Ex-Discord Product Chief Sellis Models Teams on Insurgent Cells

Snap 前产品负责人 Sellis：科技行业中位水平 PM 净效益为负 Sellis 认为，共识驱动的协作会将执行速度限制在组织中最慢的节点上，并指出可持续的用户增长完全来自于消除核心效用中的摩擦，而非追逐边缘功能。 Peter Sellis 总结了他在 Snapchat 任职七年及掌舵 Discord 产品期间形成的反共识实操手册，主张建立基于独立小组的去中心化团队，依靠坚定的理念统一而非共识驱动的协作。他阐释了为何中位数水平的产品经理净效益为负，管理者为何应在将顶尖人才推向极限的同时包容像 Nikita Bier 这样极具特长但极其偏科的人才，以及为何真正的产品品味取决于团队打造了什么、并特意选择不发布什么。通过对 Snap、Discord、Meta 和 OpenAI 的深入分析，他拆解了广告竞价的结构性现实——对比了 Snap 在年轻受众和相机优先模式下遭遇的变现阻力与 OpenAI 对基于信任指数的竞价机制的需求——并证明了可持续增长几乎总是源于为现有日常用户优化核心产品。最后，他给出了包括核心产品价值（CPV）、优秀 PM 的三大行为矛盾在内的实用框架，并坦率分享了关于野心、人生伴侣以及早期流媒体平台 Ustream 历史的思考。 如果产品组织的架构像游击作战小组而非企业狂热组织，将高度的理念契合与去中心化的单线程决策权相结合，其决策速度会更高。 (5:45) 中位数水平的产品经理净产出为负，因为该职能的基础沉没成本高，人才分布遵循极端的幂律法则，且顶级操盘手往往会转型成为创始人或高管。 (15:16) 对话式 AI 的广告变现需要一种新型的“信任竞价”机制，在数学层面上平衡赞助商影响与用户信任，摆脱社交媒体围绕原始参与度进行优化的模式。 (40:05) 科技业务持久的增长几乎完全源于为现有用户群体优化核心产品和习惯闭环，而非追逐边缘业务、相邻产品或受众群体的扩张。 (50:50)

Former Snap and Discord product executive Peter Sellis argues that exceptional product outcomes require rejecting conventional corporate consensus in favor of ideologically aligned autonomous teams, ruthless focus on core product value, and high-taste restraint. Autonomous Team Cells and Managing Anomalous Talent (5:45) Sellis advocates structuring high-growth product organizations around decentralized cells characterized by deep ideological alignment and unambiguous individual decision rights, a model he provocatively compares to insurgent networks. In his view, standard cross-functional collaboration often introduces crippling coordination drag that slows an organization to its least decisive node. High execution velocity depends on operators possessing a shared mental model of founder intent so decisions occur without bureaucratic oversight. Applying this philosophy to management, Sellis models his approach on NBA coach Phil Jackson, accommodating highly unconventional, "spiky" talent like viral builder Nikita Bier. Sellis contends that organizations fail when they force specialized operators into uniform roles. Rather than spending management cycles trying to rescue underperforming staff, Sellis counterintuitively pushes high performers to their breaking points by continually loading them with higher-order responsibilities while defending them publicly. The PM Talent Distribution and Curatorial Restraint (15:16) Sellis asserts that the median product manager is net-negative to organizational output. He attributes this to a skewed talent distribution created during the zero-interest-rate policy (ZIRP) era: high compensation attracted non-technical talent into ambiguous roles, while the highest-performing product practitioners frequently exited the craft to found companies or assume executive roles. The resulting mediocrity leads mid-tier engineering teams to treat product management as friction, whereas top-tier engineers either direct PMs effectively or refuse to work with weak ones. To counter this baseline, Sellis defines top-tier PM talent by the ability to balance three opposing paradoxes: extreme conviction paired with radical humility in giving away credit, procedural rigor mixed with ease in ambiguity, and patient long-term vision paired with daily urgency. Furthermore, as generative AI democratizes software production and floods workflows with additive feature proposals, Sellis compares product leadership to museum curation. True product taste is defined by restraint—knowing what to cut or withhold from release rather than displaying every artifact produced. Ad Monetization Mechanics and Long-Term Value (20:49) Reflecting on his seven years leading monetization at Snap, Sellis details three structural hurdles that made scaling digital advertising uniquely difficult compared to traditional social feeds: a severe CPM cliff for audiences under 18 and 21 due to privacy and tracking constraints, a camera-first interface that opens directly to creation rather than passive content consumption, and a core utility rooted in private messaging, which offers lower ad-load tolerance per minute. Despite these constraints, Snap scaled to $4 billion in advertising revenue faster than Twitter, Pinterest, or Reddit. Across ad platforms, Sellis anchors product strategy on being "long-term greedy." Because digital ad revenue is causally determined by historical advertiser return on investment (ROI), extracting revenue by artificially inflating clearing prices or cramming inventory degrades long-term net revenue retention. Sellis highlights Meta's modified Vickrey-Clarke-Groves (VCG) auction as the architectural gold standard—balancing user engagement, organic ad quality, and advertiser bids into a unified marketplace that incoming platforms like OpenAI can adopt rather than reinventing highest-bidder mechanics. Compounding Core Product Value and Active Cannibalization (50:50) Sellis maintains that sustainable growth comes from optimizing core engagement loops rather than expanding into speculative vertical adjacencies. At both Snap and Discord, growth was driven by formalizing a Core Product Value (CPV)—a precise definition of the daily utility that brings users back. Snap's CPV anchored engineering on latency and camera speed as the fastest way to communicate with close friends, leading to its post-2018 Android rebuild; Discord scaled by prioritizing voice latency and bug fixes specifically for gaming groups on Reddit. However, focusing solely on the core exposes organizations to Nassim Taleb's Thanksgiving turkey problem, where incremental safety ends in sudden disruption. To prevent this, Sellis points to SpaceX's strategy of obsoleting its reliable, profitable Falcon 9 rocket in favor of Starship. By voluntarily cannibalizing their own category-defining product to fulfill an unconstrained long-term mission, high-conviction companies bypass the innovator's dilemma through deliberate institutional self-disruption. Sellis acknowledges that his operating model creates significant friction within traditional corporate governance: public capital markets often penalize capital-intensive long-term R&D, and identifying the lived, demographic instincts needed for consumer product breakthroughs lacks a neat, standardized hiring framework.

[2:53] Peter Sellis: Oh, that person's probably about to get fired and they're going on and they're pimping themselves out for their next role. [11:19] Peter Sellis: You don't manage Nikita, you only like seek to control him or like direct him. [11:19] Peter Sellis: If your daughter gets kidnapped, like he has a very special set of skills and he's going to employ those skills... and those skills are very hard to come by and very useful in certain situations. [11:19] Peter Sellis: When you like need him to like raise the daughter or like... buy some groceries... all he knows how to do is just like kill everybody in the grocery store. [12:56] Peter Sellis: Dude, if you don't fucking answer my DMs, I'm going to go on Intro and book you so I can just have a one-on-one with you right now. [27:28] Peter Sellis: Subscription products are tough because the math around them is pretty brutal. Essentially, time is your enemy with subscription products... Anything you do takes a lot of time to matriculate in terms of impact. [32:19] Peter Sellis: There was early on, I remember at Snap, it was like so funny to see how aggressively Instagram was copying that they were actually copying some of our worst features—ones that weren't working—just because I think they just like assumed everything that we were touching was turning to gold. [33:32] Peter Sellis: I have tremendous respect for Evan and I think he's essentially been right on almost everything—just early on some stuff. Maybe early on everything, actually... All of the on-device AR that we were doing was at the time more just called machine learning, but employed some form of machine intelligence that was super avant-garde at the time. [46:32] Peter Sellis: Spend time with your winners, ride your racehorses, give a busy person something to do if you want it done quick. [55:34] Peter Sellis: Show us what's messing up. We don't care how small it is, how big it is. We will fix it. [64:31] Peter Sellis: As a product leader, like I aspire to that level of swag where it's just like, 'Yo, we're going to do something that's so aggressive and so ambitious that we can just tell you exactly what [we're] going to do because if you think you can do it better than us, like you're welcome to try.' [73:20] Peter Sellis: Snap is the fastest way to share a moment with the people you care about. [73:20] Peter Sellis: the best way to talk and hang out with your friends before, during, and after playing games. [81:44] Peter Sellis: What you haven't shipped—what you've built and tried and held in your hands, loved, but haven't brought to the world—I think is a test for that. [89:48] Peter Sellis: Time has no signposts upon which to drape duration. From nothing to nothing is no time at all.

[0:00] I asked you, what are some [0:01] counterintuitive lessons that you've [0:03] learned that go against common startup [0:05] wisdom? [0:05] >> I'm no longer beholden to any corporate [0:07] comms team right now, so I can let it [0:09] rip a little bit. [0:10] >> You like to design teams like a [0:12] terrorist organization. A good terrorist [0:14] organization essentially has two things. [0:16] One is it [music] has this ideological [0:18] culture that everyone understands why [0:21] they are doing what they're doing. And [0:23] then it has a very very clear [0:25] organizational structure of who can be [0:28] trusted with which decision. [0:29] >> You like to spend time writing your best [0:31] people into the ground versus improving [0:34] the average people. [0:34] >> If you see somebody drowning, you want [0:36] to save them. I did the absolute [0:38] opposite. If somebody is strong and [0:39] crushing it, I'm just handing them more [0:41] and more rope to hang themselves with. [0:43] something that I know you believe. Most [0:45] product managers are bad [music] and are [0:48] often net negative to the company. [0:50] >> The great product managers are exiting [0:51] the system. They're becoming founders. [0:53] They're becoming executives. It's kind [0:54] of set up to be this profession where [0:58] the median product manager is actually [0:59] probably pretty bad. [1:00] >> Snap is quite famous for not hiring PMs [1:03] for a long time. [1:03] >> The baseline is we should just be able [1:05] to operate this product without any [1:07] product manager. So, this person has to [1:09] provide value over nobody. Feels like an [1:11] interesting new skill for PMs is pushing [1:13] back on things AI adding to your [1:14] products. [1:14] >> I think I [music] saw Evan personally [1:16] say no to better ideas than I've seen [1:19] almost any other consumer startup [1:21] launch. Exercising that kind of muscle [1:23] of saying no of restraint is probably [1:25] one of the few muscles of taste you can [1:27] exercise. If a museum has all of their [1:29] collection on the walls then the curator [1:31] hasn't done anything. [1:33] Today my guest is Peter Cis. Peter is a [1:36] legend in the PM community. He was the [1:38] first PM at Snapchat where he spent [1:39] seven years building one of the most [1:41] delightful and popular consumer products [1:43] in history. More recently, he was head [1:45] of product at Discord, which he helped [1:47] grow significantly. And now he's between [1:49] gigs, which means he was finally able to [1:51] come on the podcast without a comps [1:53] team, without a PR team, which means [1:55] that he's able to be especially honest [1:57] and real about everything that he's seen [1:59] and learned and has [music] been excited [2:01] to share. We cover so much ground from [2:03] how the best teams are organized like [2:06] terrorist organizations, what it's like [2:08] to manage Nikita Beer, why most PMs are [2:11] really bad, why Snapchat [music] has [2:14] struggled to build a truly massive [2:16] business in spite of having a billion [2:18] monthly active users [music] [2:19] and it's not what you think and so much [2:21] more. This episode is full of so much [2:24] gold and that's why it went so long. I [2:26] couldn't stop digging into stuff that [2:27] Peter kept bringing up. A huge thank you [2:29] to Jack Brody and Shiram Krishnan for [2:31] suggesting topics and questions for this [2:33] conversation. Before we get into it, [2:35] don't forget to check out [2:36] lennisproduct.com [2:38] for a free year of the hottest and most [2:40] beautifully crafted AI and nonAI [2:43] products in the world available [2:45] exclusively to Lenny's newsletter [2:46] subscribers. With that, I bring you [2:48] Peter Cis. [2:53] Peter, thank you so much for being here [2:55] and welcome to the podcast. [2:57] >> Thank you, Lenny. It's uh it feels like [2:59] it's been a long time coming uh that you [3:01] and I have known each other, but I'm I'm [3:02] um yeah, I'm really excited to be here. [3:04] >> Uh you've never done a podcast before. [3:06] You've been in product. You've been out [3:08] there for so long. I'm very honored you [3:10] chose to do this. Why have you avoided [3:12] podcasts? Why have you decided to [3:13] finally do this podcast? [3:15] >> You know, it's not it's not avoiding [3:16] podcasts. I've really just been avoiding [3:18] you. No, I'm kidding. Uh it's been um [3:21] you know, I always I I think there are [3:23] two things to be totally frank with you. [3:25] A lot of times when I saw especially [3:26] early on um because you have such [3:28] wonderful distribution among so many [3:30] influential people and teens uh when I [3:32] saw somebody going on Lenny's podcast I [3:34] was like a that person's probably about [3:35] to get fired and they're they're going [3:36] on and they're they're pimping [3:37] themselves out for like their next role. [3:40] Um which was like maybe one view of it. [3:42] I was like why aren't they just [3:43] grinding? Um, now I think I I think uh [3:46] when I when I kind of take a step back, [3:48] um, you know, a lot of the people I've [3:50] worked with and a lot of teams I've had [3:51] kind of the opportunity to lead, I think [3:53] one of the things that I I generally do [3:55] is show up uh pretty authentically every [3:58] every day uh at work. There's you can [4:00] call it no filter, you call it whatever [4:01] you want. And a lot of the stuff that I [4:03] I say and do, I think it's probably [4:05] better suited for for behind closed [4:07] doors. But uh as as as you and I have [4:10] have chatted about now, I'm uh uh [4:12] between different roles. Uh so I'm I'm [4:14] no longer beholden to any uh any [4:16] corporate comms team right now. So uh [4:18] yeah, I can I guess I can I can let it [4:19] rip a little bit. [4:20] >> Okay. Amazing. These are the these are [4:22] the best conversations. No PR people [4:24] involved. No coms people. [4:25] >> Yeah. Nothing to promote either. I have [4:26] no book coming, no other podcast, no [4:29] collaboration. Yeah. [4:30] >> Wow. Just pure pure alpha. [4:32] >> Just pure Yes. About to drop some pure [4:34] alpha, baby. This episode is brought to [4:36] you by our season's presenting sponsor, [4:39] Work OS. What do OpenAI, Anthropic, [4:42] Cursor, Replet, Sierra, Clay, and [4:45] hundreds of other winning companies all [4:46] have in common? They are [music] all [4:48] powered by Work OS. If you're building a [4:50] product for the enterprise, you've felt [4:52] the pain of integrating single sign on, [4:54] SKIM, Arbback, audit logs, and other [4:57] features required by large companies. [4:59] work OS turns those deal blockers into [5:01] drop-in APIs with a modern developer [5:04] platform built specifically for B2B SAS. [5:07] Literally every startup that I'm an [5:09] investor in that starts to expand [5:10] [music] up market ends up working with [5:12] Work OS. And that's because they are the [5:14] best. Whether you are a seedstage [5:16] startup trying to land your first [5:17] enterprise customer or a unicorn [5:19] expanding globally, work OS is the [5:21] fastest path to becoming enterprise [5:23] ready and unblocking growth. It's [5:25] essentially Stripe for enterprise [5:27] features. Visit works.com to get started [5:30] or just hit up their Slack where they [5:31] have actual engineers waiting to answer [5:33] your questions. Work OS allows you to [5:36] build faster with delightful APIs, [5:38] comprehensive docs, and a smooth [5:39] developer experience. Go to works.com to [5:42] make your app enterprise ready today. [5:45] The first thing I want to talk about, [5:46] something that stood out in our jamming [5:49] on what to talk about is uh the way you [5:51] think about designing teams. The way you [5:53] described it to me is you like to design [5:55] teams like a terrorist organization. [5:58] >> Yeah. Yeah. I should probably workshop [6:00] this one a little bit. Like this is [6:01] probably an example of the one that's [6:03] better for behind closed doors, but um [6:05] yeah, when I actually talked to the [6:07] chief people officer at uh at Discord [6:09] about this and uh I think she was like [6:11] simultaneously [6:13] horrified and intrigued uh in a way uh [6:16] which credit to her for for not judging [6:18] me, you know, right off right off the [6:19] bat. Yeah, I think I think it it comes [6:22] from essentially like two um two primary [6:26] beliefs uh from me. Well, one is that [6:30] the there's actually like a a hidden [6:32] cost to collaborating that no one really [6:37] internalizes uh extremely well. Meaning [6:40] like you'll never hear really any [6:43] executive at any company be like stop [6:45] collaborating. Um, and to me that's kind [6:48] of like a like a warning sign. Um, [6:50] collaboration comes with incredible [6:52] coordination costs. And coordination [6:53] costs essentially assure that you're [6:56] moving as slow as the slowest node um in [6:59] the system. And so I try and like [7:02] aggressively uh uh internalize the cost [7:05] of collaboration uh sometimes. And then [7:08] I think the second thing separately from [7:10] this is that at high growth companies [7:12] especially and so many of the people [7:13] that you know listen to your podcast I [7:15] think aspire to be at these types of [7:16] companies product managers should be in [7:18] a position where they're making [7:20] decisions that the founders would have [7:21] made like sometimes like six months ago [7:24] definitely like a year or two ago these [7:26] are decisions that the founders [7:27] themselves would have made and so like [7:30] uh set aside violence for a moment like [7:32] I'm not saying that you know I designed [7:34] my teams for for violence uh but a a A [7:37] good like terrorist organization [7:39] [laughter] [7:40] essentially has two things. One is it [7:43] has this like uh ideological culture [7:46] that everyone understands why they are [7:49] doing what they're doing. Uh and then it [7:52] has a very very clear organizational [7:55] structure of who can be trusted with [7:59] which decision. And I think like [8:02] creating a culture where there's a very [8:04] clear like ideological almost like [8:06] religious level goal essentially that [8:08] the strategy is almost like a religion [8:11] to the entire team and an organizational [8:14] structure where people understand who [8:16] are the directly responsible individuals [8:17] who's singlethreaded on something who's [8:19] the throat to choke use whatever MBA [8:22] blah blah blah ism you want I think the [8:24] combination of those two things can be [8:26] leveraged uh in a way uh such that you [8:29] don't need to over collaborate and you [8:32] can be assured of achieving the results [8:34] that you wanted as if you had [8:36] collaborated. And so like I think the [8:39] components of what is this uh kind of [8:42] like great ideological setup. A lot of [8:45] things I focus on for example um because [8:48] I think often times the head of product [8:50] at um high growth companies where the [8:52] founders are still uh the in in charge. [8:56] The head of product is really kind of [8:57] like a vessel for their vision. And both [9:00] with with Snap and at Discord where I [9:01] had the the privilege of working with [9:03] the founders in both cases, I tried to [9:05] take and absorb and internalize their [9:07] vision and then create like small [9:10] phrases and language that could be used [9:12] and repeated that were essentially [9:14] shortcuts or almost like synctices of um [9:16] like the whole strategy. And so yeah, I [9:19] could go more into that, but that that's [9:20] kind of how I think of it. And obviously [9:22] giving it like branding that, you know, [9:24] I'm designing it like a terrorist [9:25] organization uh at least gets people's [9:27] attention. Um but maybe I need to [9:29] workshop that a little bit. [9:31] >> No, you don't. This is perfect. Uh I [9:33] this is exactly what I was hoping this [9:34] conversation would be. Uh so just to be [9:36] very clear about the elements of the [9:38] terrorist organization that we want to [9:40] keep. Uh one is this idea of autonomy [9:43] essentially kind of trusting people [9:45] further down in case the leaders get [9:48] caught or are incomun communicative. So [9:51] basically giving more autonomy to folks [9:53] further down the chain. And then two is [9:55] having a very clear mission ideology, [9:57] you might say, so that everyone's very [9:59] aligned and decisions are quick and that [10:01] kind of thing. [10:02] >> Yeah. Yeah. And to be clear, I didn't I [10:03] didn't like start with I'm going to [10:05] build a terrorist organization. Um that [10:09] I may [10:09] >> bet you built a great one by the way. [10:11] >> That would be a great one. Uh but uh [10:14] when I when I thought about the elements [10:16] of of what I wanted uh from from kind of [10:19] the organizations I was building, I [10:20] couldn't help but notice that there were [10:22] organizations out there operating like [10:23] this. and they just so happen to be [10:25] terrorist organizations. Um, so yeah, [10:27] that led to a lot of uh interesting [10:28] reading lists. I think the the Peter [10:30] Salis product reading list is like a [10:32] almost like a banned booked list at this [10:33] point. [10:34] >> It's interesting because I've heard many [10:35] people describe you want to build a cult [10:37] and this is not that different because [10:39] you know cults are very effective at [10:41] getting stuff done. [10:43] >> It is not Yeah, you can't spell culture [10:45] without cult. And so I think yeah, it is [10:48] not very different from that. Um I think [10:50] there are elements of a cult that are [10:54] more about the person leading the cult [10:57] whereas uh terrorist organizations are [10:59] more about a higher power or like kind [11:01] of like religion or ideology. Uh so so [11:04] that's why I tended more towards that. [11:06] But um yeah, I think uh I think you [11:08] could you could read up on a couple of [11:09] cults and then just take it all the way [11:11] to the the penultimate moment like [11:13] before they do the thing that makes them [11:14] well known as a cult. Uh and you [11:16] probably could learn a lot. Uh, [11:18] honestly, there's there's a lot of alpha [11:19] in that. [11:20] >> Speaking of um a higher power, uh, a fun [11:23] fact about you is you managed Nikita [11:26] Beer, [11:27] who uh came on the podcast, one of the [11:29] most popular episodes of the podcast, [11:31] quite a character. Uh, was head of [11:33] product for X recently for I think a [11:35] year. Uh, very few people have worked [11:37] with Nikita. I'm curious what it was [11:39] like to actually manage Nikita, work [11:40] with Nikita. [11:41] >> Yeah, Nikita was on on my team at [11:43] Discord. Uh, we acquired uh his startup. [11:46] Um and he was uh on the team discord for [11:48] almost or over a year um with some [11:50] phenomenal engineers that came came [11:52] along with him. Um I think managing [11:55] Nikita that's a very loose term of uh [11:57] use of the word manage. Um I think much [12:00] like yeah you don't manage Nikita you [12:02] only like seek to control him or like [12:05] direct him. Um I liken it a lot to like [12:09] I mean Nikita is an awesome uh [12:11] personality and just actual operator in [12:14] the product space. like pretty much one [12:15] of one. Um, and I have a ton of respect [12:18] for a lot of the stuff that he does. I [12:20] think of him a lot like Liam Niss in [12:21] like the Taken movie series. Like like [12:25] yeah, if your daughter gets kidnapped, [12:26] like he has a very special set of skills [12:29] and he's going to employ those skills [12:31] and um and those skills are very hard to [12:33] come by and very useful in certain [12:35] situations. Uh but then when you like [12:37] need him to like raise the daughter or [12:39] like uh you know buy some groceries or [12:40] stuff like that, all he knows how to do [12:42] is just like kill everybody in the [12:43] grocery store or something like that. Um [12:45] and so no, he's like a yeah, incredible [12:48] person to work with when you have him on [12:50] uh kind of like the areas that he is an [12:52] expert in and um truly like one of uh [12:55] for that from that perspective. [12:57] >> How do you feel about having folks like [12:58] that in your team? Um, you know, a lot [12:59] of people talk about having very [13:01] distinct, you know, person, I don't [13:03] know, roles for different different [13:05] people for different roles and sometimes [13:06] you need someone like that just [13:08] >> I think it's like essential to be able [13:09] to create an organization that can adopt [13:11] that type of talent. I I I wouldn't I I [13:13] view it as like a failure state if you [13:15] have kind of like product orgs that [13:16] can't um essentially handle really spiky [13:19] talent. I mean, Nikita took it to [13:21] extremes. like you remember when he was [13:22] on intro um you know doing that thing at [13:24] one point I like was trying to get a [13:26] one-on-one with him who by the way he [13:27] reported up into me and I couldn't get [13:29] him so I was like dude if you don't [13:30] [ __ ] answer my DMs I'm going to go on [13:32] intro and book you so I can just have a [13:34] one-on-one with you right now [13:36] >> um luckily we we didn't get to that uh [13:38] he he he answered uh I think I hit him [13:40] up on like four different messaging apps [13:41] before I I got a hold of him but um yeah [13:43] I think this ability um to handle [13:46] spikiness uh to to build teams that that [13:49] can essentially adopt the type of talent [13:51] that isn't necessarily like the perfect [13:54] nineto-five employee from a product [13:55] perspective is, you know, I look up a [13:59] lot from like a management style to um [14:02] to Phil Jackson, the the coach of the [14:04] Bulls uh the Lakers. Um like I I think [14:09] his ability to mesh talent of very [14:12] different types uh that literally does [14:14] not get along otherwise uh to create [14:17] these winning teams to have done it both [14:19] as a player uh and as a coach in two [14:21] different places to use philosophies [14:23] from different from eastern philosophies [14:25] western philosophies to combine all this [14:27] stuff like um I I I I think he's like [14:29] one of the most meaningful leaders uh [14:32] from a leadership styles perspective of [14:34] the last half century um and I think [14:36] Like the thing about product like [14:39] especially when it comes to consumer and [14:41] anything that involves more creativity [14:43] and taste is I think of a lot of product [14:45] leaders more like bands putting out [14:48] albums and like you can like love a band [14:52] but sometimes they have a dud album and [14:56] what you do with the band as a fan [14:59] during that time like do you stay a fan? [15:01] Do you still listen to the album? Like [15:02] do you ride it through? Do you do you [15:04] just listen to the old stuff? [15:06] That's kind of how I think of a lot of [15:08] the spikier product managers that I've [15:10] worked with, which is like, man, [15:12] sometimes they whiff on an album, but if [15:13] I still believe in them as an artist, [15:15] like I'll stick with them. [15:17] >> Something that kind of coming back to [15:18] Nikita, but going in a different [15:20] direction when he came on the podcast, [15:22] he had this line that he's like, I'm [15:23] honored to be on a podcast about product [15:25] management uh as someone that doesn't [15:26] believe product management is real. Uh, [15:29] and it's funny that he went on to be [15:30] head of product at XEX, but this [15:32] connects to something that I know you [15:34] believe, which is that most product [15:36] managers are bad and are often net [15:40] negative to the company. Talk about that [15:43] and just like why you think the kind of [15:45] the median PM is is not actually good. I [15:48] feel like this might have been [15:51] a contrarian take during like the Zerp [15:53] era, but now it's like kind of I think [15:55] it's kind of bleeding into the [15:56] mainstream of like holding a higher bar [15:58] >> for a product management. And it is [16:00] funny like you you you have this like [16:02] kind of because you had Tom Tom the CPO [16:04] of um whatnot as well [16:07] >> um say something similar and I think I I [16:09] agreed with a lot of the stuff he said [16:10] and I think by the way I think some of [16:12] his most senior PMs are actually former [16:14] uh SNAP PMs from from my team at Snap [16:15] which is great uh because it kind of [16:17] like I think uh verifies a lot of this [16:20] but [16:20] >> yeah now we're a bunch of self-hating [16:22] PMs who've like uh made a bunch of money [16:25] being PMs are now like oh PMs are bad. [16:27] Um but no, I think there's there's [16:29] actually like kind of a mathematical [16:31] explanation for for this. Um first of [16:34] all, I think any any craft that is uh [16:38] essentially when you're at the top end [16:39] of any like uh kind of skill that is [16:42] normally distributed, you're at the tail [16:44] end uh of the distribution. And so for [16:47] product management, if product [16:48] management is normally the skill of [16:49] product management is normally [16:50] distributed across a population, the [16:51] ones who are actually product managers [16:52] are all the way on the right hand side [16:53] of the distribution. So you lop that off [16:55] and it becomes a power distribution. And [16:57] for for any for any power distribution, [16:59] the median is going to be far below the [17:01] average. Most people are on the left [17:03] side of it. And I think this actually [17:04] gets excessively emphasized within [17:08] product management for for two reasons. [17:10] One, um especially during the ZERP era, [17:13] it became essentially the most lucrative [17:16] career in tech for nontechnical people. [17:19] So once it it has a very subjective [17:22] definition, right, which is part of the [17:23] reason why your your podcast has been so [17:25] valuable to people over the years [17:26] because they're like, "What do I need to [17:28] do to be one?" Um, and so once you're in [17:31] it, you're being compensated essentially [17:33] as if you're you're on a technical [17:35] ladder. Um, so you have kind of a very [17:37] high incentive to remain part of it. If [17:40] you're very good at it, the skills you [17:42] develop as a product leader are are [17:45] almost perfect for either other [17:48] executive roles just like overall [17:50] leadership roles or starting your own [17:52] company. So you're on the right hand [17:55] side of the distribution like the great [17:56] product managers are exiting the system. [17:59] Um they're becoming founders, they're [18:00] becoming executives etc. And on the [18:03] other side of the distribution anybody [18:05] who's in is doing anything possible to [18:07] stay in. So the distribution is be it [18:09] just becomes more weighted weighted [18:10] weighted like this and I think during [18:12] Zer like this you saw this and there [18:14] were memes about it and stuff like that [18:15] and maybe there's been some correction. [18:18] Um but yeah I think mathematically [18:21] uh it's kind of like set up to be this [18:24] profession where the the median product [18:26] manager is actually probably pretty bad. [18:29] Do you feel like it's getting better? [18:32] And do you think there's something I [18:35] think about a lot that um a lot of [18:38] people say that like PM suck, they're [18:40] useless. And I find that that's usually [18:43] the case because they've just not worked [18:44] with a great PM [18:47] and a lot of people haven't worked with [18:48] great PMs. Um and so it kind of like [18:50] makes the whole profession feel look bad [18:52] and sound bad when a lot of people have [18:54] like as you describe are working with [18:55] these kind of mediocre PMs. um thoughts [18:58] there. [18:59] >> Yeah, I think you're you're exactly [19:01] you're exactly right. I I've also like [19:03] kind of noticed this. I feel like so [19:05] much in in the world right now is [19:06] explained by like some form of of like [19:09] the midwip meme um uh in some way or [19:12] another. But yeah, I think you're [19:13] exactly right. Like that is a that is a [19:16] second order effect of of this kind of [19:18] distribution that I think governs the [19:20] the whole profession. Um, I do think [19:23] it's getting better. Uh, because uh I I [19:27] I think again like um when [19:31] a lot of people reference like the ZERP [19:34] era, but I truly feel like when when [19:36] interest rates go down, that means [19:37] opportunity costs are lower. And so when [19:39] opportunity costs are lower, uh people [19:41] are doing things like just hiring that [19:43] extra person, like doing stuff like [19:44] that. But these opportunity costs I [19:46] think are are are higher than are [19:48] internalized because of some of the [19:49] stuff we started talking about that a [19:50] lot of these costs are not are not [19:53] internalized. And so like I've been and [19:56] and we can talk about this but like I [19:57] I've been really focused on developing a [19:59] system for evaluating product managers [20:02] not just the person but the person and [20:04] the place that they are at that time to [20:07] make sure uh that the team is actually [20:09] just like net positive. And in in my [20:12] experience, and we can talk about what [20:13] some of those methods are, but like in [20:14] my experience, like the way this kind of [20:16] ends up is that I really don't hear [20:19] great engineers complain about product [20:21] managers that much because great [20:23] engineers a they know how to bend [20:25] product managers to their will. They get [20:27] the the the use out of them that they [20:28] want and b they just like don't put up [20:30] with bad ones. So they're they they like [20:32] they don't get assigned or they don't [20:33] get to work with bad product managers. I [20:35] think ironically in in kind of the midw [20:38] wee or the horseshoe theory of of things [20:40] really bad engineers also don't complain [20:41] about product managers because they know [20:43] they can just like hide behind their [20:44] decisions and stuff like that. Um so [20:46] it's really just the ones in the middle [20:48] uh [laughter] [20:49] that are complaining. [20:50] >> I do want to ask about Snap. I feel like [20:52] it's a company that hadn't doesn't feel [20:56] like it lived up to the potential of [20:58] what I think Evan and everyone imagined [21:00] it could be. I imagine the main reason [21:03] from and I had Evan on the podcast is [21:05] just the users that it's going after [21:07] just hard to monetize. Is that kind of [21:09] the core of what has kept Snap from [21:11] becoming a bigger success or is there [21:13] more to it? [21:14] >> I I'll try and be as objective as [21:16] possible, Lenny, because like this is [21:17] like a little bit of my life's work and [21:19] maybe I do have some some level of like [21:21] chip on my shoulder or or Yeah, I'll try [21:24] and be as objective as possible. So, [21:26] I'll say I'll say two things, maybe [21:30] three things about Snap from a [21:33] monetization perspective, at least an [21:35] ads monetization perspective, and then [21:36] we could talk about like they've had a [21:38] lot of success with Snapchat Plus and [21:40] subscription products in general are [21:41] doing well, but ads are truly the [21:43] magical kind of like product market fit [21:45] of consumer monetization on the [21:46] internet. And there's a lot of [21:47] mathematical reasons why that's true. [21:49] And there are three things that hold [21:50] back Snap from an ads perspective that I [21:53] think the company like we we still kind [21:56] of talk about is we deserve a little bit [21:58] more credit maybe than we get but I [22:00] understand why we we don't. Um the first [22:02] is yes the the user base is is extremely [22:04] young like Snap appear appealed to Gen Z [22:07] at an excessive rate uh early on now [22:10] even I think Gen Alpha as they like kind [22:11] of graduate from like Roblox and stuff [22:13] like that. Um, and there is a a step [22:16] change, like literally a suspicious [22:17] discontinuity in advertising CPN's [22:20] between 17 and 18 and between 18 and 21 [22:23] in the United States and abroad. Um, if [22:26] if you're under 17, the rules around how [22:29] you can track and measure uh and target [22:31] ads are extremely uh extremely [22:34] constraining and justifiably so. Um, [22:36] right to protect teenagers. Um, that [22:39] means that it's essentially much harder [22:41] to drive ROI uh from from this group of [22:43] people versus an 18-year-old, versus a [22:45] 21-year-old and 24 year old. And you can [22:47] say all you want about like, oh, that, [22:48] you know, teenagers influence their [22:49] parents, blah, blah, blah. The key thing [22:51] to think about is like those are just [22:52] harder to measure and it's harder to [22:54] build performance products around things [22:55] that are hard to measure. And so, um, [22:58] yeah, that that Snap's audience is [23:00] definitively like less valuable, um, [23:02] from that perspective versus like a [23:04] Pinterest or an Instagram, uh, or or a [23:07] Twitter. The second thing is almost [23:10] every ads business you'll ever every [23:13] every ad product you'll see that isn't [23:15] search based, um, every display ad [23:18] product opens to the screen where you [23:20] put the ads. There's a lot of studies [23:22] around attention and just mathematically [23:25] where you're going to get the [23:26] impressions where you need to kind of [23:28] like maximize that for advertising and [23:31] Snap opens the camera. Um, and that [23:34] makes it awesome as a product, right? [23:35] Like in terms of its ability to be the [23:37] fastest way to capture a moment and [23:38] share it with your friends, but in no [23:40] part of that flow, the core flow of [23:41] taking a snap and sending it to your [23:42] friends is there like a logical place [23:44] for an advertisement. It is just not a [23:46] feedbased product and it is not a [23:49] consumptionoriented product. It is a [23:51] creationoriented product, a visual [23:52] creationoriented product. Uh, and there [23:55] don't really exist uh a lot of [23:57] advertising kind of formats that um that [24:00] go after this. And we experimented with [24:01] this a lot at Snap. And probably one of [24:03] my biggest failures was the ability to [24:05] drive camera monetization. It was it was [24:07] just too hard. Like we I I couldn't [24:09] figure it out or I couldn't greenlight [24:10] the teams that could figure it out. And [24:12] um yeah, so it's a very hostile [24:15] environment for for advertising from [24:16] like that initial thing. And then at its [24:18] core, it's a messaging app. And like [24:21] there aren't really great ad products [24:23] for messaging apps right now. I think [24:26] even the the monetization of certain [24:27] messaging apps via subscription or via [24:29] business accounts. And people will [24:31] always cite like some of the Asian chat [24:33] apps as examples. You peel it back and [24:35] from a like dollars per time spent [24:37] perspective, none of these are doing [24:38] that well. Like none of these are like [24:40] exceptionally great at monetizing. Uh [24:42] message apps are incredible utilities [24:43] are incredible to have as part of a [24:45] portfolio. But Snap was trying to build [24:47] a multi-billion dollar ads business very [24:50] quickly uh on uh a lower value um [24:55] audience, a very tough format uh for ads [24:59] and and a service that was more [25:01] utilitarian than uh consumptionoriented. [25:03] >> Wow, that was fascinating. Uh like I [25:06] think it's important to note Snap is [25:07] doing it's like a many multi-billion [25:10] dollar business. I think a billion [25:13] monthly active users you know like all [25:15] the teens use it's like [25:17] >> like in the scheme of things it's very [25:19] successful. It's more just everyone [25:21] thought it was gonna be much much bigger [25:22] like the next Facebook [25:24] >> and and I think it yeah it really if you [25:26] were to peel apart Snap as a product [25:30] it's it's uh again some of my [25:32] information is dated and I I haven't [25:34] really like looked at but like it's the [25:36] number one messaging app among teenagers [25:38] in almost every western country like 70 [25:41] to 100% penetration of you know people [25:43] under 21. Uh it's their it's their [25:46] default map. gets used more than Apple [25:48] or Google Maps by by the this group. Uh [25:51] I think it's their the most popular [25:53] astrology app for for most of these. [25:56] It's definitely the most common camera. [25:57] So like imagine if you're like hey [25:58] there's a holding company. There's a [26:00] there's a a a Jenzie Bergkshire Hathaway [26:02] that has the number one camera app, the [26:04] number one messaging app, the number one [26:05] maps app, the number one astrology app, [26:07] the number one content because like I [26:09] guess Tik Tok would be the number one [26:10] content consumption one, which is part [26:11] of the problem with the ad stuff, but [26:13] you peel it apart and you're like, [26:15] that's an insane consumer product track [26:18] record. And and none of this credit goes [26:19] to like me at all, right? This is like [26:22] the that OG design team that cranked out [26:24] some of the biggest hits of like our [26:26] generation. But I think from an ads [26:28] product like maybe yeah the chip on my [26:30] shoulder is like Snap started after [26:32] Twitter, after Reddit, after Pinterest. [26:34] Uh it went public before all of them [26:37] except for Twitter. It it was quicker to [26:40] four billion I think in ads revenue than [26:43] all of them uh despite starting later. [26:46] Uh and like what is SN like Twitter is a [26:49] global town hall. Like all the most [26:51] valuable people on the planet are [26:52] talking on Twitter. Pinterest is [26:54] literally a place where you tell it what [26:55] you want to buy, right? Like these are [26:57] really great places for for ads. What [27:00] was Snap known for at the time? I I [27:02] don't know. Like I don't want to make [27:03] your podcast brand unsafe, but you know, [27:05] we can we can talk we can talk about [27:07] what Snapchat was known for at the time. [27:09] >> Disappearing [laughter] photos on [27:11] screen. [27:11] >> It was still able to scale this ads [27:13] business, which I think is like Yeah, [27:15] it's pretty fascinating. But yeah, it's [27:17] a fascinating company. And I think just [27:18] given the der of like 100 million DAO [27:20] consumer products that have been started [27:22] in the United States uh since then um it [27:25] still kind of like deserves to be [27:27] studied. [27:28] >> Yeah. When when Evan when Evan was on [27:30] the podcast I pointed out that I think [27:32] it was the last big successful social [27:36] consumer product to launch which was 12 [27:38] years ago I think other than Tik Tok [27:40] which is not like social really. It's [27:42] more of a like content media sort of [27:43] thing. [27:44] >> Yeah. Discord. Discord technically [27:46] started in 2015. [27:47] >> Okay. [27:48] >> Uh [27:49] >> but uh very very very like a a niche [27:53] product, right? Like the world's biggest [27:54] niche is gaming. Um but yeah, quite a [27:56] quite a different vibe. [27:58] >> Just like maybe a final thought here. [27:59] I'm curious to get your take. You could [28:01] almost argue that this is a example of [28:04] more PMs would have been helpful because [28:08] uh it's such an incredibly beautifully [28:11] designed product and experience and so [28:13] delightful and people love it and are [28:15] hooked on it, but it's just the [28:17] monetization and the business basically [28:19] was not solved. And you would think [28:21] that's where PMs would come in and so I [28:24] wonder if that was uh a hindrance is the [28:27] low number of PMs involved. I think that [28:30] Snapchat plus the subscription product [28:32] if it had launched subscription product [28:35] are tough because the math around them [28:36] is pretty brutal. Like essentially like [28:40] time is is your your enemy with [28:43] subscription products like [28:44] >> turn. [28:45] >> Yeah. Anything you do takes a lot of [28:47] time to to metriculate in terms of [28:49] impact. So it I think Snapchat Plus has [28:53] actually been quite successful and [28:54] there's a lot of how it's been executed [28:57] that's really really really great but I [29:00] I sometimes wonder if it had been [29:02] launched four years earlier um like [29:04] where it would be um just because like [29:07] with with messaging utilities and you [29:09] see this with with Discord and Nitro uh [29:12] as well much different than content [29:14] products like Spotify or or Netflix um [29:17] you first kind of like become acclimated [29:19] to using the product daily and getting a [29:20] lot value out of it and that's when you [29:21] see the value of the subscription [29:23] product. Um, not like right when you [29:26] start, right? You're not like you don't [29:27] open Snapchat the first time. You're [29:28] like, "Oh, I definitely need to pay, you [29:29] know, 20 bucks a month or five bucks a [29:31] month for something." Um, and so time is [29:33] like kind of like the cohorts [29:36] essentially penetration grows slowly [29:38] over time, but um, you know, grows grows [29:40] quite linearly usually. And so yeah, if [29:43] Snap had had this product out four years [29:45] ago, there might have been less pressure [29:46] on the ads business to scale as quickly [29:47] as it does. less pressure on the ads [29:48] business to scale as quickly as it does [29:50] allows for more freedom on the content [29:51] consumption side of things and they [29:53] experience and allows them maybe to [29:54] compete more with Tik Tok. I don't know. [29:56] It's a Yeah, it's a it's a big brain uh [29:59] system thinking uh like okay snap needs [30:02] more PM but uh yeah interesting [30:04] >> man systems thinking that's like I think [30:05] the seventh time in a row that term has [30:07] been used in this podcast [30:08] >> one of the um was it was it the Netflix [30:11] >> Yeah about this [30:13] >> that was a big one [30:14] >> yeah sometimes you people say systems [30:17] thinking then they go out and buy that [30:18] book uh [30:19] >> yeah with the with the slinky [30:20] >> yeah it's like I'm a systems thinker now [30:23] uh like uh yeah I think it Actually the [30:25] most underrated aspect like of of [30:28] systems thinking is is is having a [30:32] mathematical background to be able to [30:34] describe the the system in mathematical [30:37] terms. um like to actually map it out in [30:40] terms of like stocks and flows and and [30:42] you know different probability like is [30:43] this biodal or is this normally [30:45] distributed and just being able to [30:46] understand like to be able to [30:48] mathematically communicate how a system [30:50] works because otherwise like I think [30:51] it's this famous method it's a very like [30:53] MBAish type [ __ ] thing but [30:55] ironically it kind of works I think it [30:56] came out of like Toyota or something [30:58] like the five W's and you're like oh [31:00] what's the five W's must be this super [31:01] amazing system syncing method it's [31:03] apparently just asking why five times [31:05] [laughter] and like then you get to the [31:06] bottom of things And I'm like, that's [31:08] like that's how the majority of people [31:10] when they say it's systems thinking, I [31:11] think they're basically just saying like [31:12] I asked why five times and I got to the [31:13] first principle. [31:14] >> That is awesome. I I was gonna ask you [31:17] to help people understand what is [31:19] systems thinking because it's this broad [31:21] term that is hard to really know. And so [31:23] I love this answer. So you're finding [31:25] basically just ask why is that is [31:26] actually a really effective way to uh [31:28] develop systems thinking to actually [31:30] think about the bigger picture. [31:31] >> Yeah. But I would I would love it if [31:33] like maybe rather than reading the [31:35] Slinky book, there was more like Jay [31:37] Forester system dynamics, you know, from [31:39] the like manufacturing uh kind of stuff [31:41] from the 1970s and then I think just [31:43] like a mathematical like understanding [31:46] of of how the system works like if you [31:48] had to model it. Um [31:49] >> yeah, I was going to ask basically [31:51] >> just create like a spreadsheet model of [31:53] how [31:53] >> exactly. Yeah. Yeah. I think could be [31:56] very clarifying from a [31:57] >> that is actually okay. So that's a [31:58] really good way to understand systems [32:00] thinking. Model it out. put in a [32:01] spreadsheet what are all the variables [32:03] what what comes in what comes out what [32:04] can what are the levers [32:05] >> yeah yeah what are the levers [32:07] >> what are the distributions behind those [32:08] levers um [32:10] >> yeah I think it can it can be very [32:12] clarifying even if you even if the [32:13] spreadsheet goes nowhere other than you [32:15] know your your shared drive like it can [32:17] be very clarifying [32:18] >> feed your agent um I was going to say [32:20] real quick on Snapchat plus the my [32:22] favorite part of the story is that [32:24] Instagram launched Instagram plus it's [32:26] just like another [laughter] [32:28] classic Snapchat story [32:30] >> yeah I the there was early on I remember [32:33] a snap it was like so so funny to see [32:37] how aggressively Instagram was was [32:39] copying that they were actually copying [32:41] like some of our worst features like [32:42] ones that weren't working just because I [32:44] think they just like assumed everything [32:45] that we were touching was turning to [32:46] gold [32:48] >> yeah and then there are some features [32:49] that on Instagram just don't make as [32:51] much like the map and things like that [32:52] that will take time but when it comes to [32:53] stories which was obviously extremely [32:56] effective from a copying perspective [32:57] you'd be hardressed to to say anything [32:59] but the version of it on Instagram is [33:01] now much better. Um like what Instagram [33:04] has done with that format like it really [33:06] suits the product super well. Um and [33:09] that team is powered by a bunch of great [33:13] designers and uh when they start [33:15] thinking from from first principles and [33:17] and with the taste that Adam who was [33:19] also on the pod has uh like it's a that [33:24] product is well architected from a human [33:27] computer interaction standpoint in a way [33:28] that I think is actually like really [33:30] admirable on many levels. I I keep [33:32] trying to move away from Snapchat, but I [33:33] have so many I have new questions come [33:35] up as you share more. Uh something that [33:38] came up in my chat with Evan about this [33:39] is like seeing everybody copying [33:42] Snapchat for 15 years or however long [33:44] it's been. His takeaway basically is [33:46] software is no longer remote. You have [33:48] to focus on other things. Uh he [33:51] described that's why he's spending so [33:52] much time on hardware. It was his answer [33:54] and also just like the ecosystem and the [33:56] network effects of a product. Thoughts [33:58] on that? Uh, I have tremendous respect [34:00] for Evan and I think he's essentially [34:02] been right on almost everything just [34:04] early on some stuff. Um, maybe early on [34:07] everything actually too. Um, even even [34:09] stuff like avatars like u um AI like [34:13] essentially all of the like ondevice AR [34:16] that we were doing was at the time it [34:18] was more just called machine learning [34:19] but like employed some form of machine [34:21] intelligence that was like super out on [34:23] guard at the time. So, I think when he [34:25] says something like this, there there's [34:27] probably some truth to it. I think a lot [34:29] of people are are are frustrated with [34:32] Snap because it doesn't have the [34:34] financial profile to be able to support [34:36] the type of hardware development that he [34:39] wants to do, it seems. Um, and that like [34:42] there still isn't like a great answer on [34:44] the monetization front for which like [34:46] obviously like I bear like a lot of [34:49] responsibility. But from that kind of [34:51] like moat perspective, like what he [34:53] said, I I watched the pod. Um, [34:56] yeah, I I I kind of think like he's [34:59] right, but that the like maybe investors [35:01] or the the the world doesn't want to [35:03] give him the space to to be right in [35:06] some ways. Um, which has [35:09] >> Yeah. Yeah. It has like uh like artist [35:12] vibes, you know, like rejected artist [35:13] vibes uh or or um on the outskirts [35:16] artist vibes, which like I think there [35:18] there's still so many things like it it [35:20] actually reminds me like there's so much [35:22] talk these days about taste and and [35:25] things like that and [35:27] so hard to describe things like but one [35:29] of the things that I learned at Snap is [35:32] um like I I think I saw Evan personally [35:35] uh say no to better ideas than I've seen [35:39] almost any other consumer startup [35:42] launch. And it kind of made me think of [35:44] like this like my goal as like a product [35:46] leader for for founders like at these [35:49] high growth companies was always to like [35:51] turn them into almost like the best [35:53] museum, the the richest museum curators [35:55] in the world where like whatever is in [35:57] the product at any one time is just a [35:59] small slice of the overall collection. [36:01] It's whatever that curator thinks is [36:04] right for this time and and place in [36:06] society. And I think if like one way you [36:10] can kind of like get at somebody's taste [36:11] is uh when you're interviewing them is [36:13] just ask them about some of the things [36:15] they built that they never shipped. Um [36:17] not that they tested and got bad results [36:18] for but like that literally they they [36:20] held it in their hands the prototype and [36:22] they're like the world's not ready for [36:23] this or like this thing is not ready. [36:26] And if they have a lot of high quality [36:27] ideas that are on the shelf, exercising [36:29] that kind of muscle of like saying no uh [36:32] of restraint is probably one of the [36:35] easiest like one of the few like muscles [36:37] of taste you can exercise. If if a [36:40] museum has all of their collection on [36:42] the walls, then the curator hasn't done [36:43] anything. And so uh yeah, I think about [36:46] this a lot um because Evan is yeah, [36:49] probably the best at it. That is such an [36:51] interesting insight, especially these [36:53] days because AI is very bad at removing [36:57] things from designs and from products. [37:00] It always adds it doesn't say, "Hey, [37:02] maybe we should cut this piece." And so [37:05] it's so interesting how this connects [37:06] now to the there's like the we talk [37:08] about taste all the time with AI and [37:10] this is like an interesting element of [37:12] that of get being good at understanding [37:14] what to remove as AI becomes a bigger [37:17] part of the building process and [37:19] designing process. Yes. Yeah. This will [37:22] be [37:24] this will be hard I think to this is a [37:27] hard skill for for AI to replicate. I I [37:30] think like in general [37:32] it it seems like at right now it's [37:34] obviously at the code level. It's [37:36] getting to the design level. It's it's [37:38] in some cases at the video and the the [37:39] the photo level. Um it'll eventually [37:43] maybe even get to physical level as [37:44] well. But uh it's really going to [37:46] empower anybody to to be a creator. Uh [37:49] and we we've seen this in the past, [37:51] right? Like lowering the threshold for [37:52] creativity not just um kind of like [37:55] enables the existing creators to do more [37:57] uh but enables people who were otherwise [37:59] uh kind of like um bounded out or or [38:03] restricted from entering the the arena [38:05] uh to enter it. Um but uh this will put [38:09] a lot of pressure on essentially [38:12] creating stuff that deserves human [38:14] attention. Um in a way that I I wonder [38:18] if there are some analoges to this in [38:19] the past. We we'll have to think about [38:21] it. [38:21] >> And it just feels like an interesting [38:23] new skill for PMs is pushing back on [38:25] things AI adding to your products. [38:27] Engineers are like what if we had this [38:28] whatever. So it's like it's not even on [38:30] AI right now. It's on the on the product [38:31] teams basically to to cut stuff. Anyway. [38:34] >> Yeah. Yeah. Maybe the con [38:38] it'd be pretty funny if the conclusion [38:39] of every every lending podcast is [38:41] basically, yeah, the world might just [38:42] need more PMS. [laughter] [38:45] Crazy. [38:47] >> Oh man. [38:48] >> This episode is brought to you by [38:49] Mercury. Radically different banking now [38:51] with spend. I've been a Mercury customer [38:54] for so many years now. I switched all my [38:56] business banking to Mercury. And [38:58] honestly, I could not be happier. It's [39:00] what online banking feels like when it's [39:02] built by product people, not by bankers. [39:05] And now with spend, you can give your [39:07] team individual cards, set spending [39:09] limits per person or per team, and have [39:12] expense receipts automatically pulled in [39:14] from Gmail or over text. You can even [39:16] give your AI agents their [music] own [39:18] cards with their own limits and [39:19] policies. Most founders start out the [39:21] same way, one card used by everybody at [39:24] the company. It works until it stops [39:26] working. Someone goes over, a receipt [39:28] disappears, you spend two days trying to [39:30] figure out who spent what and why. Spend [39:33] is expense management built directly [39:35] into Mercury. All your team's cards, [39:37] budgets, and reimbursements. All live in [39:39] the same place as your business banking. [39:41] No chasing, no manual reviews, no end of [39:44] month scramble. The result is a team [39:46] that can move fast and a founder who is [39:48] no longer the bottleneck. Learn more and [39:50] get signed up at mercury.com. [39:53] Mercury is a fintech company, not an [39:55] FDIC insured bank. Banking services [39:57] provided to Choice Financial Group in [39:58] column A members FDIC. The IO card is [40:01] issued by Patriot Bank and a member FDIC [40:03] pursuant to a license for Mastercard [40:04] International Incorporated. There's a [40:06] couple elements here that I think [40:07] connect Snapchat and AI. Uh having seen [40:11] the ads experience of Snap and where [40:13] that has gone and not become as big as [40:16] people may have thought. Uh Chad GBT in [40:19] particular is exploring ads. I imagine [40:21] other AI tools will. Thoughts on that [40:23] opportunity? Do you think that's [40:25] actually a massive business opportunity [40:26] for these companies or do you think [40:27] they're not understanding the actual [40:29] business there? [40:30] >> I I I I still I subscribe to like the [40:35] the kind of like Eric Syfort [40:38] uh side of the world of like ads are ads [40:42] are the product market fit for the [40:44] internet. And I still think that uh from [40:47] a monetization standpoint um there [40:50] there's there's a number of reasons [40:51] again mathematical for for why that is [40:53] true and why I think like um like [40:56] subscription products and things like [40:57] that are kind of overrated versus ads [41:00] from a consumer side of things. When I [41:02] look at what OpenAI is doing from an [41:05] advertising perspective, they are they [41:07] are speedrunning all of the like ads [41:12] infra that it is exactly what you would [41:16] want and do at Discord. I I call this [41:19] like last mover advantage. Uh because [41:21] like we've we've learned so much and [41:24] we're at like a certain place with [41:25] online advertising. Um, and really Meta [41:28] and the team there from like 2010 [41:30] through 2020 really deserves I think [41:31] most of the credit for this. Like uh [41:34] like what drives a modern ads [41:36] marketplace and auction is essentially [41:39] like a solved problem and and OpenAI is [41:41] speedrunning all of the components of [41:42] that. And I do think that the the [41:46] mechanism the mathematical mechanism [41:47] which is essentially like a modified uh [41:50] VCG I think Vicky Clark's Grove uh kind [41:54] of like auction mechanism. this [41:56] mathematical mechanism is the correct [41:59] way to build an ads auction in uh even [42:03] like a chat environment in chat GPT. And [42:05] what I mean by that is I think where the [42:07] innovation came from from meta was not [42:09] just that like the highest bidder and [42:11] I'm oversimplifying like crazy but like [42:13] not just that the highest bidder wins uh [42:15] but that the organic value of an [42:17] advertisement and an advertiser to that [42:19] user should be factored into their bid [42:21] such that you essentially have to uh pay [42:24] more as an advertiser to reach people [42:26] who who may not want to be reached or or [42:28] may be harder to reach or may not [42:30] necessarily be as relevant uh to you. [42:33] And this also is great because it allows [42:35] you to mox uh both uh content, organic [42:39] content and advertising content on a [42:40] similar um in a unified auction uh with [42:43] dollars as essentially the ultimate [42:45] denominator. And the reason that I say [42:48] that this is good for chat is not [42:50] necessarily that organic value is what's [42:52] important here, like how much you like [42:54] an ad or not in a content feed, but in [42:56] chat it probably feels like this could [42:59] be replaced or augmented with [43:00] essentially trust. Like if you boil this [43:03] down to something very simple uh it's [43:05] just like hey what X should I buy? What [43:07] what mic should I buy? And in that [43:10] moment uh we have to decide uh like [43:14] should I give you the objective answer? [43:15] Should I give you a set of answers? And [43:17] how influenced should I be by a set of [43:18] advertisers that want to get in front of [43:20] you? And I think the mechanism by which [43:22] how influenced I should be should [43:24] essentially be how much trust I'm [43:26] trading off. And you kind of see us [43:29] failing at this a little bit probably [43:32] and I don't want to I don't like I [43:33] really don't want to pick on anybody [43:34] because I I know how hard it is to to [43:36] build these types of things but like a [43:38] little bit on Amazon search um in terms [43:41] of like the ads the sponsored experience [43:43] there. Um it's not like it's not it's [43:46] not threatening the product. Amazon is [43:47] such a great product but what is the [43:48] great product of Amazon? It's like the [43:49] selection and the next day shipping or [43:51] whatever. Like it is getting a little [43:53] crazy when you search for something and [43:54] you have 16 ads before you get to the [43:56] first uh thing and you're like, "Oh [43:58] shoot, is that like subtly influencing [44:01] my my trust in the system?" And so [44:04] that's why I actually think like OpenAI [44:06] is speedrunning all the right stuff. And [44:09] where the the push is going to come to [44:11] shove is uh how they instantiate it [44:15] mathematically in their auction [44:16] mechanism to maximize trust over the [44:18] long term. Um because trust maximization [44:21] will probably be correlated with user [44:23] retention um if not causal and with uh [44:28] advertiser value. And then on the other [44:30] side of it, it's probably a format [44:31] question um like what does this format [44:34] look like in in chat? Are we still going [44:37] to be what if what if it's voice? Like [44:38] we don't know. So I I think like one of [44:40] the coolest jobs right now for a PM uh [44:43] and designer would probably be like add [44:45] formats at OpenAI. Um, that'd be a that [44:49] that's like a that's like a societal [44:51] level job in advertising right now. Um, [44:55] and so, uh, yeah, I think I think it'll [44:57] be really interesting to see how they [44:58] how they work on that. Um, [45:01] yeah, maybe one thing one other thing [45:03] I'll say is like I was kind of [45:04] disappointed that they jumped right to [45:07] an adree experience for the top tier. [45:10] Um, I feel like ads could actually be [45:13] really additive to every part of the the [45:16] OpenAI, you know, not like coding and [45:18] things like that, but like actual chat, [45:19] like um, let's set aside work and [45:21] coding. Um, and like lpping off your [45:25] most valuable users so early. Um, I was [45:28] I was a little disappointed, but I [45:30] understand why people do that. [45:32] >> It is so fun to listen to you talk about [45:34] this as you take a swig of your vodka [45:36] right there. I don't know if people [45:37] [laughter] [45:38] >> It's actually really really fancy douchy [45:40] water. So, uh [45:41] >> Okay. Yeah. That you replace with vodka [45:44] so nobody can tell. This is the key. [45:46] >> That's right. If you keep me on for [45:47] another 30 or so so minutes, it'll [45:49] really get [45:50] >> Yeah. [45:51] >> Uh so, what I was saying, it's so it's [45:53] so fun to hear you talk about the ads [45:56] ecosystem and it it shows how much you [46:00] know and how deep this goes. Like I feel [46:02] like we could do a whole episode on [46:03] this. Probably don't need to do that [46:04] right now. Uh, but it's amazing. Uh, [46:07] clearly, you've spent a lot of time [46:10] thinking about ads. Uh, what I'm also [46:12] feeling now is OpenAI needs to hire you [46:13] immediately to help them with this. [46:16] >> Yeah. Yeah. Yeah. I guess free agent, so [46:18] you know, this is maybe that's my my [46:20] ulterior motive of the pot. You know, [46:22] >> this is the interview. Well, I think [46:24] you've passed. Okay. Um, I want to go to [46:26] something else. I so we were talking [46:28] chatting about what to spend time on and [46:31] uh I asked you what are some [46:32] counterintuitive lessons that you've [46:34] learned over the course of your career [46:36] that [46:37] >> go against kind of common startup wisdom [46:39] common leadership wisdom and there's [46:41] four you shared and I want to just go [46:43] through them and get your take because [46:45] they're awesome. Okay. So, the first is [46:48] uh you like to spend time writing your [46:51] best people into the ground versus [46:54] improving the average people. [46:57] Sure. [46:59] [laughter] [46:59] >> I'm realizing you write this out. It's [47:01] like uh you he's building terrorist [47:04] organizations and all he does is ride [47:05] his best people into the ground. [47:06] >> Going to be the head the title of the [47:08] episode. [47:08] >> Yeah. Yeah. You're going to need to like [47:10] there's going to be like a like a [47:12] recovery group like a therapy group of [47:14] PMs that worked for for Peter Cis. Um, [47:17] >> no, you're both probably worried about [47:19] that. [laughter] [47:21] >> They're all sitting in Discord chatting. [47:23] >> Yeah. Oh, god forbid it's a Discord [47:25] group. Oh, Jesus. Uh, yeah. The um Yeah, [47:28] I think I think this this actually kind [47:31] of flows from the the the whole um the [47:34] kind of like VORP thing. Um is that like [47:37] once you have somebody in a a position [47:39] and there there are a lot of tropes [47:40] around this too, right? Like if you want [47:41] something done quickly, give it to the [47:43] busiest person uh kind of thing. Um, but [47:46] I think it's it's very it's not [47:48] necessarily counterintuitive. [47:50] It's counter to like human nature, [47:52] right? Like if you see somebody [47:53] drowning, um, you you like you want to [47:56] save them. And, uh, like I I I did the [47:58] opposite, absolutely opposite. I I just [48:00] like kind of, um, and and maybe it's, [48:03] you know, we were right before we [48:04] started we started the podcast, we I was [48:06] talking about nominative terminalism [48:07] where like your name kind of like [48:08] creates who you are and like joking that [48:10] maybe I need to be a salesperson because [48:11] of my last name is Celis. But there's [48:13] also like a this uh you know MBA Peter [48:15] principle where you're like people are [48:17] promoted to the level of their [48:18] incompetence which definitely could [48:20] describe my career. But like that's like [48:22] kind of what I seek to do. It's like if [48:24] somebody is is strong and and and [48:26] crushing it, I'm just like I'm just like [48:28] handing them more and more rope to hang [48:31] themselves with. Um and I'm I'm I'm [48:34] pushing them as hard as possible. Um you [48:38] know, like to the to the point of of of [48:40] breaking. Um because like at the end of [48:43] the day like one of the biggest [48:45] leadership like principles for me is [48:47] that like I I will the people on my team [48:49] I will always defend in in public like [48:52] in front of exacts and stuff like that. [48:53] I never never criticize them in front of [48:56] others. And so like people who are [48:58] making consistently good decisions I [48:59] just want to load them up with more [49:00] decisions till I start to see them make [49:01] bad ones. Um, I think like there's so [49:04] much kind of like self [49:06] self-reinforcement or or self-fulfilling [49:08] prophecy of like when you have [49:09] confidence and when things are going [49:10] well, like you're able to do more. And [49:13] so, um, yeah, I think I like try and [49:15] really show that I believe in people. [49:17] Um, and when I don't, um, yeah, maybe [49:20] this is just a personal thing. I just [49:22] don't have a really good mechanism for [49:23] for helping people improve. So, I I I I [49:26] essentially like let them fail on their [49:28] own. um which could be I think probably [49:30] pretty frustrating uh for me as a leader [49:32] to work for. Um and and some somewhat [49:35] sad. Um I think the thing you need to [49:37] control for when you adopt this type of [49:39] management philosophy is that not paying [49:41] attention to somebody doesn't mean [49:42] they're failing. Um like there there are [49:44] a lot of people that I've hired over the [49:46] years, especially if I've worked with [49:47] them multiple times, where I'm like, I [49:49] know you're you I know you know what to [49:51] do, so go do it. And I I just like let [49:54] them execute. But um yeah, it's uh you [49:57] know, spend time with your winners, ride [49:59] your raceh horses, give a busy person [50:00] something to do if you want it done [50:01] quick, whatever the whatever the stupid [50:03] trope is. Um yeah, I try to pursue it [50:06] pretty aggressively. [50:07] >> Makes sense to me. And uh there's many [50:09] ways to say it. The one some are some [50:12] are kinder than others. Uh another uh [50:15] contrary take you have is that [50:16] collaboration is a pretty bad use of [50:18] time a lot of times. You talked about [50:19] this earlier, but is there anything more [50:20] to add there? I I think when you're [50:23] building a team and you talk about like [50:26] the team's principles or tenants or [50:28] whatever, if if you find yourself saying [50:31] things that like no one can disagree [50:33] with, either those aren't good [50:35] principles or tenants or maybe they need [50:38] to be challenged. Um and um and yeah, I [50:41] think collaboration like kind of falls [50:42] under under this uh this umbrella a lot [50:45] of the time. But yeah, I think I think [50:47] we talked about kind of my reasoning [50:49] there. [50:50] >> Okay. Uh, another contrarian take you [50:52] have is that growth typically comes from [50:54] the core of the business. [50:56] >> Yeah, I mean this one was learned like a [50:59] couple times uh both both at at at Snap [51:02] and at Discord. And this isn't to say [51:05] that like there aren't these like [51:06] wonderful growth strategies about [51:07] expansion into new markets or um things [51:11] like that. But I think the the nature of [51:13] like network effect oriented businesses [51:15] is that they they want to grow [51:17] implicitly. And so what you can do to to [51:21] to drive growth much more effectively is [51:24] focus on people who are like essentially [51:27] already using or wanting to use the [51:28] product for what it's for uh or some [51:31] kind of derivative like nearby of that [51:33] and just making it much easier, much [51:35] more performant, much more relevant, [51:36] etc. I think the way you could think [51:38] about this like we're using very like [51:40] you know eighth grade math where like if [51:42] you want to like I think almost every [51:44] consumer product maybe it's also a [51:47] controversial take like needs to be [51:49] optimizing for daily active use uh like [51:51] there are very few things in in [51:54] eventually you need to tie everything to [51:57] people's natural lives and cycles and [51:59] like days are real right weeks they can [52:02] be real but they're not necessarily we [52:03] real for like like you don't think about [52:06] there aren't a lot of things you do on a [52:07] weekly regimen regularly really really [52:10] effectively but like you wake up you [52:12] know like there there the circadian [52:14] rhythm is real very few things are done [52:15] on a monthly basis like a lot of B2B [52:17] stuff like paying bills etc. I guess we [52:20] could there's some biological things [52:21] that happen monthly. We could get into [52:22] that. But like there's there's something [52:24] about like the daily active use or even [52:26] hourly active use that I think is is is [52:28] essential for almost every consumer [52:30] product. And the idea like if you're [52:33] thinking about like a daily active user, [52:36] uh if you're going for somebody who uses [52:37] the product once a month, uh you know, [52:39] you're going to need 30 of those to [52:41] equal somebody who uses it daily. Um [52:43] that's like very very very very basic [52:45] math. And so that's why I think a lot of [52:47] the most effective consumer products [52:48] have extremely high DAU MAU ratios. Um, [52:51] and so a lot of growth can come uh from [52:55] essentially moving that ratio just a few [52:56] basis points versus uh trying to go [53:00] acquire uh a bunch of new monthly active [53:02] users and seeing if they if they retain. [53:04] And and by the way, like DA MAU is just [53:06] a form of retention. Like retention can [53:08] be any any period. It's just the next [53:11] day. Like given you use it today, what's [53:13] the probability you're going to use it [53:14] tomorrow? And so at at Snap, the lesson [53:18] there came from that disastrous redesign [53:21] in 2018, which was the only time I think [53:23] that Snap's overall DAU actually [53:25] flattened out um quarter over quarter. [53:28] And what we like laser focused on after [53:30] that was performance, particularly on [53:32] Android. Existing people who are already [53:33] using the app, just making it much much [53:35] faster and more performant for them. And [53:36] that led to almost like a growth [53:38] renaissance that, you know, went on for [53:39] multiple years. At Discord, I think it [53:41] was even more controversial because at [53:44] the time, Discord has Discord has been [53:46] something for a lot of people during the [53:48] pandemic, during crypto, during AI, etc. [53:50] So, at the time, you had Midjourney [53:52] being built on Discord and Midney was [53:53] phenomenal, just like such a cool [53:55] company product, everything. And a lot [53:57] of people on the outside were just like, [53:58] "Wow, this is a moment." Midjourney must [54:00] be huge for Discord. From a from a user [54:03] perspective or daily active use, like a [54:05] metric perspective, if I were to show [54:07] you like a you know, you can even look [54:08] at this online, right? like Discord's [54:10] use over time, you wouldn't be able to [54:12] pick out midjourney. Like the the people [54:14] the reason people use Discord is to [54:16] connect with their friends while playing [54:18] before, during, and after playing games. [54:20] And what we did in 2024 was completely [54:24] focus the team around this idea of how [54:28] can we be the best place uh for you and [54:31] your friends to play intentional [54:33] multiplayer games together. And it turns [54:35] out like even though Discord is a [54:37] technically it's a that technical team [54:38] is phenomenal and b it's it's it's like [54:41] near perfect in some ways when it comes [54:42] to like voice and things like that there [54:45] were still things we could do to make it [54:46] better before and after playing games [54:48] with your friends and when we focused on [54:52] gaming which at the time everybody was [54:54] like everybody who plays multiplayer [54:55] games especially on PC they already have [54:57] Discord right they already use it it's [54:59] lally 100% penetration right no there [55:02] were a lot of reasons that you were just [55:04] on the edge and you decided not to join [55:06] voice, you decided to use the in-game [55:08] voice, you decided to call your friend [55:09] on FaceTime instead, whatever it was. As [55:12] we made Discord better, as we focused on [55:14] performance, we were actually able to [55:16] drive some of the fastest, I think, if [55:17] not the fastest growth. Um, and I'm not [55:19] sharing anything that's not public, like [55:21] since the pandemic, obviously, Discord [55:22] exploded during the pandemic. And this [55:24] was by focusing on the thing that [55:26] everybody already said we had 100% [55:28] penetration of, which was like [55:29] intentional multiplayer gamers. Now from [55:31] there you have a you have a lot of [55:33] options right there are a lot of things [55:34] that look like multiplayer gaming crypto [55:36] looks like multiplayer gaming right [55:38] intense coordination costs you know [55:40] timesensitive etc honestly software [55:42] development looks like multiplayer [55:43] gaming like there are a lot of things [55:45] that you could build off of that core [55:47] but by focusing on people who play games [55:50] with their friends we were actually able [55:52] to like identify a lot of things that [55:54] brought somebody from using Discord you [55:56] know 20 or let's say 10 days out of the [55:58] 30 to 11 to 12 to 13. That's actually [56:02] monster growth when you think about [56:04] those pockets and moving up people's [56:05] illness like last seven out of seven [56:08] days. I think I stole that from Meta um [56:10] who once again wrote like a great [56:11] playbook on growth. And so focusing on [56:14] that illness of the core consumer base [56:17] um I think it just like ended up being [56:18] so so positive um in both the the Snap [56:22] and Discord um examples. [56:24] >> Uh this is a really important lesson for [56:27] people trying to grow a product. Uh the [56:29] TLDDR to me is there's always money in [56:31] the banana stand. [56:33] >> Like we were talking earlier about [56:34] building a a a terrorist organization. [56:37] Like that's a great example of a phrase [56:40] that if you could ascribe in a in a [56:42] strategy document you could scribe value [56:43] like there's always money in the banana [56:44] stand, right? Return to your core users [56:46] make sure it's perfect for them like [56:50] one there was always money in the banana [56:52] stand. I think at Discord it had [56:54] exceptionally positive externalities [56:56] because the most vocal critical [57:01] honestly like intelligent or like high [57:02] IQ user base in the world are are gamers [57:07] honestly like you you think you've built [57:09] for people who are are critical about [57:11] your product who are detail oriented [57:14] unless you've built for gamers you [57:15] haven't and to be able to turn around [57:18] and say we are focused like here's a bug [57:21] thread on Reddit show us what's messing [57:23] up. We don't care how small it is, how [57:25] big it is. We will fix it. It had this [57:28] like unbelievable positionality of like [57:29] we are delivering for this community. [57:31] And I honestly think this is why like [57:34] Discord is is so so great and so without [57:37] peer in this specific niche is because [57:39] it's so laser focused on like if you [57:41] play games with your friends and you [57:42] care about it like this product is going [57:44] to like there are people who sit in that [57:47] office who will make it perfect for you [57:49] uh forever. So, this begs the question, [57:52] when does it make sense to bet on [57:53] something new and try new things [57:54] because, you know, it doesn't always [57:57] make sense to just continue refining and [57:59] refining the the one piece of product [58:01] market fit that's worked. Uh, any [58:03] guidance to folks, and this is for [58:05] consumer in particular, any guidance for [58:06] when it makes sense to still make bets [58:08] outside that core? [58:09] >> Yeah, I think there there there are [58:10] there are two really like incremental [58:13] ways to um to think about it. Uh, one is [58:17] if you boil down the the the core things [58:20] about your product that are really [58:22] really great that people come to it for. [58:24] So for for Snap, maybe it's the camera. [58:26] For Discord, it's voice. And then you [58:28] look to other um niches, other cohorts [58:32] for which that could be useful. Uh, [58:36] that's like a it's a very like MBA style [58:38] like you know 2 by 2 type of thing. Um [58:41] that that's I think one method to do [58:43] that and you should honestly always be [58:44] assessing that because uh you know [58:46] that's a that's an easy one. And then I [58:48] think the second thing is like one of [58:49] the greatest gifts is when people use [58:52] your product in a way that you didn't [58:54] expect or like didn't you know like uh [58:57] you know didn't ordain because they're [59:00] going through extra hoops to do what [59:02] they want to do with the product. So [59:03] whatever you're offering is so good that [59:05] they're going through hoops to do that. [59:07] And like for like for Discord for [59:08] example, a lot of it centered around [59:10] communities and like how people were [59:11] using it to organize like larger and [59:13] larger communities as opposed to the 20 [59:14] person gaming friend group. And when [59:16] people are doing that, they're they're [59:17] essentially voting with their feet. [59:19] They're telling you like, I want to use [59:20] this product for this. What can you do [59:22] to make it a little bit easier for me? [59:23] So it kind of is a version of that, but [59:25] it's like expanding to the next thing. [59:27] And that that's more like thinking about [59:29] concentric circles. It's like, okay, we [59:31] got it at the core. How do we expand [59:33] expand? I feel like I'm doing something [59:34] with my hands here that's going to be [59:35] me'd, but uh work. [laughter] [59:38] >> We got it. Sorry. Okay, I'll stop. Um [59:41] but the uh I think then the other the [59:43] the the kind of like more the more [59:47] Snapchat like the more innovative way to [59:50] think about it is like more to look at [59:53] your company's philosophy around like [59:55] why it was built and then think about [59:57] the future totally unencumbered by the [59:58] present like no incrementality from the [60:00] present more like um future casting like [60:02] like backcasting instead of forecasting [60:04] and be like okay right now like the best [60:07] way to share a moment is to pull out [60:09] your phone and aim the camera at it and [60:11] record. Uh like that's very constrained [60:15] by the existing technology. If we were [60:17] to envision a future of like what's the [60:18] future look like to be the best way like [60:20] something just happened or my child just [60:21] took its first steps I'm at a concert I [60:24] just saw something really funny on the [60:25] street like what's the best way to share [60:26] that moment with my friends like I think [60:28] that's what leads to you to something [60:29] like oh probably glasses or you know [60:32] things like that. So this like [60:34] backcasting method I think is what [60:37] everybody wants to do. In in my [60:39] experience there are very few people [60:42] that have this type of like blank sheet [60:44] creativity to think about the future [60:46] totally unencumbered from the present. [60:48] Most people like to like even really [60:50] creative people like remix versions of [60:52] the present or like I think these days [60:55] most product managers like go on like an [60:57] Iawaska journey or travel to Japan or [60:59] something and then come back and they're [61:00] like I've solved it. But like very few [61:02] people I think are very good at this uh [61:04] version of it. [61:05] >> And and I think a big part of this is [61:07] you found product market fit, [61:08] something's working. You almost take for [61:10] granted how lucky you are that you found [61:12] something that works. And it may be easy [61:15] to think, oh, we'll find the next thing [61:16] now. And very few people do. Airbnb is [61:19] actually a really good example of this. [61:21] Like nothing beyond the original idea [61:24] has has worked in a big way. It's still [61:26] basically the same core idea. And you [61:29] know, I've worked there for a long time [61:30] and nothing has significantly impacted [61:32] the growth trajectory other than just [61:34] exactly what you just described, [61:36] continuing to make the core experience [61:38] better, more supply, more demand. [61:39] >> So, I think this is a really profound [61:41] point for people to to take in. [61:43] >> It's also like you can keep teams leaner [61:46] this way, things like that. Um the [61:48] biggest risk is it's kind of like that [61:50] um I think it's a Nasim Talib meme of [61:53] like a thousand1 days in the life of a [61:54] Thanksgiving turkey where it's just like [61:56] yeah everything's getting better and [61:57] better and better until something comes [61:58] along that's literally 10x better than [62:00] what you're doing and it takes you to [62:02] zero o overnight. [62:04] >> Um [62:05] >> and guarding against that I don't I [62:06] don't have any if I knew how to guard [62:08] against black swan events like that. Um [62:10] yeah I think [62:11] >> you'd be in finance. [laughter] [62:13] >> Yeah definitely be in finance. I [62:14] actually think like this, you know, we [62:16] were talking a little bit like SpaceX is [62:18] probably a good example actually of a [62:21] company that like guards itself against [62:23] this um in some ways. [62:25] >> Okay. Say more. I know your wife works [62:26] at SpaceX, so you have a really uh [62:28] unique insight into the into what how [62:31] SpaceX works and what has worked there. [62:33] >> Yeah. Yeah. No, no, no like uh no inside [62:36] information. I'm like a second party, [62:37] you I'm not first party, but um you [62:40] know, yeah, I I I sleep next to a SpaceX [62:42] engineer every night, so that that [62:44] helps. Um yeah, that's all you have to [62:46] do. Uh but no, I think like SpaceX is a [62:51] SpaceX is the greatest American company [62:53] right now. Greatest company in the [62:54] world. Uh like truly phenomenal on every [62:58] single level. And I'm talking more about [62:59] SpaceX like XAI and all also doing [63:02] insane stuff, but like just that. And I [63:04] think the best example what I mean by [63:06] like guarding against us the future I I [63:09] don't think any company does anything [63:11] close to what SpaceX does here and [63:13] there's a recent example literally I [63:14] think in the last week of this perfectly [63:16] where they essentially announced some [63:18] version of um they are taking the Falcon [63:21] 9 which is their primary launch system [63:24] the greatest launch system in the world [63:26] ever by far uh reusable most cost [63:29] effective kilograms to lower earth orbit [63:31] to to geostationary orbit that um every [63:35] country on the planet would would give [63:37] up their GDP to have this technology to [63:40] be able to launch and reuse rockets and [63:42] they're sunsetting the Falcon 9 and this [63:46] is like Cortez level burn the ships upon [63:49] arriving because they're essentially [63:51] saying like in order to reach our vision [63:53] in the future we need uh like a heavier [63:56] lift launch vehicle that is Starship uh [63:59] so you know starting in a few years cuz [64:01] you know they have customers and things [64:02] like that uh you'll no longer be able to [64:04] buy rides on the Falcon 9. I don't think [64:06] this is for everything, but you know, [64:07] they're gradually basically telegraphing [64:09] that like the Falcon 9 is the past. By [64:11] the way, the Falcon 9 would be the [64:13] future for literally every other country [64:14] nation state on the planet. It's the [64:16] it's the it's it's incredible machine. [64:19] And yeah, I think this is the like Elon [64:22] has built this culture that like, you [64:25] know, Clayton Christensen is dead, but [64:27] like he probably would would shed a tear [64:29] over how happy it would make him about [64:31] avoiding the innovator's dilemma where [64:33] it's just like just because we have the [64:34] thing that is the best, we're completely [64:36] willing to build the the thing that will [64:38] compete against ourselves and make the [64:40] thing that we built irrelevant. And I [64:43] wish I had more insight into what drives [64:46] that beyond maybe just Elon or or that [64:49] whole organization, but like what what [64:52] I've witnessed top to bottom is like [64:54] everybody you talk to there is like, [64:56] "Yeah, it's going to be awesome." Like, [64:58] "Yeah, we're going to do this." Um like [65:01] it's uh it's definitely part of the [65:02] culture. U maybe the cult um that that [65:05] they can do this even though on paper [65:08] and now as a public company it like kind [65:10] of looks insane. Um, and so yeah, I [65:13] think that's like actually a really [65:14] underrated part about SpaceX is their [65:16] ability to essentially burn the ships [65:17] and and focus on ne what's next. This to [65:20] me comes well back to the original [65:24] beginning of this conversation building [65:25] an or like a terrorist organization. Uh [65:28] in particular the mission like clearly [65:30] the mission of SpaceX make uh humanity [65:33] multi-planetary get to Mars and with [65:37] that mission I can totally see how they [65:40] may be working backwards from in order [65:42] to do this we need this larger rocket it [65:45] makes no sense to keep investing here. [65:47] So to me it feels very clear how they go [65:49] about making these decisions. [65:51] >> Yeah. I think there's another thing [65:52] about both of Elon's like major like [65:54] SpaceX and Tesla. I know less about like [65:56] um Neuralink Boring Company, etc. But [65:58] like you look and like on the website, I [66:01] mean now I don't know where it is, but [66:03] like they publish essentially like [66:04] 10-year plans. Like nothing SpaceX is [66:07] doing is is not something that they [66:10] didn't talk about 10 years ago. Um, like [66:12] Tesla talked about exactly what they [66:14] wanted to do with the with the initial [66:16] roadster to fund the S to fund the three [66:18] like the in some ways like the product [66:21] roadmap is is published, right? And [66:24] that's like as a product leader like I [66:27] aspire to that level of swag where it's [66:29] just like yo, we're going to do [66:30] something that's so aggressive and so [66:34] ambitious that we can just tell you [66:36] exactly what you're going to do because [66:37] if you think you can do it better than [66:39] us, like you're welcome to try. But like [66:43] this um yeah that that's like uh yeah [66:46] that that's a mission right that you can [66:47] publish and and turn into a product [66:49] roadmap say you're going to do it do [66:50] what you're going to say and then say [66:51] like yeah copy us and in the Tesla case [66:53] like oh here the patents too [laughter] [66:55] >> right I was just going to say they [66:56] released their yeah all the patents for [66:58] the Tesla electric stuff [67:00] >> yeah so next level [67:02] >> it's so interesting so one meme across [67:04] this podcast over the past couple months [67:06] has been systems thinking the other has [67:08] actually been ambition and how much that [67:11] matters now with companies and startups [67:13] because AI is making it so much easier [67:15] to build. You just tell it build this [67:17] thing, it does it. Uh, and what that [67:20] means is the level of ambition has to [67:21] keep going up. And that's like a new [67:23] habit people have to build because we're [67:25] not naturally that ambitious. We haven't [67:28] evolved to try things that seem [67:29] impossible. And it feels like we need to [67:31] learn that skill now as builders. [67:34] >> Yeah. I think there's like the status [67:36] quo is the status quo for a reason. like [67:38] the like whatever again whatever [67:41] framework you're using local versus [67:43] local optima versus you know global [67:45] optima like you know uh opportunity cost [67:48] like it's all basically just trying to [67:51] build a muscle to help you gain the [67:53] courage to just like make tomorrow not [67:55] look like today. And that's actually [67:58] just not natural for most human beings. [68:01] And there are a lot of systems in the [68:03] world uh designed to to stop you from [68:06] doing that like like the world doesn't [68:09] necessarily just like want to change. [68:11] And so yeah, it takes like an a powerful [68:14] nudge to beat that uh kind of like [68:16] static friction of of where you are uh [68:18] today. And it's not it's not just it's [68:20] not like it's not just the position, [68:22] it's the vector, right? Like like again [68:24] the Falcon 9 is is up and to the right [68:26] like this. Um, but it's like, okay, I [68:29] don't want the I don't want the actual [68:32] position. I don't even want the speed. I [68:33] don't want the velocity. I want the [68:35] acceleration. And like, in order to get [68:37] that acceleration, I need impulse, [68:38] right? I need the third derivative, I [68:40] guess. Like, I need to push it. And uh, [68:42] yeah, this is just not natural. And [68:44] people who can do that regularly and [68:46] repeatedly, but not just like [68:48] disruptively all the time, like, you [68:50] know, just always asking for more, blah, [68:51] blah, blah. I think it became very [68:53] trendy as like Steve Steve Jobs's kind [68:55] of like um ethos grew but like yeah the [68:58] types of people who can do that within [69:00] an organization especially a large [69:01] organization are pretty rare. I just had [69:03] Tara s on the podcast. She's had a [69:06] product for chat GBT work although [69:08] they're always shifting people around. I [69:09] don't know what her role is these days. [69:11] Uh and she pointed out the job of PMs [69:13] now actually more and more is to elevate [69:15] the ambition of the team to remind them [69:19] we could go bigger. What would it take [69:20] to think bigger? What would a thousandx [69:22] of this idea be? [69:24] >> Yeah. Yeah. I mean it's again like if [69:27] you can do that credibly repeatedly, [69:29] >> you know, not not just by like showing [69:30] up to meetings and be like, "What if we [69:32] did more [laughter] like [69:33] >> what if we did it faster?" [69:35] >> Yeah. Exactly. Like no, I got Yeah. I [69:36] have [69:38] got white male, gray hair, uh, you know, [69:41] fancy degrees. I can just come in any [69:43] meeting and be like, "Have you thought [69:45] about doing this quicker?" Like, okay. [69:48] Like, why do I [laughter] just Have you [69:49] thought about punching yourself in the [69:50] face? Like, yeah, maybe. [69:54] Oh man. So that's the skill is saying [69:56] that but in a way that doesn't get you [69:58] punched in the face. [69:59] >> Yeah. Yeah. To be fair, I've been [70:01] punched in the face a couple of times. [70:02] You don't you don't get a nose like this [70:04] naturally. You get you get punched in [70:05] the face a couple times. [70:06] >> Is this for real? Is that a real thing [70:07] that you got punched in the face? [70:08] >> I have gotten the [ __ ] kicked out of me [70:10] a couple of times because I run my mouth [70:11] non-stop. But uh is there is there a [70:14] story there you can share? [70:15] >> Or is that off? [70:18] >> That's that's more ancient history. We [70:20] like to move past these things, Lenny. [70:22] Come on. [70:22] >> I feel there's trauma there. We need to [70:23] get in. we need to uncover. Okay. Um, [70:27] final topic. You when we were starting [70:30] this podcast, you told me there's some [70:32] celicisms uh out there. Things that you [70:36] how do you what's the word you use? [70:37] What's the phrase? Cellisms. [70:40] >> Celicisms. Things that you often share [70:43] with folks that kind of become a thing [70:45] that people that work with you know. Uh, [70:48] are there any that we haven't covered? [70:49] things that you think are kind of [70:51] recurring pieces of advice or frameworks [70:53] or even lessons you've learned that you [70:54] think are important for people to to [70:56] hear that might be helpful. [70:57] >> I don't I don't think there's like too [70:59] too many. I mean, I think a few of them [71:00] came up. I think when we were talking [71:01] about ads, um, like one that I and I [71:05] think I I might have, uh, like I don't [71:07] know if I deserve original credit for [71:08] this, but I've I've used it quite a bit [71:11] is like this idea that, um, ads is such [71:15] a such a cutthroat like hand-to-hand [71:18] combat, very revenue goal- driven [71:21] product area that it's it's very easy to [71:23] become very short-term oriented. And one [71:26] of the things and this hearkens back to [71:28] kind of like the ideological stuff from [71:30] you know building the type of [71:30] organization you want that we really [71:32] focused on um over the years both Snap [71:34] Discord and even some of the companies [71:36] I've advised on the ad stuff is like [71:38] with ads you really want to be long-term [71:40] greedy and that was a phrase we used and [71:43] long-term greedy is a is a is a is a [71:45] funny phrase but it was actually backed [71:47] by essentially the the kind of like [71:49] mathematical realization that um which [71:52] again is not it's not rocket science but [71:53] like uh long revenue tomorrow for [71:58] advertising is essentially causally [72:01] related to the ROI you drive for [72:04] advertisers today. So you need to focus [72:07] your your focus should be on driving [72:09] advertiser ROI. The the issue is that [72:12] the the easiest way to drive ROI is [72:15] actually to to to decrease the [72:17] denominator the the I so you can you you [72:20] could lower prices. right now in modern [72:22] auctions price is an output not an [72:23] input. But the the kind of like powerful [72:26] learning there is that like the way I [72:28] will grow this business longterm, the [72:30] way I will hit my goals long term, the [72:31] way I see compounding growth because of [72:33] retention and increasing budgets um like [72:36] net revenue retention for for some of [72:38] these like SAS products is essentially [72:40] by delivering ROI for people today even [72:43] at the cost of potentially short-term [72:45] revenue. And so this type of think, this [72:48] type of phrase allowed for gave space [72:50] for a type of thinking that was [72:51] otherwise like can get pretty tough. And [72:53] people say like, "Oh yeah, it's thinking [72:55] long term." Again, I'll take it back to [72:57] like there there you're going to [72:59] encounter times in your career where [73:01] thinking long term is actually like [73:02] really really [ __ ] tough. And a lot [73:04] of this time will be when you are a [73:06] public company, like when you are [73:08] missing numbers, when like you may have [73:10] to do a layoff if certain things don't [73:12] happen. and to be able to have an [73:15] ideological setup uh for which you can [73:18] like use it as cover to think long term. [73:20] And that was that was what long-term [73:21] greedy was is like we're not a charity [73:24] like we're not giving this stuff away [73:25] because we want to make people happy. [73:27] We're giving this we're giving this [73:29] away. We're making this happen today [73:30] because we know they'll come back [73:31] tomorrow even bigger. And um yeah, that [73:34] one was was was really really really [73:36] effective. I think the other thing was [73:38] like at both Snap and Discord, we [73:40] created this concept of core product [73:41] value and this was and this is related [73:44] to my my DAO comment was like you need [73:47] to define both from a theoretical or [73:49] like a conceptual conceptual uh level as [73:53] well as a a metrics level like what it [73:56] means for somebody to come back to your [73:58] product every single day. And so like [74:00] Snap I think at the time was like Snap [74:02] is the fastest way to share a moment [74:04] with the people you care about. So it's [74:05] like fast share a moment people you care [74:08] about and you could actually like break [74:09] down what does fast mean like time to [74:10] load what does share mean it's like okay [74:12] you're opening to the camera with [74:13] sharing quickly what does uh people you [74:15] care about mean focus on best friends [74:16] like you could you could actually break [74:17] this thing down and so you had fastest [74:19] way to share moment with your best [74:20] friends like so you're if you're [74:21] building a product roadmap on the [74:22] consumer side and you're like oh should [74:24] we work on this is like well does it [74:25] help us be the fastest way to share a [74:26] moment with your best friends kind of [74:29] tenuous link to that like maybe that's [74:30] not what we should prioritize I think [74:32] Discord was like uh the best way to talk [74:34] and hang out with your friends before, [74:35] during, and after playing games. So, [74:37] it's like, you know, talk, hang out, [74:39] games before, during, after. Like, you [74:41] can break this thing down. And we [74:42] actually had metrics definitions, which [74:43] I think are like kind of proprietary [74:44] internal, but like we had metrics [74:46] definitions that we were looking for. [74:47] Those metrics definitions made their way [74:49] into the growth books into the [74:50] different, you know, into the hacks like [74:52] like, okay, does did this impact CPV? [74:55] And this whole like ideological like [74:57] language that everybody shared you could [74:59] onboard new data scientists to um I [75:01] think got just got people on the program [75:04] really really fast. These are awesome. [75:06] This idea of core product value is such [75:08] a nice way to frame it's like combines [75:12] mission with like the levers essentially [75:15] of the business. And basically the idea [75:20] here is come up with just like a phrase [75:21] that describes the mission that happens [75:24] to also connect with how we grow this [75:27] business. And it feels like three [75:28] components is a consistent pattern here. [75:31] >> Yeah. Three minute three minute. Yeah. [75:32] Maybe it's just like a rule of threes. [75:34] Yeah. Yeah. That's but the way you [75:35] characterize it is exactly right. And [75:36] like you can't overly rely on one or the [75:38] other, right? Like if it's just metrics, [75:40] it's kind of meaningless to people. If [75:41] it's just a phrase, it doesn't get [75:43] actioned day-to-day by data scientists, [75:45] product managers, and engineers. So [75:47] having both and building a story around [75:49] that um and telling that story [75:50] repeatedly like for for for core product [75:53] value at Snap um we actually had an [75:56] onboarding presentation which I [75:57] delivered to every single new hire for [75:59] seven years straight every week or every [76:00] other week later uh that was about this [76:03] because like that was part of the cult [76:05] right like this is how we deliver core [76:07] product value. This is why people choose [76:08] Snapchat every single day in the face of [76:11] tons of alternatives and a bunch of [76:12] savage competitors. Um, and I think the [76:15] same same goes for Discord. [76:16] >> And the the work is coming up with that [76:18] phrase that actually is clear, simple, [76:20] connects to all these things. Not not [76:22] easy, I imagine. [76:22] >> And the and the metrics that are like [76:24] causally demonstrated with with [76:25] retention. [laughter] [76:26] >> Those that's tough, too. That I think I [76:28] think no data scientist likes working [76:30] with me because I'm like, "Hey guys, we [76:31] got to figure this out." [76:32] >> Interesting. Okay. One other take that I [76:34] uh you mentioned when we were chatting [76:36] ahead of this is this idea of how a [76:37] great PM is an oxymoron. [76:39] >> Yeah. I can't I can't help but think [76:41] like this is maybe why it's so hard to [76:43] like come by some of them because like [76:46] you see an overexpre like there's [76:47] basically these three and yeah I would [76:50] love I bet you could actually come up [76:52] with this from like a synthesis of so [76:54] many of the amazing product leaders [76:55] you've had on here but there are these [76:57] three kind of like oxymorons I think [76:59] like there there are two wolves inside [77:01] of the PM three times uh one is like I [77:05] think to be a really good really good PM [77:08] you kind of have to be like really [77:10] confident and that borders on arrogance [77:15] a lot of the time. Um especially at [77:18] certain levels. And so that has to be [77:20] balanced with like some form of humility [77:23] like really serious humility where you [77:25] kind of again are so confident that [77:29] what we're doing is the right thing to [77:31] do and that we have conviction on this [77:32] path and also personally that I will be [77:35] recognized for my contribution. [77:38] I can go out there and give credit to [77:40] exactly who deserves it. The the [77:42] designer who made that last second [77:44] change that made a huge difference. The [77:45] engineer that stayed up all night to to [77:47] get out the the thing on time for the [77:49] partner event. Like I think that this [77:51] also ties to being a great leader. But [77:52] like having the humility and the [77:55] attention to detail to know who to give [77:57] credit to actually comes from this sense [78:00] of confidence that like you kind of know [78:02] that PMs are going to get too much [78:03] credit no matter what. And so like I [78:05] think that's the first one. The second [78:07] one which I like really like and really [78:10] try and test for in interviews and I'm [78:12] not I'm not the best interviewer um is [78:14] like you want these people who are like [78:17] extremely organized and very like [78:20] process driven like they understand the [78:23] value of good process and not too much [78:25] not too little etc. and then are just [78:27] like super comfortable with ambiguity [78:31] almost to the point of like despising [78:34] authority like like like almost like [78:36] pathological demand avoidance. Uh and so [78:39] it's just like I have a checklist for [78:41] everything, but I'm also comfortable if [78:42] I have to throw it out and redo it all [78:44] the time or something like that. And I [78:45] think like these two wolves are like if [78:48] you can demonstrate that through some [78:49] stories or or through some um kind of [78:52] like examples that's like so attractive [78:54] to me uh in a candidate. It's like, [78:56] yeah, I'm I'm really comfortable with [78:57] ambiguity, but that doesn't mean I just [78:59] like am not organized or like I I hate [79:01] process. Another one of those things [79:02] where like every it's really trendy to [79:04] just say like, oh, process is useless, [79:05] but it's like I I like somebody who has [79:07] both those walls. Um, and then I think [79:10] this the last one is a little more [79:12] obvious, but like you really a lot of [79:15] the great PMs that I've worked with [79:17] really just always have some form of [79:19] long-term thinking in their mind. like [79:22] they just they just know how they want [79:25] the world to look 6 months, a year, two [79:28] years from now. And they're patient and [79:31] they they they know what building blocks [79:33] need to come into place for the system [79:34] to work. But yet on a day-to-day basis, [79:37] they're their bias for action is just [79:40] like um borderline annoying, you know? [79:44] Uh you know, they're just like they're [79:46] just such good people at setting the [79:48] pace of the entire company. like you [79:50] know when they wake up in the morning [79:52] like the company wakes up uh the group [79:54] wakes up the team wakes up whatever it [79:56] is and um yeah I think that combination [79:58] of like urgency and and bias for action [80:01] with long-term thinking um yeah those [80:03] are my three my three oxymorons the [80:05] three three pairs of wolves inside every [80:07] great PM [80:08] >> amazing description of what makes a [80:10] great PM such a cool way to think about [80:12] it uh Nquille Singal was on the podcast [80:14] a couple times and he talks about the [80:17] it's like a different way of thinking [80:18] about it but he talks about the shadow. [80:20] You have a superpower, but then there's [80:21] the shadow. Know what the downside of [80:22] that is. It's a different thing, but but [80:24] I love this. It's just like a really [80:26] interesting way to just think about what [80:27] is a great PM. Okay, one one final uh [80:31] thing before we get to a very exciting [80:32] lightning round. I want to take us to [80:34] Contrarian Corner, which is a recurring [80:37] podcast uh segment where uh I feel like [80:41] you're going to have some good stuff [80:42] here. Uh the question is just what's [80:44] something you believe that most other [80:45] people don't? [80:46] >> Being contrarian is tough these days. [80:47] very trendy to be contrarian, which is [80:49] ironic. Uh, so maybe I should just be [80:51] conformist. Uh, the Can I Can I break [80:54] this into like a like a light spicy, [80:56] medium spicy, heavy spicy? Love that. [80:59] Let's do it. [81:00] >> Okay. Okay. Uh, so my light spicy, I [81:03] don't think everyone would disagree with [81:04] this, but some I think as a product [81:07] manager in particular, I I don't think [81:10] we have time to go into everything, but [81:11] like I think the most important career [81:14] decision you make is is who your partner [81:17] is, like who you marry, uh, who you [81:19] spend your time with personally. Um, I [81:22] think this like has both growth and just [81:25] like knock-on effects that are really [81:27] hard to quantify. And um yeah, you'd be [81:31] there's no there's no amount of Lenny [81:33] podcasts you can listen to and [81:35] internalize that will match um you know, [81:38] finding a great partner uh who will like [81:40] help you support you and and help you [81:42] grow. Um and be like kind of like [81:44] intellectually a sounding board for you. [81:47] Um so I think that that's probably my [81:50] light take because I think a lot of [81:51] people would kind of like agree to that [81:54] in some degree. I know my I know my [81:57] super spicy one. So that one I don't [81:58] think of on the medium spicy like [82:03] this is kind of related to the my my [82:04] museum curator comment. Um I think a lot [82:07] of people who I look up to in terms of [82:12] taste and like I maybe it's just being [82:15] in Los Angeles but like I care a lot [82:17] about like stuff like art like art [82:20] behind me stuff like that. um is I've [82:23] noticed that that uh you know like the [82:28] difference especially in wealthy people [82:30] between those who have like taste and [82:31] those who are seen as like um gouch or [82:34] something like that is uh the ones with [82:36] taste like kind of like stop one step [82:38] before everybody else does. um like like [82:43] you know maybe they're collecting like [82:45] art in a certain trend uh but then like [82:47] they they realize like when things are [82:50] whatever the the old phrase is jumping [82:51] the shark or like when things are like [82:53] hitting a certain level where it's like [82:54] okay now like now is the time to kind of [82:56] like stop and look for something new and [82:59] I think that like what you haven't [83:02] shipped what you've built and tried and [83:04] held in your hands like and and loved [83:06] but haven't brought to the world I think [83:09] is like kind of a like a test for that [83:11] in in some ways. Um, and so yeah, I [83:15] don't know if that's Yeah, that that's [83:16] like a super contrarian, but I think you [83:19] can measure a lot is like people with [83:21] great taste seem just like know when to [83:23] take something and get passionate about [83:24] something and then kind of like [83:26] >> when it gets cringe. [83:27] >> Yeah. Yeah. Let it Yeah. Maybe cringe is [83:29] it? Yeah. Let it go. My spiciest one [83:34] and like this is like this we'll bring [83:37] it back here to the start of the [83:38] podcast. like why why am I doing this [83:40] podcast now? Like you know you and I [83:42] have known each other, been in group [83:43] chats, chatted all the time for a long [83:45] time. This one's probably a take where I [83:47] don't even know if I'd be super [83:48] comfortable saying it depending on my [83:50] employer. Um I think that [83:55] to truly create [83:58] great consumer products for certain [84:02] cohorts of of humans [84:05] of of users. [84:07] um [84:09] the product managers [84:12] and designers that you put on some of [84:14] those products. Um there are certain [84:19] uh inalienable traits about those [84:22] product managers that cannot be learned [84:24] uh cannot be acquired that are [84:28] that are essential to creating truly [84:30] great consumer products. And by that I [84:31] mean literally gender, um, race, uh, [84:36] ethnicity, [84:38] uh, socioeconomic background. Um, and [84:42] what I mean by this is not like, oh, if [84:44] you want to create a product for, um, [84:47] you know, young black men, you need to [84:49] hire young black men. Like, I don't mean [84:51] stuff like that. I just mean um and I [84:55] don't think I don't know how comfortable [84:55] I am giving like examples but I do think [84:59] that we are not comfortable as a society [85:00] admitting this but there are certain [85:02] things about um how people are brought [85:04] up uh who they are biologically uh [85:07] things like that that uh suit them to to [85:09] better working on certain consumer [85:11] products because at their core like most [85:15] consumer products are and and Nikita [85:17] talks about this too are like hitting [85:18] against certain like parts of Maslaw's [85:21] hierarchy of needs and um in order to [85:25] like understand those needs really well. [85:27] I think there are certain people with [85:29] with you know backgrounds, upbringings, [85:31] etc. that make them better at that. Um [85:35] that doesn't mean you can hire for it. [85:37] Uh that doesn't mean you you can judge a [85:39] book by its cover, things like that. And [85:41] that that's why it's such a spicy take [85:42] is like it's not necessarily even [85:44] actionable in in many situations. But um [85:49] yeah, maybe I'll leave it at that before [85:50] I give myself. [85:51] >> Yeah, like the the real spicy version [85:52] would be like that grid of here's the [85:55] here's who you need for this type of [85:57] product, but we're not going to go [85:58] there. Uh before we get to our very [86:02] exciting lightning round, is there [86:05] anything else that we haven't touched [86:07] on? Anything else you wanted to share? [86:08] Anything you want to double down on [86:10] before we get to exciting lightning [86:13] round? We covered a lot of ground, so [86:14] there might not [86:15] >> No, I think I think we Yeah, we covered [86:17] a ton of ground. Um, it's it's really [86:20] great. [86:21] >> Well, Peter, with that, we reached our [86:23] very exciting lightning round. I've got [86:24] five questions for you. Are you ready? [86:27] >> All right. Um, I will I will admit to [86:30] our our viewers and listeners that I [86:32] have not prepared for these, which I [86:34] guess was the point of the lightning [86:36] round. [86:36] >> That is the point except they're the [86:38] same questions to everybody except the [86:39] last one. [86:40] >> Okay, cool. So there's no uh it's not [86:42] too much of a surprise. [86:43] >> Okay. Then then I love these questions. [86:45] All right. [86:45] >> Okay. And I love it when people aren't [86:47] prepared. Uh okay. Two or three books [86:48] that you find yourself recommending most [86:50] to other people. [86:50] >> I'm going to butcher the title a bit. Uh [86:52] how to get filthy rich in rising Asia [86:57] and [86:59] when genius failed. [87:02] >> Oh, haven't heard of these? Uh amazing. [87:05] >> And then the third one uh so how to get [87:08] filthy risk rising Asia. When Jesus [87:09] failed is about the fall and the rise [87:11] and fall of long-term capital management [87:12] in late 1998 caused like one of the [87:14] crashes or nearly averted crashes. And [87:16] then my third one which I actually give [87:18] to a lot of consumer PMs is um [87:20] Einstein's Dreams by Alan Lightiteman [87:23] the physicist from MIT. [87:25] >> Very uh unique pick. I haven't heard [87:26] these on the podcast before. Favorite [87:29] recent movie or TV show you've really [87:30] enjoyed. [87:31] >> So I do not consume uh TV uh and I [87:36] rarely I have I haven't watched as many [87:38] movies I used to. I used to be a big [87:39] movie buff. So, if I could flip this um [87:42] the to something that was closer to my [87:44] heart, gaming, I think if you haven't if [87:48] you if you don't play games, video [87:50] games, um you should uh give last year's [87:55] game of the year, um Clare Obscure [87:58] Expedition 33, uh a shot. um just a [88:02] wonderful indie studio out of France [88:04] that constructed a story with music and [88:07] and sound that is just like um like like [88:11] Louve level like beauty and really [88:15] reminds you that games are a form of art [88:18] that maybe we still as a society haven't [88:20] yet like realized how to appreciate. Um [88:23] and I I think the people behind it are [88:25] phenomenal and I think what they produce [88:26] is amazing. [88:27] >> Wow. Uh, I want to watch a Twitch stream [88:29] of you playing it. How fun would that [88:31] be? That could be our follow-up. [88:33] [laughter] [88:34] Is that um is that on desktop or is that [88:36] like a console thing? [88:37] >> Uh, it's it's available on console as [88:39] well, but I think best probably best [88:40] experience on desktop. I think it works [88:42] on Mac and and PC. [88:43] >> So cool. I think that's the first game [88:44] recommended. I like how uh different [88:46] this round is than most. Favorite [88:50] interesting AI product right now. [88:51] >> In general, a lot of the AI products [88:53] still feel like we're throwing spaghetti [88:54] at the wall. Um, I'll take a more and [88:57] I've heard a couple people actually [88:58] mention reference this on the pod. The [89:00] one that I found that changed my [89:02] workflow the most in at work [89:07] um was Hex. Um, it was the first time [89:11] that I just like marketkedly found [89:14] myself doing something differently [89:15] immediately. And it was it was one of [89:17] the only multiplayer AI experiences that [89:20] that worked. like literally me and and [89:22] the co-founder of of Discord, Stan, like [89:24] in a meeting in a hex thread interacting [89:26] with Hex as we were having discussions [89:28] in the meeting and getting data back and [89:31] things like that. Obviously like latency [89:32] and inference time and stuff like that [89:34] is still an issue, but um I find that [89:37] product to to be like um really really [89:40] great um and really changed like really [89:42] just enabled me to feel like a super [89:44] super human. [89:44] >> I love how every one of these answers I [89:46] don't think anyone's ever said. Two [89:48] final questions. Uh, do you have a [89:49] favorite life motto that you often find [89:52] useful in work or in life? [89:54] >> There's this quote that I'm going to [89:55] butcher because I don't have it in front [89:56] of me. Um, that [90:00] that is from John Steinback in um, East [90:02] of Eden, I think, about the passage of [90:05] time. And I'm just like generally [90:07] obsessed with the passage of time [90:08] because the people in my life, whether [90:10] it's been like personal or or [90:12] professional, who I've just always been [90:14] literally like infatuated by, like like [90:16] I've just wanted to spend more time. [90:17] Like I've I've like fallen in love with [90:18] these people. They oftenimes like seem [90:21] to float through time and perceive time [90:23] very differently than than normal [90:26] people. And this quote from Steinbeck um [90:32] uh it's uh it's about like how time [90:35] passes differently when you experience [90:38] more or less. And essentially uh the the [90:41] quote is that like duration has no [90:45] signposts uh sorry time has no signposts [90:49] upon which to drape duration. Um from [90:52] nothing to nothing is no time at all. [90:55] Uh, and I really I really want to get [90:57] the full full quote for you. Uh, but um [91:00] because it's like literally one of those [91:01] things where you're like, "How could a [91:02] human write this? It's so beautiful." [91:05] Um, but yeah, from nothing to nothing is [91:07] no time at all. And so like a time a [91:09] life that is even scarred with tragedy [91:11] or like essentially just do it, right? [91:14] Like just just get out there, meet new [91:16] people, experience things because um the [91:18] best way to to slow the passage of time [91:21] uh is essentially to to do something [91:22] with it. Um, and so I'm like I'm pretty [91:25] generous I think with my time uh and [91:28] like you know how I spend it with people [91:30] and just like trying to to do things [91:32] that I want to do as soon as possible [91:34] because um yeah, you just you just never [91:37] know. And um yeah, those time periods [91:39] that really seem to pass slowly in a [91:41] good way are the ones where um you've [91:43] just experienced a lot. And so yeah, [91:45] this idea that from like from nothing to [91:47] nothing is no time at all uh like is is [91:50] so scary and motivating to me. That is [91:53] beautiful. Again, uh a quote nobody has [91:56] shared before. Final question. Okay. So, [92:00] I was looking at your LinkedIn, looking [92:01] at your career history. Uh funny enough, [92:03] you were head of product immediately. [92:04] That was your first product role at a [92:06] company called [92:06] >> I was a product manager for a little [92:08] bit, but [92:08] >> Okay. Okay. You've overlooked that in [92:10] your LinkedIn. Uh so, it was a company [92:13] called Ustream, and you described your [92:15] role as product leader for the Sheba Enu [92:17] puppy cam. We need to hear more. [92:20] >> Yeah. Yeah. Ustustream. Ustream is a [92:22] funny funny company. It was like the [92:24] first live streaming uh it was one of [92:26] the first live streaming companies [92:27] alongside live stream the company [92:29] actually called live stream and Justin [92:30] TV with Justin Khan and and him like [92:33] kind of a golden age of of live [92:34] streaming that was first enabled by [92:36] essentially like higher bandwidth and [92:37] things like that. Um sometimes we talk [92:40] about failures like Ustream was [92:42] essentially a failure because we we we [92:44] should have focused on gaming. Justin TV [92:45] focused on gaming became Twitch etc. But [92:48] um one thing about Ustream uh or I'll [92:51] tell you really quick I know it's [92:52] lightning round uh the number one [92:54] product adviser to Ustream this is 2009 [92:56] was Josh Elman uh which like that was [93:00] like my come up in in product was 17 [93:02] years ago like having a mentor who was [93:04] Josh Elman uh which was really really [93:06] phenomenal to to learn product and then [93:08] the actual first product manager at [93:10] Ustream uh was this guy Matt Schlick uh [93:13] who had dropped out of high school I [93:15] think to work on the product uh 2008 [93:18] 2009. Uh Matt just sold uh Moltbook to [93:23] uh to Meta uh the kind of the social [93:25] network for for agents. Uh and so he is [93:28] like one of the OG uh social hackers. Uh [93:31] so it was really really awesome to see [93:33] him realize that. But um yeah uh Ustream [93:38] became really popular uh from the Sheba [93:41] Inu puppy cam which was somebody taking [93:43] a Logitech camera and aiming it at a [93:45] litter of new Sheba Inu puppies and 247 [93:49] just streaming these puppies growing up [93:50] acting crazy etc. And we we should have [93:54] learned there were there actually three [93:55] 247 cams at the time. I think there was [93:57] um there was uh the Sheba puppy cam. [94:00] Then there was the Eagle cam that showed [94:02] the birth of baby eagles in Dora, Iowa. [94:05] And then there was uh Jason Calcanis uh [94:08] on uh just like 24/7 broadcast. He was [94:10] the only person who could afford like a [94:12] broadcasting his this weekend startups. [94:14] >> So two beautiful things and one hideous [94:16] thing. No, I'm just kidding. Uh but uh [94:18] yeah, the Shiba Union puppet cam I I [94:20] there was a lot to learn about that. H [94:22] people use this as their background. Uh [94:24] just to have it in the background while [94:25] they worked on their on their headphones [94:27] just listening to these puppies was very [94:29] soothing. And we got so many emails and [94:31] people being like this saved my life [94:33] like blah blah blah. So um [94:35] >> yeah. Yeah, it was uh it was really fun. [94:37] Live streaming is still kind of like an [94:38] uncracked product category. Um still [94:41] technically challenging, quite [94:42] expensive, etc. But I I learned a lot [94:44] from like a consumer standpoint from [94:46] that. [94:46] >> Incredible. There's also a on your [94:49] LinkedIn you worked at a company that [94:50] incubated Dollar Shave Club and Liquid [94:52] Death. We're not going to get into that. [94:54] Uh Peter, you mentioned briefly a second [94:56] ago how generous you are with your time. [94:59] Uh I really appreciate you making time [95:01] to be here. Uh we got through so much. [95:03] Uh and this is one of the longer [95:05] episodes I've done, which says a lot [95:07] about just how much there is to talk [95:08] about and how much you have to share. [95:09] So, thank you for being here. [95:10] >> Yeah, I I um it's so great to to talk [95:13] about some of this stuff. And um you [95:15] know I haven't uh expressed it uh [95:18] externally in a lot of ways and um [95:21] really I think it says a lot about like [95:23] you and your your uh your podcast and [95:26] your your your readership and [95:27] viewership. Um it's I think you like [95:30] probably have done more than any person [95:31] to really shape uh what this career [95:34] looks like and how it's improved over [95:36] over the years. Um so I'm really [95:38] grateful for the time. I'm grateful for [95:39] everybody who listened. [95:41] >> Uh final question. How can listeners be [95:42] useful to you? I I think like people if [95:46] they listen through this podcast, they [95:48] probably got an idea of like not [95:49] necessarily like the contrarian takes, [95:51] but just some of the aspects of like how [95:52] I think about um product that are maybe [95:55] just like a little bit different than [95:57] the typical product leader that came up [95:59] in Silicon Valley or something like [96:00] that. And um yeah, if they feel if they [96:02] feel like something resonates to them or [96:04] that they disagree with something and [96:06] want to correct it or or they'd have [96:08] beef with it or they just like something [96:10] fascinated them about it, um I would [96:12] love to hear about it. I'm not uh yeah, [96:15] I would I would just love to hear um I I [96:17] just love I love hearing when I love [96:18] disagreeing with stuff. I love hearing [96:19] about stuff like that. [96:20] >> Sweet. And we'll point your to your [96:22] socials if folks want to disagree with [96:23] you. Peter, thank you so much for being [96:25] here. [96:26] >> Thank you for having me, Lenny. Really [96:27] appreciate it. Bye everyone. Thank you [96:30] so much for listening. If you found this [96:32] valuable, you can subscribe to the show [96:33] on Apple Podcasts, Spotify, or your [96:36] favorite podcast app. Also, please [96:38] consider giving us a rating or leaving a [96:40] review as that really helps other [96:42] listeners find the podcast. You can find [96:44] all past episodes or learn more about [96:46] the show at lennispodcast.com. [96:49] See you in the next episode.