to you about some of the things that I'm seeing, some of the things that are changing and some of the things that I think fundamentally have stayed the same. So there's this you know maxim in product management in the world which is like build something people want. And again just as I am using product manager loosely, I am also using people loosely. So we can make this build something agents want but it's it's fundamentally the job of product of a founder who's focused on product any of you um is to build something where there is desire for that product there's a real problem to be solved and people want it and if you've ever built something people want it's such a like visceral feeling that cannot be replaced but it's it's not what we spend most of our time in fact We have been coming up with creative ways to avoid building things forever. I'm getting a laugh over here. See, here's the problem with the lighting is now I can just point to individual people.

1:52 Yeah. So, we've been spending like decades and decades and decades inventing increasingly creative ways to not build stuff. And some of this was actually pretty reasonable. Um, in in the past, you know, like two years ago, uh, building was very expensive and slow and risky. And so it actually was logical to not build stuff. And in fact many product managers I gave a talk you know two years ago and I said there are like almost 999 million results for how a product manager can learn to say no.

2:26 It was almost the entire job was to not build things. And what were some of the creative ways that we structured our not building of things and some of this will if if you haven't been victimized by any of these diagrams? You are young and youthful and I'm very happy for you. Um I unfortunately have been victimized by all of them. But yeah, we had like agile these loops and you're gonna have, you know, this is like a loop, sometimes it's a waterfall, sometimes it's something different, but like this step by-step path and like every color you had to like pass the gate and if you didn't pass the gate, you went back to zero to the the backlog or below the cut line. So, we we had agile um we had like these things called backlogs. I don't know if you all remember those. I don't have one any I have a backlog of PRs to review which apparently cursor is going to help me with but like these backlogs of ideas that you weren't sure if you had the capacity to build and then they had to you know go into these meetings where you debated numbers in a spreadsheet and then they went to these things called sprints was all very exciting. Um yes PRDs like these complicated documents where you had to describe how buttons worked. Um this is how precious engineering was 10 years ago. They could not decide how the button worked themselves. Was not possible. You had to write down how these button worked. This thing who knows this thing somebody.

3:44 Okay, we got we got a couple people. If you don't know safe, you really haven't lived. Um, but this is apparently an effective way to write software. I don't know. Um, and then we have multi-agent loops and autonomous uh coding. A Hold on. Wait a sec. Wait a second. We're all doing weird stuff now. And it's maybe in service of building, but my challenge to you all is a lot of what we have going on right now is not in service of building something people want. And so it's in service of building for sure. I do a lot of this. I pay, you know, I pay my my token tax. But I think these are elaborate systems that are really just descriptors of how work gets done. Very similar to the agile loop, very similar to the backlog, you know, flowchart, very similar to whatever safe is. This is I think an expression of the same impulse which is we are still very far from building what people want and so we invent other productive feeling things to do especially as product managers sometimes as engineers and we do that before we know whether or not the output would be valuable and I think this is the fundamental challenge of this moment in AI which is you know 10 years ago we had scrum and story points and backlogs and now we have agent swarms and token maxing and we're all working 24 hours a day and I think what I'm seeing and okay cursor people cannot answer this question you keep your hands down this is not for you who is writing more code than ever right now okay they' revenue is going up proportionate to the amount of PRs they're shipping okay again cursor arms down um and so I think the trends do change We love talking about the work, especially product people. Oh god, we love talking about the work. What a glory. Not to do the work, but to talk about the Lord.

5:49 But a lot of this, I think, is is is an expression of the exact same thing we've seen before. And so I want to talk about what we should actually do right now. What are the inputs to why this is happening? What should we do differently? And what does it look like to build amazing product when kind of anything is is possible. And I think what we saw on stage this morning is just showing the acceleration of what is possible is happening more and more faster and faster and we should presume that capacity will no longer be our issue. Okay. So going back building was hard and expensive. So these old systems existed because engineers were a super scarce resource. Product managers became like these very skeptical gatekeepers.

6:37 And you know, not not only did we say no, we cut scope. Um, so many so much scope being cut. Um, and and really the job was to protect engineering capacity. And so product managers and in in fact like whole businesses became very good at this funnel. It became very good at becoming skeptical of ideas um so that we could protect our very precious engineers. And what's different now is that funnel no longer needs to exist for capacity reasons. And we're seeing this constraint breaking. So this is a chart Anthropic put out. Um engineers are now shipping eight times more code per quarter than they were in the past. You can just you can debate whether or not this is a perfect measurement, but I think we all know code is being shipped more and more by most teams. And so the amount that one person can ship is going up significantly and the amount a team can go up is going up significantly and I think it's about to outpace what I showed before which is this backlog of great ideas and so the the constraint has shifted and therefore I think product teams product people wherever wherever you are whether you're in an engineering role a product role a founder role has to change and I I just think this is is straight facts you know We asked in in the room who's shipping more code but the revenue isn't going up. We're seeing more things. We we're in an age of more more and more more and like I love more. There's somebody in this room that has worked for me. They know I love I love to ship everything. I am a yes and product person. Backlog, no thank you. Scope cut, no thank you. I want to ship it all. And this was before AI. So imagine now that I have this sort of like infinite engineering intelligence on demand. What a total nightmare I am. Um but luckily I'm in good company which is everybody is building more. We are just doing more.

8:35 But I think you all have seen this chart which is while app store you know submissions and apps in the app store have gone up installs and ratings and use have not. And so there's this big gap between what's being shipped into the world with all this excess capacity and what's being consumed by what we put on this first slide which is people. There is just not consumption of our code at scale in the market. And I think that is the problem for ambitious builders right now. There's so many more things that I can execute on right now. There are so many more things that I can imagine. There's so many more things that I can put into production.

9:16 That is not my problem anymore. Production is exploding, but demand is not proportionate to production. And I think that's a real that's not just a problem for all of us builders out here. It's something that we need to think about with platforms we're using, how we think about our job to be done, and what we ultimately do with all these tools that are being put at our fingertips. So I think we have more like we're living in an age of abundance. We have token abundance. We have intelligence abundance. We have code abundance. There is more capacity. There are more builders including non-technical folks.

9:54 There is certainly more code. There's so much more code that we need to GitHubs. Um that's what I heard today. And there are just like more things. There is more more and more. I just like this is so obviously true. But there are not more things that people want. And I I think even more than this, there are not more things that people will pay for, which is what I like. And so I really do think that the this is the foundational product problem. This is the foundational business problem of the age is we don't have anywhere to land our capacity. And we have an entire generation of product thinkers and builders trained on a completely different way to think about their job and how you approach this problem because the constraint has shifted so quickly. So there's a skills problem here. There's a framework problem here. And I think there's like a market gap problem here because there's just simply nowhere for all this capacity to go. Good thing product managers, product people love they they love a problem. This is a problem statement. Um they love a problem. And so there are things I think you can focus on that will help you get leverage out of compute, that will help you get leverage out of code. And I hope that by walking through these things, you all will take away the kind of foundational steps and product practices that are going to make you all, you know, $60 billion wealthier. So that's that's my goal out of this. You're going to go, I went to Clarvo's talk and six months later she really to the moon. Um, so that's my goal for for this talk. So again, code is abundant. The market is not and that is the problem that product people need to solve. So, what do I think product management should be? I'm going to give you give you all the answer. Um, so it's definitely no longer thinking about engineering capacity.

11:59 That's a very boring job. I I don't actually think it's about prototyping. Um, I think that's like a lazy interim step. That's interesting to upskill yourself, but I don't actually think that's necessarily the core of product work. I think the core of product work is finding direction, deciding what's worth building and proving that it can matter. I describe this a little bit differently than most people do because I like ARR. Uh yeah, that's right. Woo. For the recurring revenue SAS people, give me some love. Okay. So, I think building what people want is is a job, but that's just sort of a maximum that you say.

12:38 What does it mean to have something people want? It is not like, "Oh, I have a problem. It would be nice if somebody built this." That's personal software. We're all going to have that. I built like a 64 song emo bracket for March Sadness. There is no market for that other than me. It's something people want, but it's not product market fit. And so, I'm talking about um real commercial concerns, real businesses. So, I think there are two things that you need to think about as a product person in terms of defining what people want. The first one is commercial value.

13:10 Will people pay to solve this problem? Do they perceive this to have any value in their life so that they will exchange that value for money? And I think so few product people put this problem for. It's like a little distasteful um to think about commercial value, but I actually think this is the problem of of the era. Yes, we have all this intelligence, but my my hypothesis is that the rate limiter is actually going to be on commercializable ideas. Not can I ship this feature and will someone use it, but can I ship this feature and will there be an exchange of value that results in money? And I think more the more that product managers and product thinkers and founders can think about this problem um the money will not always just come um I think is very very important. And then the second part of building something people want is can you drive behavior change to adopt that product. And this is where I see a lot of people failing. They can identify a problem.

14:18 Claire needs a way to determine the best emo song. They can build a solution. We're going to vibe code this bracket app. Um but can I get a hundred people to use that? Can I get a thousand people to use that? Can I get them to log in and vote and like like that piece of behavior change I think is so underappreciated in this moment where features can fly out of your IDE on demand. And I think we see so much we think that if I ship the button, if I ship the command line tool, if I ship the platform and it has reasonable value, people will of course adopt it.

14:57 And my entire experience in almost two decades of building products is that is rarely true. It is actually quite magical if you can ship a product and meet a market where behavior change actually happens. And I think turning the screws on experience, turning the screws on positioning, turning the screws on distribution to figure out behavior change is one of the hardest jobs of product and something I don't see people talking about a lot. I see a lot of people talking about oh wouldn't it be cool if this existed and I'm going to manifest it not who is actually going to change how they work or how they live to use this. So this if you do one thing well two things if you do two things as a product manager you do this side please. But I think there's something else and something that AI cannot help us with or at least in my opinion AI cannot help us with and that is manifesting novel ideas from first principles. I have tried very hard to brainstorm with AI, very hard. And it can make my ide my existing ideas good.

15:59 And it can um explore adjacencies fairly effectively, but it is actually not that great at manifesting a future state of the world that I think exists, that I think I can manifest through behavior change, that has commercial value, that has no basis in reality. And I think copying products are is no longer an advantage. Shipping faster is no longer an advantage. I think having a unique vision of the world and being able to articulate what the shape of that product looks like and build it to that vision absent any evidence it and and with a lot of conviction. That is the job of product managers. And I feel like product people have almost let this muscle atrophy so long because building used to be expensive. And so instead of saying I'm going to be bold and without evidence I'm going to ship something, we said I'm going to be incremental and only with evidence will I like ek out a little engineering capacity for this issue. And so I think this is a muscle that people need to develop. And no, like no model is going to confirm for you that your brand new novel idea is going to work in the market. Like there's just nothing in I mean I guess you could like pick pick a nice one and it will tell you you're a genius and it'll definitely work. But there is just no knowing in so much of product management. I think I've had so much exper I I've did a experimentation startup. My friend up here in the front row knows I did a startup all about AB testing. I have worked at basically every every AB testing company in the world. And guess what? There are some things you just cannot know through data and having conviction and having vision I think is the job of the new product person especially when capacity is unconstrained. And then the last thing that I think product needs to think about is distribution. Product just does not end at problem and solution. We can all build problems. We can all discover problems and we can all discover solutions. I think what we really need to think about is bringing people along in like why are we all here? Because we're part of this cursor community. We are bringing people along a journey of saying this is where I think the world is going. I would like you to hitch your wagon to mine and come along. And that is not just customers or community. That is teammates and employees. And that is whole markets. whole markets need to follow you into your vision in order for you to capitalized uh capitalize word intended on the opportunity at hand. And I still think this is a job for the human. Um, I have been known to say RZ is the only moat, which is like truly I get a g giggle up here. Like truly if you have like anybody can code, if anybody can ship, if anybody can build an agent, you know, pretty soon if like anybody can put data centers on the moon, like what is truly what is the mode? And it is who convinces the the best talent in the market, who convinces customers and who convinces an entire market to follow them along their journey and invest in that vision. And this again is a skill that has to be developed. It's a skill that needs to be exercised internally and it certainly is a skill that has to be exercised externally. And I just think with these three things, sort of knowing what the job of a product manager is, how to decompose what people want into commercial value and behavior change to develop net new novel ideas from first principles absent good data. Um, and being able to inspire people to follow you along the way. Those are the things that are going to allow you to get leverage out of all this code. Those are the things that you need to be doing.

19:53 Those are the things we should be obsessing more about. Those are the things that we should have, I guess, like fancy flow diagrams for. And I wish as an industry we spoke more about that. You know, I think we don't speak about it because it's rare, because it's hard and it's full of failure. Um, but I do think it's the truth of what makes really good product management, a really good product thinking, and an amazing company. So, what does that mean for folks? Like what does that Okay, Claire, I like you said all these big words.

20:21 Fine. What does that mean for me individually? Does that mean we're going to have like this is what people really want? Like Claire, you said product management is dead. Like are we going to have PMs or not? Can I fire them or not? Are my engineers going to be them or not? What is it going to be? Give me the answer. And I did say a couple years ago I thought AI would collapse the product stack. So I did, you know, I hypothesized that the people that could do this job could come from any function. They could come from classic product management if they were willing to adapt their skills. Could certainly come from engineers. I think engineers, you being told you don't have product thinking is a bunch of hoie. Maybe you don't want to have product thinking, that's fine. Um, but that, you know, the fact that we've been convinced that you can't make decisions about buttons for a while is is is no longer. Um, designers can certainly do this. And so I did I did and still believe that the boundaries between roles will collapse.

21:14 Um, but I do still think there is a place for these jobs. And so my updated kind of projection is, you know, people are still hiring for product managers. You say something really extreme and then two years later people like they're still hiring product managers. Product management is not dead. I do think the titles will survive. I think organizations are really comfortable with titles. I think companies are really comfortable with titles. I think the title will survive. I think the job will not. I think the job of a product manager will be very different. I think the job of an engineer already is very different and will continue to evolve over time. I think the job of a designer is very different from when I was a designer working in I promise Photoshop files like artisally rounding corners and doing shadows by hand. I think that while these titles may continue to exist, especially in larger organizations that are slower to adopt new models of working, I think the job is so different. And so in particular product folks, in particular people that work with product teams, if your job does not feel different than the job you were doing a year ago or a year ago, and you're not focused more and more on these things, I think you will end up with a very big codebase and a very little company. And that's the thing that I think we all need to avoid. So, you know, in this world where kind of like anybody can code, npm is dead, what's what stays true? I I think, you know, I put these two icons up here because any chance I could put my like little chat pod next to 60 billion cursor, I'm going to do it. But, you know, like what what do you do when functionally like all our jobs have been compressed into icons on our desktop?

22:55 all of our identities, all of our skills, what what then matters and I think at the end of the day like what matters are users and again as an industry and as a market I think we have spoken so much about the code so much about the infrastructure so much about the how we're building um a little bit about what we're building but very little about who's consuming it and why and I would love to hear us more put all of this intelligence both human and AI against this problem of human behavior change, commercial value, new things that should exist in the world and getting people to believe in that vision. Now, this is not to say that all this like intelligence is not worth anything. Again, I have people in this room have have worked with me. I like to ship. That is not what I'm saying. I'm not saying like I I'm sure what how many people 13 PRs during this talk have been have been, you know, shot up for review.

23:52 You know, I I I love this moment of speed. I have been known to say fast beats right. There's almost nothing in a product manager's toolkit, a team's toolkit, than getting stuff into market fast. That being said, I think you need to know where to aim your slop cannon. Like where is it going? Where are we shooting all these PRs? And you know, I was talking to a friend uh before coming in here and I am sitting on 64 64 PRs right now. Some of which are like good product ideas, some of which are like things that I've just forgotten that we built that maybe we should build, maybe we should not. Some of which we should just definitely merge. They're ready to go. And I think we have this we're we're at risk of developing this churn and this token consumption that actually does not land anywhere in the market.

24:39 And so like I really want people to think about where they're deploying this um this capacity. So, you know, in conclusion, I think you can use AI as a shield just like we use process as a shield. You can use it to ship code. You can use it to talk about activity. You can come up with fancy ways to build. Fine. But I really think you want to think about AI as an accelerant to a vision to impact in the market. And then my favorite and very subtle um c like money. What are we all doing this for if not to build amazing businesses, not just amazing products? And I think um we've seen an example of company that's done just that. I think more and more of us thinking in that model will make sure that we can get great outcomes out of this amazing technology that we're all using. So here's my manifesto.

25:29 Product people, product thinkers, find the money, manifest novel ideas and and then bring people along. That's the job. That's the job when anybody can build and don't just write code. Build build companies. Thank you so much.