Lenny's Podcast: Product | Career | Growth
00:24:25 · Lenny Rachitsky, Casey Winters
[00:24:25] S1: Mhmm. Well, I don't know. Another thought is there's this mean that g p t four is gonna... Or five or six or 10 is gonna replace product managers. And anytime some new feature comes out, people put out these videos of, oh, look. They're doing all the PM's jobs. What's your take on the future of PM in in AI, if any? [00:24:41] S0: Well, I think if you thought the PM job was just filling in the latest, you know, Reforge or Shreyas framework and then getting, you know, that automatic Thang promotion every year and a half [00:24:52] S1: Mhmm. [00:24:52] S0: Then, yeah, you're gonna get replaced by AI. But I think the real PM job is the least likely to get replaced over time because you need, you know, real subject matter expertise. You need to, you know, be trading off a lot of different types of things and and making good decisions for the company and and for your customers. So in terms of using AI now as a PM, I'd actually be cautious in, like, the current iteration of the cycle for PMs. Yep. It's a tool that's trained on sounding smart rather than always necessarily being smart. I was at Eventbrite the other day, and someone was telling me how they were loving the new Notion AI integration. And she asked me if I'd used it. I said, no. She's like, oh, you totally need to. Hey. Go ask you your bio. It's really cool. And I said, alright. Let's just do it on [00:25:39] S1: your screen.
00:25:39 · Lenny Rachitsky, Casey Winters
[00:25:39] S1: your screen. [00:25:40] S0: So she did, and, like, my bio said I started my career at Google and worked on Google Trends and a bunch of other products, none of which happens. Like, it's just complete nonsense. It did have some stuff that did happen. It knew I worked at Eventbrite, and it knew I worked at, you know, like, Pinterest and stuff. But, like, like, half of it was just completely made up. It sounded plausible, but it wasn't actually, you know, true. I think where I'm more inclined for PMs to try to get leverage out of something like GPT four is a lot of the tedious work that's maybe not their specialty to begin with. So, you know, if you're not great at, like, Excel or Google Sheets and you need to model something, you can ask it how to format something. That'll probably be perfect. Or, like, how to get some Zapier integration to work. Like, it probably knows how to do that really well. So I think at this point, it's a pretty good no code, low code tool, and it's obviously gonna become a lot more than that. But for a lot of these other use cases that people are touting, there's a decent chance confidently tells you the wrong thing. And and that would scare me as a PM. [00:26:45] S1: I've had the exact same experience. I had it, describe what I've done, it was, like, 70% mistaken. Yep. Is there anything you've found value in yet with GPT of any kind in your your work, or is it still kind of just poking around in novelty for you?
00:26:45 · Lenny Rachitsky, Casey Winters
[00:26:45] S1: I've had the exact same experience. I had it, describe what I've done, it was, like, 70% mistaken. Yep. Is there anything you've found value in yet with GPT of any kind in your your work, or is it still kind of just poking around in novelty for you? [00:27:00] S0: Yeah. I've had a couple of those things where it's like, hey. How do I actually get this weird thing done? And it's been like, you know, clear bullet points, take these steps, and I'm like, oh, great. Like, this was actually really hard to find on Google. So things like that, I actually quite like it for. So that's where my head goes to, like, more confidently use it. Because, of course, I'll try it, and if it's wrong, I'll learn where So it's it's it's low risk. But it's been right on some of those things, and it's like, oh, cool. I couldn't figure out how to do that. There's something on Google Docs that I need to figure out how to do that was kinda fancy. I asked ChatGPT. They told me how to do it, and I was like, oh, great. [00:27:39] S1: I had a similar experience where I just needed a formula in Google Sheets, and I just told it, like, I need this thing, and it just gives me the actual formula. [00:27:45] S0: Yeah. That's that's so magical when it happens.
00:27:45 · Casey Winters, Lenny Rachitsky
[00:27:45] S0: Yeah. That's that's so magical when it happens. [00:27:47] S1: Yeah. I don't know. I'm not that good with sheets formulas. Shifting course a little bit, you also wrote this really interesting post on founder intuition versus team expertise, and there's a lot to it. And I think it's a really interesting topic because it's this classic discussion that startups have. How much should a founder trust their gut and how much should it be top down? Here's what we should be doing versus the team kinda bottom up, figure out what to build. And so I'd love to spend a little time here. So maybe just to start broadly, what is this... You kinda came up with a framework of how to think about when founder intuition should overrule, say, team expertise and how that changes over time. So maybe just talk about that.
00:28:25 · Casey Winters
[00:28:25] S0: Yeah. If founders are lucky slash good enough to find product market fit, They've usually built up all this intuition about their customers, about their product, and their business that are really hard to explain to others and may even be, like, subconscious. So when they start to hire people, you know, these people will come in with a lot of, you know, excitement, ideas, maybe even really relevant experience, especially, you know, senior leaders. And I've seen some founders are like, cool. I need to let these experts start to own these areas and get out of the details. And I think that's actually the opposite of what you want because none of these people know the business as well as you get. You actually need to direct them until they show you they've really got it and are making better decisions than you would. So this is not a Founder example, but I remember when I hired Erika Warren to build out our loyalty program at Grubhub. She now leads growth at at change.org. And after a month or so, she was like, dude, you're in all my meetings. I got this. And, you know, that let me know she was in a different part of the situational leadership framework than I had put her in. So I backed away, and, you know, she crushed it from there. Founders will put a lot of new leaders in that delegate bucket too quickly. It's like, oh, it's theirs. They got it. They know more than me. When founders actually know the right answer and should just tell them. And this can, of course, change. Like, in the case of Erika, you know, if you hire the right people, they they do get smarter than you. You can and should start delegating. And many founders don't notice the signals where their founder intuition has been usurped by team expertise, or maybe it's just ebbed in effectiveness over time. And and that, of course, makes sense. Right? Founders have all these different things they're dealing with, and then your employees can really dive deep in certain areas. So I encourage employees to proactively signal both directions, like, hey. I need direction from you, or hey. I got it from here. Trust me on this one.
00:30:27 · Lenny Rachitsky, Casey Winters
[00:30:27] S1: Interesting. So you're saying as a founder, the heuristic should be if you feel confident about a decision, generally, you should be clear, like, Hey, I think I know what we need to be doing here. Don't just make people feel better necessarily by just saying, Okay, you tell us what you want to do. And then on the flip side, an employee at the company, if you feel really confident about something, make it clear to the founder, like, I really think this is the right move. Is there anything more you wanna add there? [00:30:55] S0: Yeah. Well, I mean, there's obviously the feedback loop of that, of, like, you know, what happens when you do either of those things. So, you know, depending on the the founder's response to that or the employee's response to that, there's kinda different approaches, you know, you can take. [00:31:12] S1: And part of this, yeah, this cool chart of just, like, over time, the founder expertise becomes less and less relevant. And I guess is that because they just don't spend as much time with the customers because they have other stuff going on where they hire people that are smarter. And then over time, team expertise goes up.