Notion Drops Resumes and CMO Role in Shift to AI Talent Model

Notion转向AI人才模式：弃用简历并取消CMO职位 CEO Ivan Zhao 认为，人工智能使传统企业架构过时，要求企业实行去中心化运营，并依据主观品味而非从业经验进行招聘。 Ivan Zhao 阐释了他将公司视为灵活的爵士乐队而非刻板的军乐队来管理的理念，并详细介绍了 Notion 如何面向 AI 时代调整其架构与人才模式。他概述了公司两次截然不同的“再创立”历程——第一次是在京都寻找产品与市场的契合点，第二次是在坎昆转向生成式 AI。Zhao 还谈及拆除诸如 CMO 等传统高管架构、拥抱杠铃型人才模式、纠正早期市场拓展失误，以及他为何将企业文化视作一种共同的信仰体系。归根结底，这场讨论凸显了一种正在转变的企业运作模式：AI 充当着结构钢的角色，赋能更精简的团队以极高的自主性与灵活协同，在战时速度下高效运转。 Ivan Zhao 提出以“爵士模式”（Jazz Mode）重塑组织文化，主张用即兴协作、高自主性与个体卓越取代僵化的“行军乐队”式科层管理，公司内部目前聚集了 50 至 60 位前创业者以践行该理念。 (0:00) Notion 曾通过两次破釜沉舟的“二次创业”（Refounding）完成蜕变：2015 年裁员后迁往日本京都重构产品以寻找 PMF，2023 年在坎昆体验早期 GPT-4 后彻底转向 AI 原生路线。 (0:44) 赵认为基于大语言模型构建软件更像“酿啤酒”而非“造桥梁”，充满不可预测性；为此 Notion 采取“杠铃型”人才架构，跳过中级工程师，由极具品味的资深架构师直接带领高能动的初级工程师与代码 Agent。 (9:59) 为消除传统汇报带来的沟通瓶颈并紧跟高速的产品迭代，Notion 彻底拆除了传统的 CMO 组织，将市场团队解构为紧贴产品端的“叙事团队”与主打线索转化的“GTM 销售支持团队”。 (13:18)

Notion co-founder and CEO Ivan Zhao argues that the emergence of large language models requires software companies to abandon traditional SaaS management playbooks in favor of a flexible, experimental operating model built around taste, agency, and ongoing "refounding." Refounding Notion From Kyoto to Cancun (0:44) Zhao characterizes major company resets as physical necessities rather than routine strategic pivots. In 2015, lacking product-market fit and running out of seed capital, the founders laid off their entire staff and relocated from San Francisco to Kyoto, Japan, to rebuild the software from scratch. That routine of intensive focus—rooted in an admiration for traditional craft and early computing pioneers like Douglas Engelbart and Alan Kay—allowed them to reach profitability with fewer than ten employees before hyper-growth hit during the COVID-19 pandemic. Beer Brewing Over Bridge Building (5:03) While traditional software engineering behaves like bridge building—where specifications can be handed predictably from product managers to designers and engineers—Zhao compares developing with language models to brewing beer, an organic process where outcomes cannot be strictly dictated to the underlying technology. Because raw capability in writing and coding is increasingly commoditized by AI, Notion shifted its internal talent formula to prioritize innate taste and individual agency over accumulated experience. Dismantling Silos and the Barbell Org (9:59) To prevent communication bottlenecks, Notion eliminated its traditional Chief Marketing Officer organization, splitting responsibilities between a product-adjacent storytelling team and a direct go-to-market demand team. The company adopted a "barbell" hiring distribution that skips mid-level roles to pair senior architectural talent with high-agency junior individual contributors who manage automated coding workflows. The Pragmatics of Wartime Operations (17:09) Operating in what Zhao and the host describe as "wartime," Notion relies on roughly 50 to 60 former startup founders who can execute decentralized projects without territorial friction. Traditional multi-month product roadmaps have been abandoned for week-by-week improvisation, though conservative financial planning remains intact. Zhao also notes that because corporate knowledge work primarily consists of simple information retrieval rather than high-end reasoning, Notion can utilize cheaper open-weight or second-tier models to protect gross margins against high AI compute costs. Enterprise Realism and Organizational Invariables (53:17) Despite championing unconventional structures, Zhao acknowledges limits to reinventing business practices from first principles. He considers his multi-year resistance to a conventional enterprise sales force a mistake, noting that buyers inherently demand human trust. Notion ultimately solved its go-to-market execution by pairing rigorous systems architecture with an aggressive outbound sales leader. Looking forward, Zhao expects companies to remain hierarchical due to legal accountability and human nature, with AI serving as structural coordination rather than replacing human leadership entirely. While Zhao and the host contend that circular, AI-centered organizations moving at rapid cycle speeds will outcompete rigid corporate hierarchies, the model relies on cultivating subjective "taste" and sustaining an intense, wartime cadence that may prove difficult for conventional enterprises to replicate.

Notion联合创始人兼首席执行官Ivan Zhao指出，传统SaaS管理模式在人工智能时代已显僵化，企业需以高自主性的“爵士乐队”模式重塑组织架构、人才标准与产品开发流程。 二次重塑：从京都绝境求生到全面转型AI (0:44) Notion的成长历程经历了两次关键的“二次重塑”（Refounding）。第一次发生在2015年前后，当时公司尚未找到产品市场契合点（PMF），且资金即将见底。Zhao与联合创始人Simon Last裁撤了全部员工，离开旧金山迁往日本京都，依靠转租旧金山办公室和公寓实现现金流转正，在极度聚焦的环境下重写产品。这次重写耗时一年半，最终在疫情初期迎来用户爆发。第二次重塑则发生在2023年墨西哥坎昆的公司活动期间，Zhao在体验OpenAI的GPT-4后产生强烈信念，随即推动公司全面转向以AI为先，尽管内部初期存在质疑，且2022年底起尝试的AI Agent产品曾因技术不成熟经历了一年半的失败探索。 摆脱传统管理：从搭积木工程走向“酿造啤酒” (5:03) 在经历疫情期的高速扩张时，Zhao曾在投资机构的传统理念压力下引入职业化管理体系，但后来发现这稀释了公司的原生文化。Zhao将传统软件工程比作“建造桥梁”，即需求由产品经理流转给设计师，再由工程师按部就班实现；而开发大语言模型产品则如同“酿造啤酒”，底层模型的行为难以精准预测，团队无法指挥酵母朝特定风味演进，必须依靠优秀人员直接在模型前持续调试。这一变化促使Notion打破了岗位边界，传统部门壁垒瓦解，转而依赖能写代码的设计师、具产品思维的工程师以及高频调用API的产品经理。 爵士模式与组织扁平化：AI充当结构性钢材 (0:00) 针对硅谷关于管理模式的争论，Notion提出了“做爵士乐队而非军乐队”的运营理念（被节目称为Jazz Mode），强调高能动性、即兴应变与个体技艺的协同，而非僵化的科层制与重复分工。为加速信息流动，公司彻底取消了首席营销官（CMO）层级，将营销拆分为紧邻产品线的叙事团队与服务销售线索的获客团队。在约千名员工中，Notion引入了50至60名前创业者来主导去中心化项目。Zhao认为，传统知识工作中充斥着低效的信息搬运，而大语言模型与软件的结合就像现代建筑中的结构钢，能够支撑企业在保持扁平组织的同时向上规模化，降低沟通损耗。 哑铃型人才模型与以品味为核心的用人标准 (9:59) 随着大语言模型将编程与写作等基础技能平民化，Notion更新了其人才评估公式（能力与经验相乘，再乘上品味与价值观，以及意志力与能动性）。由于底层技能已被工具拉平，公司将难以改变的“品味”与工作韧性视作核心区分度。为此，Notion在招聘上采用了“哑铃型”结构，重点吸纳资深架构师与刚毕业的高能动性新人，跳过了中层工程师；由资深专家把控架构品味，指导初级工程师，再由初级工程师调度多个代码智能体（Coding Agents）。此外，面试不再依赖传统简历，候选人需直接通过Notion提交实际构建的作品。 知识工作的成本曲线与销售落地的试错 (23:02) 面对大语言模型高昂的算力成本对传统SaaS高毛利（如86%）带来的冲击，Zhao认为公司处于必须接受利润率压力的“战时状态”。但他指出，与依赖昂贵前沿模型的代码生成类产品不同，企业内部大量的日常知识工作（如规章查询与信息搬运）并不需要顶尖模型，未来可通过更廉价的开源权重模型或中国模型来降低推理成本，维护长期毛利。在商业化方面，Zhao承认公司早期曾犯下试图以“第一性原理”重新发明销售体系的错误，延误了约两年的企业级销售推进；直到引入GitHub前首席营收官与经验丰富的销售主管形成体系与执行的互补，销售业务才走上正轨。 尽管行业内存在打造全自动化企业的激进构想，Ivan Zhao基于人类科层天性与法律责任体系，预测未来的组织形态仍将保留人类层级，但核心决策机制将深度交织AI。对于增长陷入停滞的成长期初创企业，管理者面临的核心抉择在于是否有勇气打破现有的局部舒适，承受短期业绩代价以执行彻底的二次重塑。

[0:00] Ivan Zhao: We want to be a jazz band, not a marching band. [5:03] Ivan Zhao: building classic software is like engineering a bridge—if you can design it, you can predictably build it by passing requirements from product managers to designers, and then to engineers. Building with language models, however, is like brewing beer. [5:03] Ivan Zhao: You can't truly predict the underlying thing... You cannot tell the yeast, 'Hey guys, go go go, go towards that flavor profile more.' You can't do that, right? You sort of have to throw your best people in there, see what the model, see what technology gives you. [9:59] Ivan Zhao: Talent = (Capability * Experience) * Taste/Value System * Agency/Will [9:59] Ivan Zhao: It's to each of their own. There's no right answers. [18:21] Ivan Zhao: Oh, wartime is way more fun. [21:47] Ivan Zhao: there's no freaking plan. [24:28] Ivan Zhao: sometimes your body just tells you you have to do it [28:05] Ivan Zhao: the story we tell ourselves [32:50] Ivan Zhao: the world is stopping. [36:45] Ivan Zhao: little decalcified meathead machinery just trying to break things [36:45] Ivan Zhao: a one and a two and then 50 tied for third. [42:09] Ivan Zhao: It's invented. It's not as old as fire or language. [45:35] Ivan Zhao: to lead the group of human... people want to hear from you. [50:49] Ivan Zhao: might as well just be an artist. [59:04] The host: Be yourself, everyone else is taken.

[0:00] We start a mantra in the company. We [0:02] want to be a jazz band, not a marching [0:03] band. I love that. [0:04] >> It's like a self-reflecting value thing. [0:06] It's auto equilibrium with me. So, it's [0:08] like, okay, I'm a jazz band person. So, [0:10] we've been hiring more and more jazz [0:11] band person. Also, AI happened in the [0:14] past two two and a half years. So, more [0:16] jazzband people can really shine during [0:19] this moment. [0:19] >> Okay. So, we had manager mode, we have [0:21] founder mode. There's this new mode. [0:23] What should we call it? [0:24] >> Jasmo, maybe. [0:25] >> That's not bad actually. [0:44] Hey everybody. Today we have Ivan Jawan, [0:47] the co-founder and CEO of Notion. I like [0:50] to think of him as the refounder. Uh he [0:53] started the company a long time ago and [0:55] in 2015 he didn't have part of market [0:57] fit. He and his co-founder moved to [0:59] Kyoto of all places and kind of [1:01] restarted the company and rebuilt it. [1:04] After that, they got product market fit [1:05] and it really took off. He did the same [1:07] thing in 2023 [1:10] around generative AI in an offsite down [1:12] in Cancun and literally restarted the [1:15] company and in my mind he's the [1:17] canonical example of how a SAS company [1:20] can move and become an AI company. So [1:23] that's super interesting. The second [1:25] thing that's interesting is just hearing [1:26] how he builds the company. He's very AI [1:29] native in the way he organizes the [1:31] company, the way he compensates people, [1:33] the way he builds today. And it rhymes a [1:36] lot with some other CEOs that have [1:38] interviewed like Jack Dorsey a couple [1:39] weeks ago. I hope you like it. What do [1:42] you think of my outfit, by the way? [1:44] >> I think your outfit it's uh your pants [1:46] could be looser. [1:47] >> Looser. Okay. On the bottom or just in [1:50] general? [1:50] >> In general. Uh I like your sneakers. I [1:52] don't know what they are. Okay. They're [1:54] Italian. Premiada. [1:56] >> That's cool. [1:56] >> You're known as a very fashionable [1:58] fellow. Where where you shop these days [1:59] here in New York? [2:00] >> New York has a good store near here in [2:02] like Soho Traca called Lakasan. Um not [2:06] like a designer designer, but more like [2:08] a European, [2:11] easy to wear. Um [2:13] >> I like the vibe. [2:14] >> Thank you. [2:15] >> Um I got up my game. Ivan, [2:17] >> just channel some nice. [2:18] >> Let's go shopping. You want to go [2:19] shopping? [2:20] >> It's right around here. Afterwards, we [2:22] can go. [2:22] >> Okay. Hey, I want to talk about CEO [2:24] stuff, not just fashion. I've been doing [2:26] this sequoia thing for a couple years [2:28] and a lot has changed in CEO. Like the [2:32] whole manager mode, founder mode thing [2:34] happened and now it's founder mode and [2:35] some new mode seems to be happening. [2:37] Let's just reflect back when Paul [2:39] Graham's founder mode memo came out. My [2:41] my person [2:42] >> When was that? Since 2022, right? 2021. [2:45] >> No, no, no. It's like it was later than [2:47] that. It's like 2024. [2:49] My reaction, by the way, was relief [2:51] because I was a quirky kind of founder [2:53] mode type of person. Always being [2:56] everyone's trying to talk me out of [2:57] doing that. What was your reaction? [2:59] >> I didn't have too much of a reaction to [3:01] be honest. Yeah, [3:02] >> nowadays, uh, you know, Jack Dorsey [3:06] kind of redesigning his company to put [3:08] AI in the middle and Brian Armstrong [3:10] doing the same thing. My sense is you're [3:12] kind of heading this way as well. Um can [3:15] you talk about your journey there, your [3:18] company's journey? What's new? How are [3:20] you thinking about the organization? [3:21] How's it changing? [3:23] Yeah, my journey actually I have a small [3:25] detour off the founder mode um which is [3:29] like try to delegate trying to grow your [3:32] notion has very been small as a company [3:34] in the early days like we had product [3:36] market fit 2019 early 2020 [3:40] uh we're sub 10ish people 2019 and um [3:44] company become profitable because we're [3:46] small right um [3:48] >> did you raise venture in the early days [3:49] or you bootstrapped it [3:51] >> we raised venture [3:52] >> you Okay, that must have been tough. [3:54] >> Um, was tough like first five years just [3:57] seage money dragged us for five years. [3:59] >> Yeah, HubSpot was like that. Took us a [4:01] long time to get the product market. [4:02] >> It's very tough. But [4:04] >> yeah, it's a different type of tough [4:06] compared to now. It's back then there's [4:07] no product market fit. It's despair. You [4:09] just black. You don't know where to go, [4:11] right? Um so we have been very small [4:13] from the get-go and we sort of taste the [4:15] sweet of it because it's you're small [4:17] you're profitable easier to raise money [4:19] you're default alive and um we've been [4:22] resisting growing team size and and [4:25] during co a lot of companies really blew [4:28] up remember that right [4:29] >> I'm sure you did [4:30] >> we relatively yeah blew up [4:32] >> you got to product market fit at the [4:33] right time [4:34] >> we grow to product market around the [4:35] same time um there's a lot of eyes that [4:37] you got to grow so we did grow um [4:40] relative to other company still [4:42] resistant to grow a lot uh in terms of [4:44] headcount. Um but that's the moment I [4:47] sort of learned oh you need to delegate [4:48] need to have introduced in professional [4:50] manager management [4:51] >> why who is telling you that [4:53] >> it's a good question who te tell us some [4:56] investors [4:57] you didn't tell me that I don't think so [5:00] uh [5:00] >> it was conventional wisdom amongst the [5:01] investor class [5:03] >> I would say there's a pro and con for [5:04] that like we sort of eyes resist that [5:07] for a long time so we did not build a [5:09] sales team in retrospect I should have [5:12] started sales team earlier because I [5:14] thought I can first principle the whole [5:16] way like create new kind of PLG some [5:19] kind of brand new motion that's a [5:21] mistake um the other part is product [5:24] marketing engineer those functions sort [5:27] of I wouldn't call mistake but we're [5:29] upping the traditional way of building [5:31] it [5:31] >> did you hire been there done that people [5:33] in that phase [5:35] >> we we did [5:36] >> and do you have any of them still [5:38] >> um many of them are still with us yeah [5:40] our CTO it's last three years before [5:44] that there's another CTO during that [5:47] phase. Um but I would say some part of [5:50] notion DNA are defined by the classic [5:54] SAS executive. Now the thing changed in [5:58] the past three years since we're [6:00] building with AI like we're one of the [6:03] earlier [6:04] um adopter of language modeling or [6:07] product and you start think things are [6:09] breaking building with this new [6:11] technology the classic boundary between [6:14] rows doesn't quite work the analogy I [6:16] like to use back then it still is is [6:18] like building classic software is like [6:20] engineer bridge if you can design it you [6:23] usually can build it fairly predictable [6:26] um like product manager talk to [6:29] customers pass to designer pass to [6:31] engineers right building with language [6:34] model back then and somewhat still is [6:36] it's like brewing beer [6:38] >> you can't truly predict the things the [6:40] underlying thing good analogy you cannot [6:42] tell the east hey hey guys go go go go [6:44] go towards that flavor profile more you [6:46] can't do that right you sort of have to [6:48] throw your best people in there [6:50] >> see what the model see what technology [6:52] gives you so it's not customer first [6:54] it's more experiment with this [6:56] technology first. Although the last year [6:58] or so model gets so much better now but [7:01] still the spirit is technology [7:02] firstdriven development rather than [7:04] customer-driven first development. Um [7:07] that really forced us to change the at [7:09] least how we build product. Um designer [7:13] is the most productive thing are [7:14] designer and engineer and product people [7:16] all just sit in the same bucket then [7:19] work with evals work with experiences [7:21] and figure out [7:22] >> inside your company has one of those [7:23] personas kind of dominated and the other [7:25] ones falling aside or is there they all [7:28] kind of do the same thing now [7:30] >> we always started by like our DNA in the [7:33] early days always blurry like we are one [7:35] of the first to hire designer who can [7:37] code I'm a designer who can code right [7:39] so we we build a company based um my [7:41] image and my company my co-founder's [7:42] image so those profile really shines [7:45] designer who can co- engineer have [7:47] strong product sense um PMs don't not [7:51] afraid to roll up sleeve and try things [7:54] like one of our heaviest token [7:57] consumption person in the company is a [7:58] pad for example [8:00] >> those people really shines and we have [8:02] quite a few of in the notion [8:04] >> okay so you read the Dorsy stuff you [8:06] read the Armstrong stuff my summary of [8:09] it is they Think of like a big triangle [8:11] or chart. It's very inefficient flow of [8:13] information up and down and across it. [8:15] It's been around for thousands of years [8:17] and it's time to rethink it. Basically [8:19] stick AI in the middle and start [8:21] delegating more and more over time and [8:23] letting the system make decisions and [8:26] feeding it context in good ways. That's [8:28] kind of my highle summary of it. What's [8:31] your take on what those two are doing [8:33] and are you doing basically the same [8:35] thing? If some sense we're creating [8:37] product for other company to do that [8:39] >> but yourself [8:40] >> the high premise is similar at least my [8:42] interpretation of it it's so much [8:45] knowledge work is paper pushing in the [8:47] past it's really hard require software [8:50] to push information from one person [8:52] another but so much of a knowledge work [8:54] is fuzzy so human need to be in the [8:55] middle to pushing information from one [8:57] to the other and turns out language [8:59] model can do really good at two things [9:01] one thing is to write the software to [9:04] design the perfect pipes for your [9:06] company to pushing information and [9:08] decisions. Second is like doing this [9:10] kind of mini decision making in the stop [9:12] between one decision point to another [9:14] decision point. Right? So a lot of used [9:17] to companies to a lot of people just in [9:19] the middle to pushing information [9:22] decision around to sustain a business at [9:25] scale that could be rethink. [9:28] >> Okay. The analogy I like to use, we've [9:31] been using somewhat is like before steel [9:35] buildings cannot be higher than five six [9:37] floors and [9:40] New York City are largely six floor tall [9:42] iron and stones and bricks [9:45] >> after steel there's you information or [9:48] way structure can go much higher since [9:51] sometime language model plus software is [9:53] that it's the steel for organizations. [9:56] What does the organization look like now [9:57] that we have the steel? [9:59] >> I think [10:00] >> I know buildings got taller. What what [10:01] happens to or organizations? [10:03] >> Uh I think everybody's trying to figure [10:04] out right like Dorsy's structure is like [10:06] three layers or something or three type [10:08] of people, right? [10:09] >> Yeah. Three types of people that sit [10:11] around the AI system that mostly train [10:13] it um extremely flat and I think Brian [10:16] Armstrong said it's five layers and [10:18] everyone's a player coach which I think [10:20] is pretty interesting. [10:20] >> I think all the spirits are right. Yeah, [10:22] >> I don't think and I believe it should be [10:24] a hierarchical like I think people tried [10:27] in the in the past like flat [10:30] organizational structure I don't think [10:32] that's part of human nature human nature [10:34] are hierarchical [10:35] >> and do you think that's changed now we [10:37] have AI that it can be flatter or you're [10:39] like that's BS [10:39] >> oh definitely can be flatter like we are [10:42] getting flatter and flatter [10:43] >> okay how many direct reports you have [10:45] >> I have 7A [10:48] >> um in generally what is it [10:49] >> the companywise probably around 7 8 9 [10:52] people per person. [10:53] >> A lot of people are going to 15 20 25. [10:56] >> Yeah. Many of our best people 15 20 [10:59] something. [10:59] >> Got it. How have you changed your [11:02] company? How you compensate people? How [11:05] you build your org? How you hire people [11:08] just in the last year as all this stuff [11:10] has developed? [11:11] >> Yeah. Two years ago, we start to um hire [11:15] different shape of engineers. [11:16] >> Okay. um no longer optimized for [11:18] people's experience, optimized for [11:21] people's um agency, energy, optimism, [11:25] curiosity. A rough framework we're using [11:28] like okay what is the talent in the [11:31] company? A talent equal to this person's [11:34] capability s experience times this [11:37] person's um taste or value system times [11:40] this person's um agency will right um [11:45] what language model change is like just [11:47] like Google everybody can access [11:48] information language model allow [11:50] everybody to be a pretty good writer and [11:51] programmer so capability got normalized [11:54] democratized and taste becomes still [11:57] important like what's your value system [11:58] what do you want to bring to the world [12:00] which direction do you want to that you [12:01] cannot change as easily and you will how [12:05] hard do you work this you cannot change [12:06] so we optimize for the latter to today [12:09] >> okay [12:09] >> that means in engineer science we're [12:12] hired for much younger people okay [12:13] >> internally we're call like IC early out [12:16] of school IC [12:17] >> this is the opposite of the way most of [12:19] the CEOs I'm interviewing are going it's [12:21] interesting [12:21] >> I think more people are thrilled [12:23] >> they say that like a junior engineer [12:24] powered by the stuff's more productive [12:26] but if you got a senior engineer that [12:27] understands the whole system been there [12:28] a while and they're pled they're like a [12:30] 100x. [12:31] >> Yes. That's so higher for super junior [12:34] super senior. [12:34] >> Oh, interesting. [12:35] >> So the shape it's not it's more like the [12:38] >> barbell barbell shape, right? Um because [12:40] the senior provide taste there's still a [12:43] lot of auto distribution [12:45] information in language model [12:46] architecturing and language models. No [12:49] one very bad with that. So they need to [12:51] provide a direction. It's almost like [12:54] this a good engineer managing four to [12:57] six coding agent at once, right? Um but [13:01] a good really senior architectural [13:03] engineer can managing two to three [13:06] junior interns or engineer they each [13:08] manage two or three then that's the [13:10] number is higher right and they can [13:12] train the next generation. Uh I think [13:14] that might be a more optimal shape than [13:16] bunch of senior people manage coding [13:17] agents. [13:18] >> Okay and you I assume you're still [13:19] hiring designers and still hiring PMs. [13:21] Is the ratio changing or are you just [13:24] hiring people do the whole thing? The [13:26] definition of designer PM are different [13:28] like we're intentionally hired designer [13:30] who work as PM PM can usually the other [13:34] way usually it's designer who can work [13:36] as PM it's harder for PM to work as [13:37] designer because it's the [13:39] >> the visual genetic [13:40] >> thing it's more like a visual craft [13:42] takes years to train right um but we're [13:45] >> but designers want to talk to customers [13:46] and get feedback [13:48] >> yeah rare but uh don't stereotype [13:51] designer though they do exist [13:52] >> so if I look at your website you're [13:54] hiring engineer junior senior are you [13:56] hiring PMS and designers or do you have [13:58] some [14:00] P I don't know how to combine those two [14:02] words together [14:03] >> we hire PM designer by the way we [14:04] interview them we evaluate them like [14:07] we're in talk with two to three design [14:09] leaders they're kind of we value them as [14:12] PMs basically are you how in can you [14:15] drive things yourself it's not just come [14:18] up good ideas um how much are you [14:20] interfacing with customers besides [14:22] you're good with craft design craft Some [14:25] of this stuff is like the fashions [14:27] change. Like for the longest time it was [14:28] experience. Then the word everyone used [14:30] was slope. We hire for slope and now [14:33] it's kind of taste and curiosity and [14:36] agency. [14:36] >> That's kind of slope though. That's like [14:38] the [14:38] >> slopeish. [14:40] >> How is it different? [14:41] >> Slope to me is just horsepower smart. [14:44] >> Well, intelligence is first intelligence [14:46] multiple dimensions [14:47] >> like GPU speed like how'd you do on your [14:51] freaking SAT? [14:52] >> But there's a plenty smart people are [14:53] lazy. That's for sure. [14:55] >> Do you call that high slope or low [14:56] slope? [14:57] >> I don't know. I just always bucket it in [14:58] to hire the the the the the people who [15:01] learn who learn fast and are very smart [15:05] and I think it's changed to the word [15:07] taste comes up a lot and you're you've [15:09] been talking about taste for a long [15:10] time. The industry talks. Why is that [15:11] all of a sudden a thing? Why is that [15:13] word everywhere now? [15:15] >> Because taste is now in language models. [15:17] >> Yeah. taste is in in if you zoom out a [15:21] long long time maybe is but right now [15:23] it's not. Yeah. [15:23] >> And it taste is more rooted based on [15:26] your value system. Are you drinking [15:27] sparking water or still water? Are you [15:29] drinking diet coke or iced tea? [15:30] >> Yeah. [15:31] >> It's to each of their own. There's no [15:32] right answers. [15:34] >> Um the rest of you, you've got more than [15:36] just engineering. Uh you've got a bunch [15:38] of HubSpot people here which is great. [15:40] Uh but let's just take marketing. Is the [15:43] criteria change there too? We actually [15:46] changed our marketing department. We no [15:48] longer have a CMO organization. This is [15:51] this year. Last year, this year. Yeah. [15:53] We break up our CMO organization into [15:56] storytelling. [15:57] >> Yep. [15:58] >> Which is closer to product, closer to [16:00] because the product is changing so fast. [16:02] It's like you just classic marketing [16:04] can't keep up. We haven't shipped this [16:05] fast before, right? So let's just have [16:08] that sit next to product directly [16:10] connect to social where people are [16:12] discussion are happening and then the [16:14] other part is like serving the sales go [16:16] to market function like demand genen [16:19] >> lead genen just like um we really really [16:22] focus on that so rather than information [16:25] round trip to a CMO then figure out how [16:27] to serve both the product side and the [16:28] go market no more just like let both [16:31] side okay [16:31] >> figure it out themsel okay [16:33] >> not themsel but [16:34] >> more decentralized [16:35] >> and are there any other of classic [16:37] marketing functions in there. It's sort [16:39] of those are the two. [16:40] >> Those are the two. Um the designers [16:43] there are also designer and creative [16:45] part on the storytelling side [16:47] >> like community. [16:48] >> Yeah, community is in there. It's a [16:50] community where as part of a community [16:52] ecosystem uh or organization which is [16:56] startup where is one of our most [16:58] important communities um notion fans [17:02] around the world and figure out how does [17:04] this to move our market through those [17:06] people right um that's the bucket of [17:09] people [17:09] >> and has the criteria changed sales [17:12] marketing all the different other [17:14] functions have you changed the way you [17:16] think about the hire like most companies [17:18] have a form when you interview someone [17:20] has that form [17:20] >> change we're actually in the middle of I [17:22] just described engineer one right [17:24] engineer designer it's the we changed [17:26] two years ago yes um go to markets [17:29] marketing would change no six monthsish [17:32] ago sales were changing right now [17:34] >> and when you hire a sales rep they have [17:35] to be AI [17:36] >> pil yes there we're actually changed our [17:39] first interview no longer look at your [17:42] resume build something for us send a [17:44] notion link send something that you [17:46] built we look at what you built How [17:49] about compensation? Is that changing? My [17:51] sense is we we had an environment where [17:53] at least in HubSpot we people are much [17:55] more productive than the average person. [17:57] Now it's even more so. Have you [17:59] rethought? [18:00] >> We are changing that as well. Okay. So [18:02] they need to be extremely lot more [18:05] meritocracy. You can peanut butter [18:07] things. Um I think the SAS era is kind [18:11] of like playbook peanut butter [18:14] everybody. It's like um pretty peaceful [18:18] era. Um and right now it's a wartime [18:21] >> totally [18:22] >> actually make the SAS feel like what are [18:24] we doing back then for [18:25] >> well nothing changed for the longest [18:27] time like [18:28] >> nothing right like notion we had [18:31] >> that's the reason I left HubSpot it's [18:32] like okay we're building this platform [18:34] we need to build build more apps we need [18:36] to keep opening the platform we need to [18:37] hire more reps and just like keep going [18:40] >> until a couple years ago now it's [18:41] completely changed [18:42] >> or you can't well come back to become [18:43] operator if you like [18:44] >> that's why I joined Skoya I'm like wow [18:46] there's so much exciting things going on [18:47] >> when we hit product marketing CO [18:49] happened. It's more like um world [18:50] changing but not like a people building [18:53] things and technology changing. [18:54] >> Do you prefer wartime or peace time? [18:56] >> Oh, wartime is way more fun. It's like [18:58] >> what's the difference? What do you do [18:59] differently? [19:00] >> You feel more alive. [19:01] >> Okay. But what methods do you use that [19:03] are different? [19:04] >> I need to take care of my body. I need [19:05] to exercise every day. [19:07] >> I didn't think you were going to say [19:08] that, but okay. [19:08] >> Yeah. It's like it's uh um more fluid. [19:13] >> Are you more top down versus bottoms up? [19:15] >> I think the company has a good spot. [19:17] notion right now is a good spot with me [19:20] in a sense that I can dance around [19:22] really smoothly either direction. [19:24] There's enough people I trust really [19:26] well. They can bottom up and it's [19:28] constant surprise me like we're a [19:31] company a thousand people we have about [19:33] 50 60 founders in the company [19:34] exfounders. It's like they really can [19:37] lead things right. um and and they're [19:40] welcome me to top down into jump into [19:43] any part I can work on, I'm interested [19:45] in where I can provide you new value and [19:47] they don't feel territorial. This is in [19:49] a good place that people who didn't feel [19:51] that way either no longer at notion or [19:53] let them go. Um so the company I me in a [19:56] pretty interesting equilibrium at the [19:58] moment. I interviewed Parker Conrad from [20:00] Ripling and he's similar has a lot of [20:02] founders and the way it kind of strikes [20:04] me is like the bigger HubSpot got [20:08] almost all the important things cut [20:09] across all those organizations and it [20:12] was just very hard to find someone who [20:14] could really drive something across the [20:17] org as opposed in their org. So we did [20:20] not do that. I I regret not doing more [20:22] of that. That's my real remember I say [20:25] right after co we scale and then we try [20:27] to scale more traditionally I realized [20:29] that's not us right [20:31] at that point this is was three years [20:34] ago we start a mantra in the company we [20:37] want to be a jazz band not a marching [20:38] band [20:38] >> I love that I I do the orchestra versus [20:40] a jazz band I like the marching band [20:42] >> yeah like I realized I cannot be a [20:44] marching band person that's just not me [20:45] I will feels like if everybody I [20:48] delegate everything everybody do thing [20:50] what do I do I feels like I I can't do [20:52] that. Um, so it's like a self-reflecting [20:56] value thing. It's auto equilibrium with [20:58] me. So it's like, okay, I'm a jazzband [21:00] person. So we've been hiring more and [21:01] more jazz band person also AI happened [21:04] in the past two two and a half years. So [21:06] more jazzband people can really shine [21:09] during this moment. [21:10] >> Okay. So we had manager mode, we have [21:12] founder mode. There's this new mode. [21:14] What should we call it? [21:15] >> Jazz mode. Maybe [21:16] >> that's not bad actually [21:18] >> because Jasmo has some structure, right? [21:20] Yeah, [21:20] >> you don't want to be like you want other [21:22] people to contribute. You don't want you [21:24] want people to have fun, not just you [21:26] having fun or dictating everything. [21:27] >> Okay, let's give that a go. Let's push [21:29] that out on on X and see if it sticks. I [21:30] mean, pretty good. Yeah, because I I put [21:32] a survey out the other day like, what [21:34] mode is this? And I was like, is it [21:35] Dorsy mode? Is it native mode? And I [21:38] said, is it other mode? And somebody put [21:39] in there, it's dep mode. [21:43] >> Dep [21:47] mode. I think that's a good idea. One [21:49] other question on this new world. Um, I [21:52] work with all these startup founders. [21:53] The planning's tricky and I I sat [21:56] through your pitch at Sequoia and you [21:57] had a plan and you had a P&L like what [22:00] is it actually like internally planning [22:02] wise? [22:03] >> Well, group has our plan which most [22:05] company at this point if it works if [22:07] group has your plan, right? So um I [22:10] would say the financial planning is [22:11] still at least I find useful to give you [22:15] a like you know you're on a treadmill [22:18] you know how fast you're running right [22:19] that's like a financial plan in my mind [22:21] it's like oh am I running seven eight [22:23] one nine what number that is but product [22:26] strategy it's oh there's no freaking [22:28] plan no plan uh because what changes the [22:33] market the technology is so fast it's [22:36] not in the six month it's not three [22:38] months It's week by week. So this you [22:40] have to jazz this thing. Financial [22:42] things march that right. Um then at [22:46] least we try to be conservative mid to [22:48] conservative with financial things and [22:51] ex take risk and there's a buffer like [22:55] cost becomes another dimension like if [22:57] you're just building software don't [22:58] think about cost. [22:59] >> No [22:59] >> now you have to like at least I have to [23:01] really think about cost. [23:02] >> Can I ask you about that? Like a lot of [23:04] these startups gross margins are awful. [23:08] And like I companies like HubSpot were [23:10] like 86% gross margins. That's sort of [23:12] SAS normal. What are happening to your [23:15] cost and your gross margins? And where [23:17] do you think it settles out for [23:18] companies like Gurus? [23:20] >> I don't have an answer for you. [23:21] >> I'm assuming your gross margins have [23:23] gotten worse, but maybe you didn't hire [23:25] as much headcount, so maybe operating [23:27] margins are better. And how willing are [23:29] you to be like gross margins to get [23:32] radically worse? [23:34] >> Oh, you have to. [23:35] >> You're you're down for that. you like [23:36] what are you doing then you're not in [23:38] the war mode right I think notion is for [23:41] knowledge work which might have a [23:42] different curve than coding products [23:45] coding product you run frontier models [23:47] usually the more the smarter the better [23:50] um and knowledge work you you can we [23:54] already start to see this you sort of [23:56] can settle with the the second year the [23:59] Chinese models the open way ones they're [24:01] not super smart like a lot of this kind [24:04] of paper pushing information in your [24:05] company oh I spill coffee on the carpet [24:08] can HR team come someone to tidy this up [24:11] next week that doesn't require opus to [24:14] file tickets right so that's a lot of [24:16] problem we are best at solving at notion [24:19] so that will have a different growth [24:20] margin profile than [24:23] >> yes [24:23] >> the biggest product market fit today [24:25] which is a coding agent [24:26] >> yes okay [24:28] you're kind of in my mind the king of [24:30] refoundings [24:32] >> you've done it I think twice at least [24:35] Claude thinks you've done it twice. [24:38] Uh, your first one, you and your [24:41] co-founder moved to Kyoto, locked [24:44] yourself in a very small room and [24:46] rebuilt. [24:46] >> It's not that small, but yeah. [24:48] >> Okay. You know, more recently, [24:49] obviously, you you this everybody knows [24:51] the story. You went to Mexico and you [24:53] had early access to the models and you [24:54] turned into an AI company. Talk about [24:56] the different refoundings and what was [24:59] your mindset at the time. There's lots [25:01] of early stage founders who need to [25:02] refound and there's a lot of bigger SAS [25:05] companies that definitely need to [25:06] reound. [25:07] >> Yeah, I like as an idea as a the best [25:10] artists reinvent themselves. Um Miles [25:14] Davis has three at least three face, [25:16] right? Picasso two to three faces. Uh [25:21] Michael Jordan the second one probably [25:23] not as good. No, but the the best one [25:25] truly Steve Jobs the king of refunding. [25:28] Uh the first version of notion is like [25:30] we took us four five years to find [25:33] product market fit. In the middle of [25:35] that we sort of find the shape of the [25:37] product we want to build but we'll be [25:40] running out of money building it. So my [25:42] co-ounder se and I said let's just lay [25:45] off everybody just go by the two of us. [25:47] That's the Japan story. Uh we rebuilt [25:49] notion from Japan. Um [25:52] >> how hard was that the laying off [25:54] everyone? It's not easy, but sometimes [25:57] your body just tells you you have to do [25:58] it. Like you know you're dead if you [26:00] don't do it. Uh you know just there's [26:03] like a panda energy just like [26:05] >> why the hell did you go to Kyoto of all [26:07] places? [26:08] >> Well actually we never been to Japan cuz [26:11] when you just say goodbye to your [26:12] friends all those people you work with [26:14] notion by all means not big five [26:16] peopleish. Um it's morale so low right. [26:20] So S and I was saying like let's go [26:22] somewhere that we just changed the mood [26:25] a little bit and we never been to Japan. [26:28] So initially I picked Tokyo on Airbnb [26:31] realize the the house are so small. I [26:35] don't I love Simon but I don't want to [26:36] cram into a small apartment with him. [26:39] And Kyoto the apartment much bigger [26:42] because they didn't get burned on the [26:43] World War II. [26:44] >> They weren't bombed. [26:45] >> They weren't bombed. They spared that [26:46] city. too much treasures and the house [26:49] is much bigger so and cheaper so we [26:52] that's why we picked Kyoto and we ran [26:54] our place in SF uh and our office in SF [26:57] we actually become cash flow positive [26:59] doing that um that's the first time but [27:03] once you land you start coding just [27:06] those kind of rep coding eating coding [27:08] eating um it's actually really [27:10] liberating [27:11] >> I lived in Japan for a couple years was [27:13] there something about Japan that was [27:16] inspirational [27:17] in your refounding or the way you built [27:19] the product. There's a certain aesthetic [27:21] there that sort of rhymes with notion. [27:24] Or could you have been in anywhere? [27:26] Could you have been in Thailand? [27:28] >> Well, the usually we like [27:31] summer warm. Simon likes warm climate. [27:33] Ideally, he can swim. Um, Japan just [27:36] happened to be were both curious about [27:37] it. But once you go to Kyoto, I don't [27:39] know, you probably been to Kyoto. It's a [27:41] very special city. [27:42] >> Very special. [27:43] >> Yeah. [27:43] >> Very inspirational. [27:44] >> Very inspirational. Those things are [27:46] hundreds everything. Oh, this shop 150 [27:49] years old, right? I'll open that door. [27:51] The the shrine gone to a beautiful place [27:53] in Thailand, Phuket, whatever. Do you [27:55] think you still would have pulled it off [27:57] or was there something magical about the [27:59] location that really worked? [28:01] >> Most of things are a story we tell [28:02] ourself, right? [28:03] >> Of course. What story is the true story? [28:05] >> The story we tell ourselves is like [28:07] Kyoto is a special place. If you can [28:10] pull off anywhere, you can pull off from [28:12] Reborn in Kyoto. Then it's it's also [28:14] inspiring just look at the surrounding [28:16] the craft is probably craft capital of [28:19] Asia right so just make you okay if [28:23] notion fundamentally is tool is a tool [28:26] for humans and those people are building [28:30] knives ceramic cups tatami seats like [28:35] they're the tools other tools for [28:36] >> craft craft central [28:38] >> it's a craft central and like how do we [28:40] how can you not be inspired by building [28:42] a better software Okay, you still [28:43] haven't answered my question, [28:44] >> which is [28:45] >> if you had gone to [28:48] whatever, let's pick some place in the [28:50] Philippines and it was pretty, [28:52] >> would you have pulled this off or was [28:54] there something exceptionally important [28:57] about being immersed in such a craft [28:59] ccentric place? [29:01] >> I think I pulled it off, but yeah, but [29:03] the story is the better story. [29:05] >> Okay, fine. Okay. So, your big [29:07] refounding, not easy. [29:10] Didn't immediately hit product market [29:12] fit after. Still not easy. [29:13] >> No, it took a while after that because [29:15] it took a year and a half to rebuild [29:17] from scratch. [29:17] >> Why didn't you just quit and start a new [29:19] company at that point? [29:21] >> We never start a company for starting a [29:23] company. You want to build something [29:25] like I got obsessed about this notion [29:29] shaped tool problem since last year [29:31] college and Simon and I were both in [29:34] this kind of tools for thought community [29:35] which is you can trace the lineage back [29:37] to grateful that era is the the first [29:40] generation computing pioneers around the [29:42] Bay Area like the lineage trace way back [29:45] directly back to there. I want to [29:47] continue that lineage. Um so to me it's [29:50] not [29:50] >> it wasn't like you had another idea [29:51] kicking around. this is your life and [29:53] this is like bigger than I can ever chew [29:55] on this idea. So I wouldn't do anything [29:57] else. He's in that community. He's [29:59] obsessed with this idea too. So if we [30:02] don't work on this, we'll work on [30:04] another company doing the same thing. [30:05] >> Okay. I'm going to I'm going to get back [30:07] to the reffounding, but just on the [30:09] Grateful Dead, [30:11] uh you study very important computer [30:14] scientists from the 60s. You're a [30:16] student of Steve Jobs. I'm sure there [30:19] was sort of a craft ethos about the [30:21] Grateful Dead and the music back then, [30:22] but certainly about the company, [30:23] certainly about jobs. [30:26] Is it still here? It seems like it's a [30:28] lot more commercial than craft today in [30:30] Silicon Valley. Have we lost that? [30:33] >> I think most people don't have it. Most [30:35] company don't have it. [30:36] >> I I I agree. like Steve Jobs I look up [30:39] to him or people like him is like [30:42] interception of humanity with technology [30:46] versus Silicon Valley at least tech is [30:48] largely a technology and tinker culture [30:51] and tinker culture doesn't even know the [30:53] history of it's not even science it's [30:54] tinkering right it's like people like to [30:57] build things in their garage and [30:58] something works and maybe has a product [31:00] market fit and grow a business out of [31:01] this [31:02] >> to be fair W and W was tinkering in the [31:04] garage [31:05] >> yeah but you required to meet someone to [31:08] to bring the human side of it, right? Um [31:11] I think tinker culture the the risk of [31:13] us like you don't know what comes before [31:16] you. Science at least you respect the [31:18] history. Like if we ask people who um [31:21] Douglas Anglebar or Alan K those [31:23] computing pioneers are most people in [31:26] tech have no idea. [31:27] >> Honestly the only reason I know them is [31:28] I work for this guy Ray Azie. [31:30] >> Ray Azi like [31:30] >> who you would know of course because [31:32] he's the father of notes but he was [31:34] obsessed with those two and talked about [31:35] them constantly. reality uh it's like [31:37] Microsoft CTO for a while too right so [31:40] um that lineage is broken by the way the [31:42] people who remember this 807s and 80s [31:47] they were sort of people don't even yeah [31:50] they no longer work in tech today um so [31:53] tech is like industry doesn't know its [31:55] past if you don't know his past you [31:57] don't know history which is humanity I [32:00] think your your your point of view is [32:03] like what's in front of you what your [32:04] competitor is doing Versus if you bring [32:07] humanity, you have all other discipline [32:10] around you. You have all the history [32:12] behind you. There's way more good stuff [32:14] you can steal from. Um I wish more tech [32:19] founders [32:21] steal from aside or from behind. [32:23] >> I think they'd benefit from it too. Um [32:26] back to the refounding, [32:29] lots of founders are going sideways. Uh [32:32] lots of companies that started in 2021 [32:34] are going sideways. They've got 10 [32:36] people. It's like maybe they're cash [32:38] flow break even. What's your advice to [32:40] somebody? Is it a founder been going [32:42] sideways for a long time, got venture [32:44] dollars in like you and working on [32:46] product market fit? Like should [32:49] everybody just be like rip the band-aid [32:50] off and shrink it and move to Kyoto and [32:53] and refound [32:54] >> if their body tell them so? Like I for [32:57] me it's like there's you just feel you [33:01] have to do something drastic and um [33:04] Simon during that time was he broke his [33:07] he got his wisdom teeth taken out or [33:09] something. So I'm by him myself. I was [33:11] like I remember I was at my [33:14] ex-girlfriend's studio. I just like [33:16] inner urge like there's no way out. I [33:18] have to do this. And it's like then then [33:22] you feel liberated once you land in [33:23] Japan. I guess your advice would be is [33:26] probably more people should do it and [33:29] they should listen to their inner body [33:30] and be more risk-seeking on that because [33:33] they can get stuck for a long time in a [33:34] ventureback startup. They can spend 15 [33:37] years of their career on something going [33:38] sideways. [33:39] >> I think there's no better time than now [33:41] to do that because the market dynamic is [33:44] wide open. [33:45] >> Yeah, I agree. Okay. You refounded a [33:47] second time when you had a thousand [33:49] employees. Uh [33:50] >> back then it's probably 500. Yeah. Okay. [33:54] 500 boys, different animal without [33:56] telling the same story you've told a [33:58] thousand times because I think everyone [33:59] knows the lore. What was it like kind of [34:01] behind the scenes? What did you feel? [34:03] Was it the same thing? Your body felt it [34:05] like what was the story? [34:07] >> So, the second time was in Mexico [34:09] Cancun, [34:09] >> right? [34:11] >> Happen to be out somewhere, [34:12] >> right? You're always in a nice place. [34:15] >> We just got early access GPT4 model. It [34:18] feels like the world is stopping. Like [34:20] when you first interaction with the GBD3 [34:24] is fine. GBD4 is a religious experience [34:26] for me. It's like holy It's a it's [34:29] a full body religious experience. Like [34:31] you have to do something with this. This [34:33] is what change everything. Anything you [34:35] do if you don't do this it will be [34:36] meaningless. Um I think the company has [34:39] been fairly there is be people doubt [34:41] because remember that around three years [34:43] ago that's the crypto era lending right? [34:46] A lot of people is this another crypto [34:48] thing. There were a lot of doubters. Um [34:51] um some of them obvious no longer at the [34:54] company but the conviction is so clear [34:58] at least for me and Sam early days like [35:00] we just have to do this. [35:01] >> Did you take two steps back in the [35:03] business to go 10 steps forward? Like [35:06] did revenue really slow while you're [35:07] rebuilding everything? [35:08] >> Our first products actually has a [35:11] meaningful revenue boost. We launched [35:13] our uh first AI product actually two [35:16] weeks before chat GPD happened. um it's [35:19] like AI writing product that has a boost [35:22] but after that we've been pretty much [35:23] want to build agent product in 2022 end [35:27] of 2022 and that's a slo that took us a [35:31] year and a half to really [35:32] >> didn't really work back then [35:33] >> didn't work at all like Simon [35:35] >> I would say my problem and Simon's [35:37] problem we might be living too much into [35:39] the future y [35:40] >> it bites you with new technology [35:43] like we try everything built a model for [35:46] us work openai finetuning or another [35:50] just none of them work that will like a [35:53] year and a half just go with no error [35:55] and morale is definitely low only start [35:57] to work a year and a half our revenue [36:00] grow inflection this around the same [36:02] time right when the AI product [36:05] >> the underlying got better [36:06] >> yeah it's like um but that was the that [36:10] was a pretty painful period [36:11] >> was that the most painful period in [36:13] notion history for you [36:14] >> different type of pain [36:15] >> what was the most painful [36:17] just when those two three early days [36:20] pre-product market fit is despair I just [36:23] like [36:23] >> despair is that the word okay [36:26] >> you don't know where it's coming from [36:28] you don't you don't know where to go [36:30] right um when your growth rate low like [36:33] during that period of time where we were [36:35] try to rebuilding our AI foundation [36:37] multiple times it's [36:40] yeah that's not great um right now just [36:44] everything on fire it's a different type [36:45] of pain Nice. So, a lot of people, 500 [36:48] person companies, a lot of SAS companies [36:53] are doing the AI thing, [36:56] but certainly haven't refounded. Advice [36:59] >> start with your product. We just talk [37:01] about treat it like brewing beer, not [37:04] building bridges [37:05] because that's end of the day most tech [37:08] companies is the product, right? So um [37:12] start with the products cuz then you the [37:15] founder has to be really hands-on to [37:16] experience that way of building things. [37:18] >> Can nonfounder-led companies pull it [37:20] off? [37:21] >> I don't know. Would be tough. I think [37:23] founder has the moral high ground to to [37:26] change things. People are more tolerant [37:28] and forgiving with founders, right? Um I [37:31] think the person need to be [37:35] use language modeling in in just know it [37:38] like you got to feel the AGI feel the [37:40] AI. [37:41] >> Yeah. You can't read about it. You can't [37:42] watch YouTube. You got to build [37:43] >> feel you have build it. Building is one [37:45] of the best way [37:46] >> for your product or build for your [37:48] internal systems. [37:49] >> Build for yourself. [37:50] >> Yeah. Tinker on the weekend tinker on [37:53] the side. you have to do that because [37:55] like at the end of the day business are [37:56] kind of like pathf finding entities [38:00] right like we're in this thing called a [38:02] market there's equilibrium for your [38:05] company with the market with with your [38:07] shareholder with your company like we're [38:09] just finding local optimas local [38:12] maximums for different our niches when [38:15] this new brand new ingredient got [38:17] introduced into the market in this case [38:18] language model doors are path that are [38:21] opened up and we are Each company at [38:24] different point in the market each you [38:26] have different angle you have different [38:27] niche you have to taking this [38:30] internalize this new ingredient figure [38:33] out which path got opened for you the [38:36] truth is with the customer with yourself [38:39] if you don't know how this this can do [38:41] you cannot find a new path I think you [38:44] have to do something you have to feel [38:45] this you feel the AI feel AGI [38:47] >> I think you're right you use a word that [38:49] I cringe at but I think is good called a [38:52] lot of companies calcify as they get [38:54] older. Let's just say you're a company [38:56] that's calcified. How do you get [38:58] uncalcified? [38:59] >> We intentionally do a lot of [39:01] acquisitions. We have 50 60 founders at [39:04] notion. [39:05] >> And were they aqua hires? [39:06] >> Aqua hires. Equires acquisitions. Um [39:10] >> founders are are kind of this kind of [39:12] like little decalcified meatthead [39:14] machinery just trying to break things, [39:16] right? So [39:17] >> keep your regenerating. [39:19] >> I agree with that. [39:19] >> Yeah. Um I find that really helpful. So [39:22] founders found for a reason and now all [39:23] of a sudden they're in a thousand person [39:25] company. How do you keep them motivated [39:27] and not leaving and starting something [39:28] new? [39:29] >> Well, first some of them work on a [39:31] domain that they were working on before [39:32] like the person who leading our AMO has [39:34] a startup of AMD notes before the [39:36] company the person leading our [39:38] enterprise search is a founder of [39:39] enterprise search product. So you give [39:42] actually give them a better platform [39:44] support to do this. They people feel [39:47] appreciated like I just had lunch with [39:50] the founder who start enterprise search [39:51] product before it's like [39:53] >> you give them leverage they give them [39:55] >> you give them leverage they actually [39:56] people care about the thing they're [39:57] building right versus fighting for [39:59] product market think really crowd the [40:01] market they don't you can argue that it [40:03] might be more difficult to start new [40:05] companies today compared to even a year [40:08] ago [40:08] >> okay [40:09] >> because the world's getting more noisy [40:12] like the shipping cadence from existing [40:15] companies are just increasing. It's like [40:18] the PE everybody got Twitter fatigit, [40:21] right? So, [40:21] >> and the SAS companies have got the memo. [40:24] Most of them got the memo and are [40:25] shipping fast now. You were ahead of [40:27] everyone, but [40:28] >> yeah, people like it's still [40:29] >> like HubSpot's cranking. [40:31] >> That's good. No, that people are still [40:33] kicking that they're they have to kick [40:36] in different gear. You know, [40:37] >> my thing these days is like it's it [40:39] could it's never been easier to start, [40:41] man. has never been harder to scale [40:42] because like in every little micro [40:44] segment if there's a little traction [40:46] like there's hundred competitors two [40:48] days later and the thing I've also [40:50] noticed about the startup land is almost [40:52] in every one of these verticals it's a [40:54] duopoly there's like a one and a two and [40:57] then 50 tied for third they're way off [41:01] when you look back in time that the the [41:03] the great companies were often founded [41:05] in shitty times because people who are [41:07] really grinders and people who really [41:09] want to start a company and really have [41:10] a vision didn't start in like 2022 2023 [41:13] a terrible time to start in times like [41:15] 2026 you get a lot of tourist [41:17] entrepreneurs [41:18] and so I've heard people say you know [41:20] 2026 ventures is going to be crap even [41:22] though the valuations are [41:23] >> like crypto crypto vibe energy [41:25] >> yeah a little bit [41:26] >> a little bit um I don't disagree [41:30] market will do the thing market is kind [41:31] of beautiful in that way [41:32] >> yeah advice to founders of reasonable [41:36] sized companies who are a little scared [41:38] to refound You know, I I'm CEO of a 500 [41:42] person company. It's going pretty well. [41:43] I'm doing some AI stuff. Uh my board's [41:47] kind of happy with me. Um [41:50] but there's a birdie inside of me [41:51] listening to Ivan and maybe I should [41:53] refound advice. Does a person need to [41:56] feel the AGI first? If you feel you [41:59] realize most tech industry is support [42:02] different flavor knowledge work [42:05] that we're at the beginning of how this [42:07] can do due to this industry. I agree [42:09] with you. How do you think organizations [42:11] look like? So they're changing a lot [42:13] now. What do you think they look like in [42:15] like three, four years? Is your [42:17] knowledge work? You're right at the [42:19] center of this. I think truth is [42:21] probably somewhere [42:24] there's a Dorothy Jack Dorsey's thing. [42:26] Uh [42:28] there is the traditional hierarchical [42:31] ones. Um [42:34] I like to think what are invariables [42:36] like let's look invariable in this case [42:39] human nature are rather static like we [42:42] all [42:43] >> we don't change that fast [42:44] >> we don't human don't change humans as a [42:47] species doesn't change that's where I [42:49] think it should it will be a [42:51] hierarchical organization if just [42:53] observe bamboo chimpanzees right it's [42:55] like very if you watch nature program [42:57] there's hierarchy that doesn't change [42:59] this every hasn't changed left us since [43:02] we left Africa. Um, division and labor [43:04] make sense. Um, so um people have [43:08] different interests. You you look at all [43:10] the different like personality tests, [43:11] right? They're just different way to [43:13] saying like people are different. [43:14] They're not 10,000 flavor of them there, [43:17] but there are probably a dozen flavor of [43:19] humans. They cluster around this [43:20] different dozen things. And we'll know [43:23] that our legal system well at least [43:27] right now there's no autonomous [43:29] companies, right? It's the CEO CF will [43:31] have to sign up for something. So it [43:33] will be a small team of human to start. [43:36] So those to me are the invariables. So [43:38] this invariable tells me you need to [43:40] figure out how those group of human [43:42] collaborate. The group of human all have [43:44] different biases, preferences, different [43:46] values. Um and it will be somewhat a [43:50] small some kind of hierarchal system [43:52] there. Then how do we use this new techn [43:56] new ingredient to passing information [43:58] decision around a small midsize group of [44:00] humans? That's how I think about it. [44:03] >> I kind of think about it as like right [44:04] now in companies the AI system makes [44:06] very few of the decisions. Let's call it [44:08] 1%. But as that system gets more context [44:11] using notion for example, it'll make [44:14] better decisions than humans. It'll have [44:16] more context and more history and then [44:19] humans will be more training that system [44:22] and giving it context and giving it [44:24] taste and I think over time the system [44:27] makes more and more of the decisions [44:28] instead of the humans but the humans are [44:30] still really necessary. [44:31] >> Yeah. People say AI researchers in 18 in [44:34] 15 months will not be able to make [44:36] decision to improve AI models, right? [44:38] Kind of the same thing but they're just [44:39] in a bleeding edge for this. [44:40] >> Yes. By the way, I think you're [44:42] incredibly well positioned for all this. [44:44] >> Oh yeah. I I [44:45] >> I I worked for a while with this guy [44:46] Rayazi. It was a company called Groove [44:48] Networks. We were in the knowledge [44:49] management industry, which is a [44:52] industry cuz chief knowledge officers [44:54] have no money and it was never urgent. [44:57] In theory, it sounded great. Now, [45:00] knowledge management is actually [45:01] incredibly important and works [45:03] incredibly well. [45:03] >> I know it's like what's that industry [45:06] called in the ' 90s? People hire [45:08] consultant to map out their orchart and [45:10] information flow. that become like what [45:14] AI and for deployment and [45:16] >> totally and your platform is ideally [45:18] suited for it. [45:19] >> Yeah. Um I think we always forget that [45:21] modern knowledge work is only about 150 [45:24] years old. [45:25] >> It's invented. It's not as old as fire [45:27] or language. It's only 150 years old. [45:30] Yeah. [45:31] >> Right. So [45:32] >> why can I be a new flavor of it? [45:34] >> Okay. I want to talk about you. [45:36] >> Okay. [45:37] >> I don't know you super well. You're [45:39] softspoken and I thought you were an [45:41] introvert, but in getting to know you [45:44] better. Was I wrong about that? [45:46] >> I think I'm still introverted. [45:47] >> Okay. [45:48] >> Yeah, I like one-on-one conversation one [45:50] too many. I train myself to do company [45:52] all hands just by doing a lot. [45:54] >> A lot of founders are introverts. The [45:56] world is sort of built for e I'm kind of [45:58] extroverted extroverts. [46:01] Advice for founders who just don't like [46:04] talking to other you don't like the [46:07] whole extrovert game. I have to evolve a [46:09] lot like I hate doing [46:12] >> force yourself like how to work a [46:14] significant other push you [46:15] >> no I force myself okay um to lead the [46:18] group of human you need to do one to [46:20] many communications otherwise people [46:21] don't trust you they want to hear from [46:23] you [46:24] >> writing is one mode which is like [46:26] sometimes ling num but even more so is [46:28] face to face to the all hands so for a [46:32] while I try to delegate all hand to my [46:34] co-founders to other exeacts no way I [46:36] have to now at every all and I'm open [46:38] the mic. I do it. I lead all hands. Now [46:41] I'm pretty comfortable with it. Uh I [46:44] still don't love it, but I'm well it's [46:46] not I don't dread it. [46:47] >> Okay. Advice for people doing all hands. [46:49] Everyone seems to do it differently. [46:51] >> Oh, get a teleprompter. [46:54] >> I'm the opposite. I can't I can't read [46:56] fast enough for the teleprompter. [46:57] Really? Yes. I hate a teleprompter. [46:59] >> Well, that changes me because like Well, [47:00] English is my second language, so it's [47:02] like if I have to think and speak, it's [47:04] hard. So I just prefer to spoken before [47:08] and right now all those like uh speech [47:11] to text model locals are so good. [47:13] >> They're so good. [47:14] >> So good. I literally like okay the night [47:16] before or then I just like talked to [47:18] myself and got what I'm roughly want to [47:20] say and the time doing all hands [47:23] teleprompter there. [47:24] >> Do you do them once a week? Once a month [47:26] once like what's your cadence? [47:27] >> We're actually changing our cadence. [47:28] used to be once a month. Right now it's [47:30] every other week uh for all hands and [47:33] every other week again for AMA. AMA. So [47:36] every week now. So [47:38] >> you ever step on yourself on an AMA? You [47:40] ever say something you're like I got a [47:42] crush for that? [47:43] >> Yeah. Everybody has it, right? Or just [47:45] you miss a good opportunity. You can [47:46] actually rally people about something, [47:48] but you you give a non-answer. That's [47:50] the worst. and non-answers. [47:52] >> We in at least at HubSpot in the company [47:54] meetings in the AMAs, there would be a [47:56] Slack channel on it and you go back and [47:57] look at it and be like, [48:00] >> what [48:00] >> just like if I said something stupid or [48:03] I missed an opportunity like it would [48:04] really show up in there. [48:06] >> That's good. [48:07] >> How do you manage your time in your [48:09] schedule and your deep work? Like how do [48:13] you manage your energy? Like for me, my [48:15] energy is pretty good right now. It kind [48:16] of goes up and down during the day more [48:19] than I would like. Is that the case for [48:21] you? How do you deal with that? How do [48:23] you deal with your introvert versus [48:25] extrovert? Like, how do you do it? [48:27] What's your schedule like? [48:28] >> It's actually not that crazy. Um, [48:33] wake up 7 7 a.m. make coffee for my [48:36] wife. Um, [48:38] I usually like to do work in the [48:40] morning, do some thinking, writing at [48:42] home, come to office. I come to work [48:44] every day. Uh, come to office. Um, [48:47] sometimes I'm a new habit I'm trying [48:49] right now is doing the gym during the [48:52] day. [48:53] >> Okay. [48:53] >> Yeah, [48:54] >> that's a good idea. Perk you up. [48:55] >> Elevate my energy during the middle of [48:57] the day and work in the office till [49:01] 7ish. [49:02] Usually social dinner with my wife, back [49:05] to work, sleep out till midnight. Around [49:07] midnight. [49:08] >> When do you have time to get down the [49:09] claw rabbit hole? [49:12] >> We can't We can't My happy time. No. Um [49:15] >> you describe that as your habitat. [49:17] >> Yeah. You can you can be led by your [49:21] curiosity rather than by your [49:22] responsibility. Right. And like usually [49:24] Saturday, Sunday morning, I I have a bit [49:27] time to read the books I want to read. [49:29] Um sometimes it's for work related, [49:32] sometimes purely just for my guilty [49:33] pleasure. [49:34] >> And during the day, like half hour like [49:36] I'm seeing found one of my founders here [49:38] in New York say who it is. He's got his [49:40] schedule in 10 minute blocks [49:43] >> meeting wise. [49:43] >> Yes. Like what does your day look like? [49:46] >> A lot of 25 minutes blocks. I usually [49:49] like five minutes between meetings and [49:51] 25 25. [49:54] My team pack my meetings together. So I [49:56] have one or two hour chunk to think but [49:59] most my thinking time is in the morning [50:01] or in the evening. Usually I get too [50:03] tired in the evening. I can't think [50:04] really well. Um morning is better. And I [50:07] do a lot of Slack. My my wife joke my [50:09] Slack is my social media. [50:11] >> Why doesn't not build a slack killer? [50:13] >> We're shut. Definitely. [50:14] >> Yeah. [50:15] >> A lot of your team said you're humble [50:17] and I think you are too. Why? [50:21] >> I wouldn't use the word humble. I use [50:22] the word [50:23] >> your team uses that [50:24] >> truth seeeking probably [50:26] >> they use humble. [50:28] >> I think you seek truth you know you're [50:29] just one perspective of the truth or [50:32] pluralism. There are multiple flavor of [50:34] truth in the world. You know the story [50:36] of blind people touching elephant. [50:38] >> Yes. [50:38] >> As um where just different blind people [50:41] touching different part of elephant. [50:43] Sometimes could be the tail, the [50:44] elephant's a snake, sometimes the trunk [50:46] is a treat, right? That's my worldview. [50:49] >> Okay. Tonight we're you we're going to [50:51] have dinner with a bunch of founders and [50:52] some are somewhere like 10 employees and [50:55] they hope to go and build a thousand [50:58] person company in the next several [50:59] years. What advice do you have for them? [51:02] >> It's a game. It's a game that's a a [51:05] social game. [51:07] It's like there's all the good part and [51:09] bad part of social game. like humor or [51:12] status is human thing, right? like um [51:15] power is this the CEO game is full of [51:18] that game just like entertainment just [51:20] like sports you have that like high [51:22] school college lunch sessions right you [51:25] you watch that in real in in the tech [51:27] scene it also could be a pursuit of your [51:31] personal values this is something I [51:34] realized more in the past three four [51:36] years it's like I need to be in [51:39] equilibrium with my own value of what [51:42] this company I want to build with that [51:45] give me more gen more sustainable energy [51:48] source. I cannot just chase something or [51:51] competition for competition's sake like [51:54] I I'm really competitive. I need to [51:56] compete to win. That's part of me. But [51:58] at the same time, I need to be true to [51:59] my own values. the value of building [52:01] good tools, the craft, the human side of [52:04] things, the social part of being a CEO, [52:07] the market dynamic just compete for [52:09] competing sake and the value the person [52:13] I wouldn't say there are this is some [52:15] slightly this is somewhat a zero sum [52:18] game. So you need to figure out where do [52:21] you want to sit in this spectrum and [52:23] your personality might be different. [52:24] Some people might just compute for [52:26] comput sake doesn't care about what [52:28] they're building. Some people just want [52:30] to be popular. If you just only care [52:32] about craft that you don't build a [52:33] business, it might as well just be a [52:35] artist. [52:36] Took me a while to realize where's my [52:38] equilibrium. [52:39] >> Are you someone that that that like [52:41] tries to fix your weaknesses or you just [52:43] lean into your strengths and what's your [52:44] advice on that? [52:46] >> Definitely lean on your strength. [52:47] There's a necessary thing you have to do [52:50] like one too many communication likely [52:52] you need to do with your company scale. [52:55] Maybe you can lean more on on writing [52:56] rather than speaking. But there's [52:58] there's certain invariable that works [53:00] right. Um within the boundary of that [53:03] you should amplify what you're good at [53:05] because capabilities are become more and [53:07] more commoditized with machines. So it's [53:11] your own point of view your strength [53:13] that you can live with that's that's [53:16] more lasting. [53:17] >> Okay. Something I assume that isn't your [53:20] strong suit is building an enterprise [53:23] sales or which seems like you're in the [53:25] process of building that. And when I [53:27] hear it's going exceptionally well, how [53:29] did you get your head around it? Um, how [53:32] did you hire? Sounds like you hired the [53:34] right person. So many companies are PLG [53:37] and are desperate to move to the [53:39] enterprise and the founders are similar [53:41] mold to you. Can you talk about that? [53:44] >> We had a mistake. I mentioned like we [53:47] sort of did not buy traditional sales [53:49] motion for a good two-ish years, right? [53:52] If we because notion sales could start [53:54] earlier but we want to first principle [53:56] our way into it which is we're going to [53:58] design our own system that doesn't work [54:01] like again it's a reflection of human [54:03] nature a lot of things you want to talk [54:05] to a seller there's a lot of the reason [54:07] the modern sales playbook has been [54:09] around for twoish decades right so [54:12] >> I think enterprise sales is like the [54:14] last function like [54:14] >> I agree with you it's like that you want [54:17] to see a human you feel comfortable [54:21] >> you don't want to have be a doctor from [54:22] a robot? No. Then why should you buy a [54:25] expensive thing from a robot, right? So [54:28] um it's not that hard I would say. Okay. [54:31] But you need to know like a lot of smart [54:34] people might think they can create [54:36] something new like fundamentally you [54:38] should not reinvent new things unless [54:40] it's really absolutely necessary. I was [54:43] trying to reinvent a new go to market [54:45] motion. That's stupid. [54:46] >> Okay. Like [54:47] >> I did mine. [54:48] >> You really build Hubspot. [54:49] >> Yeah. Right. May that's but that's the [54:51] thing made hotspot right like this [54:53] inbound [54:53] >> that was the thing [54:54] >> you should only have each company should [54:56] only preserve your innovation point to [55:00] few places go a couple places you cannot [55:03] spread everywhere like otherwise you're [55:06] spread too thin you're like you're [55:09] trying to like reinvent the wheel well [55:10] too many times we try to put our in [55:13] innovation point onto sales back then [55:16] just just go classic sales it works um [55:19] but you and make a lot more efficient [55:21] today with a lot of tools we're building [55:23] internally. Um, that's different. [55:26] >> Okay. So, you woke up one day, one day [55:28] you're like, we got to get good at [55:29] enterprise sales. What' you do? What are [55:32] the mistakes everyone makes when they do [55:33] it? Did you miss a hire or two before [55:35] you got the right person? [55:36] >> I thought this somewhat solved today, [55:38] right? [55:39] >> I kind of do, but people struggle with [55:40] it. [55:41] >> Um, [55:42] >> enterprise sales has changed very little [55:44] since I did it in the 1990s. [55:46] >> Yeah. So we tried um initially only half [55:49] uh PLG flavor of sales like system [55:53] thinker sales system sales leader um [55:57] didn't quite work. It's just like then [55:59] you're just taking orders and no like [56:02] the PRG give you the luxury that your [56:04] customer want to buy your product. You [56:05] just [56:06] >> easy to sell that you you it's like [56:08] >> fulfilling demand. [56:08] >> You cannot get your sellers to do [56:11] something difficult when there's so much [56:13] easier thing they can do. just doesn't [56:16] work. Then we realize okay yes we still [56:19] need to respect the system. So our CRO [56:22] is Erica's one she was CRO of GitHub [56:25] before it's a system thinker right we [56:28] pair Erica with our head of sale PVesh [56:30] which is the rah rah like the [56:33] >> meat eater [56:34] >> meat eater right so beans in his teeth [56:36] >> per perfect pairing and so we have [56:39] sellers with raoy proves but also [56:41] respect working with Erica's overall [56:43] system this is when things start to work [56:46] in the past a year and a half [56:47] >> got it and so many of the organization I [56:49] deal with the engineers look at how much [56:51] money the sales people make and look at [56:53] the rah rah and they roll their eyes and [56:54] they complain. I assume that happened [56:56] here. [56:57] >> No, we we actually notion culture is [57:01] always pretty receptive for different [57:03] people are different. So we never have [57:05] an organ rejection of sellers in the [57:07] company. Our our engineering team love [57:09] work with sales team. [57:10] >> Okay. Awesome. Speaking of culture, I [57:13] saw online some people refer to you as a [57:15] cult leader and some people referred to [57:18] me as a cult leader back in the day. I [57:20] didn't particularly like that. How does [57:22] that land with you? [57:23] >> I like that. [57:24] >> Okay. [57:25] >> A company is [57:28] it's sort of a religion is you have a [57:31] point of view. You have your value [57:32] system project to the world through a [57:34] business, through a product. Like my [57:36] favorite saying is Catholic Church is [57:39] one of the most successful companies of [57:40] all time. 2,000 years. They have their [57:43] ritual, their business models. They're [57:46] still around kicking. great founder in [57:48] Jesus and a great head of sales in Paul, [57:51] >> right? [57:52] >> Yeah. [57:52] >> And there are a lot of sellers and uh um [57:55] referral system, [57:56] >> viral marketing, [57:57] >> viral marketing, OG viral marketing. [57:59] >> Yeah. And to me, religion or belief [58:03] system or value system are again one of [58:05] the few dozen thing that's ingraining [58:07] human nature. Like people want to [58:10] believe in something to find meaning and [58:12] purpose. That's why wartime make this [58:14] more fun, right? your purpose and [58:15] meaning got amplified. Um work and [58:19] company give people that camaraderie [58:22] give people that. That's why we play [58:23] sports, right? It's the same energy. [58:26] >> The first four words in culture are [58:28] cult. [58:29] >> Yeah, I probably I need to look up the [58:30] roots. But yeah, it's like what is all [58:33] hand? People go to church every Sunday. [58:35] All hands every week. Same thing. We [58:38] used to have a thing with our all hands [58:39] that Dsh and I wanted it to be so [58:43] inspirational that people would well up [58:45] or cry during the all hands. Like we [58:47] really really worked on our all hands. [58:49] >> Yeah. Nothing I don't think anything [58:51] wrong with that. [58:52] >> No. [58:53] >> Okay. Thank you for coming on the pod. [58:55] Thank you for building notion. Thank you [58:57] for being inspiration for tons and tons [58:58] of founders. [58:59] >> I'm glad hopefully to be helpful for [59:01] others. [59:01] >> Appreciate you. [59:02] >> Thank you for your time. [59:04] >> Okay. I hope you like that episode with [59:05] Ivan. One of the things I like about [59:08] him, there's a there's a famous quote by [59:10] Oscar Wild. Be yourself, everyone else [59:13] is taken. I like to think I am myself. [59:16] Like I'm pretty quirky and I've embraced [59:18] that quirkiness. And I think at the end [59:22] of the day, companies really reflect [59:24] those CEOs. And I don't think CEOs [59:26] should be afraid to really be themselves [59:28] and kind of build the company around [59:30] them. Ivan's really done that nicely. [59:32] That might be a good plan for you, too. [59:34] The second kind of takeaway I had is, [59:38] you know, I work with lots of startups [59:40] and there are so many companies, let's [59:43] say less than 500 million in revenue [59:45] growing like 30%. [59:48] They're kind of stuck and there's [59:50] infinite numbers of stuck companies out [59:52] there. If you're in that boat, I'd [59:55] recommend a refounding moment. And the [59:58] refounding is harder than it looks. It [59:59] typically involves like a big step back [60:02] and two steps forward. It's going to be [60:05] hard. Your numbers will look worse in [60:07] the short term. Ivan didn't do this, but [60:10] when we wanted to do a refounding at [60:11] HubSpot at one point, we started a new [60:14] company inside of HubSpot that was [60:16] completely independent from it, and that [60:18] served us well. That's how Apple did it [60:20] back in the day. But if you're stuck, [60:22] and a lot of companies are stuck, be [60:24] risk [60:26] and do what Ivan did. like he was a ye [60:29] old SAS company and now he's a rocket [60:31] ship AI company. I think he's on to it. [60:34] The last takeaway I have on Ivan kind of [60:37] starts back in a podcast interview I did [60:39] with Jack Dorsey a few weeks ago. And it [60:43] does seem to me there's a new way to [60:46] build and run companies today. That's a [60:48] pretty big departure from the way people [60:50] have done it the last 30 years. Jack's [60:53] iterating on it. Ivan's iterating on it. [60:56] a lot of the call it vintage 2025 [61:00] startups are iterating on it and this is [61:03] something I'm thinking a lot about. I [61:05] think there needs to be a new playbook [61:06] on how to be a CEO in AI native in the [61:09] kind of era we're in. Um and I'm early [61:12] in the iteration but just roughly the [61:14] way I kind of think about the way [61:16] they're organized is historically there [61:18] is a triangle-shaped org chart. Those [61:20] org charts have been around a long long [61:22] time since the Roman Empire. They're a [61:24] very inefficient way to run a company, [61:26] but they've been the best way to run a [61:28] company uh because there's been no [61:30] better one. I think the problem with [61:32] them is the bigger that org chart gets, [61:35] the more that information gets distorted [61:37] across and up and down and the slower it [61:39] gets. And so it's pretty inefficient. [61:42] the new way that Dorsy described and I [61:45] didn't quite get into it with Ivan but I [61:47] suspect it's similar is it's a circle [61:49] and in the middle is all of the context [61:52] in the AI system and initially humans [61:55] are making all the decisions but over [61:56] time as you give that that kind of [61:58] centralized AI system more context it [62:01] gets smarter and you're feeding in [62:03] context and you turn over more and more [62:05] of the decisions to them I don't think [62:06] it happens overnight I think it's [62:08] gradual but it's an IT task to get it [62:11] right you have to be legible. I don't [62:12] love that word, but you have to have a [62:14] legible set of systems. You have to kind [62:15] of record everything. But I think this [62:18] is the new way. Companies who are doing [62:20] this seem to be just moving a lot faster [62:23] than the companies that aren't. Um, and [62:25] Ivan is moving at breakneck speed. And I [62:29] think there's a lot of things that fall [62:30] out of it. They don't need as many [62:32] people. Um, they move much faster [62:36] because of that. They have much shorter [62:38] planning cycles. All of those kind of [62:40] one-way doors that Bezos used to talk [62:43] about are really two-way doors cuz [62:44] nothing takes that long to do. Anyway, I [62:48] think there's this new playbook. I'm [62:49] interested in it. I'm going to be [62:50] talking about in the podcast and on X. [62:52] If you're interested in it, too. Hit me [62:55] up on Ballagan. Thanks for tuning in.