读完一早上的访谈.
主题综述

创始人模式 (Founder Mode)

主题综述

主题页（活文档）· 最近更新 2026-05-25 · 取材 6 篇访谈

更新日志

主流共识

第一点：层层组织会主动制造"与现实的距离"——这是真实存在的、可识别的病

不同的发言人用不同的名字描述同一个东西：

"Bureaucratic 'game of telephone' that often plagues scaling companies."
「在不断壮大的公司里反复出现的官僚式'传声筒游戏'。」
Brian Chesky · Brian Chesky - AI Founder Mode
"My favorite quote of Toad's is, he once told me that the founder's job is to fight the entropy of ambition."
「我最喜欢的 Toad 语录是——他曾告诉我，创始人的工作就是对抗雄心的熵增。」
Mike Cannon-Brookes · 20VC: Atlassian CEO

Horowitz 把同样的现象换成"创始人怎么坏"的视角——失败的创始人 CEO 共同点不是不聪明，是犹豫；这场对话把由此累积的瘫痪叫作"决策债"：

"The second thing that you see often is just hesitation. 'I don't know what the right answer is. I have a suspicion, I'm like 52-48.' … That lack of confidence, that hesitation — those things are really what caused founders to fail at the CEO job."
「你经常看到的第二件事，就是'犹豫'。'我不知道正确答案是什么，我有个猜测，大概 52 对 48。'……那种不自信、那种犹豫——正是这些让创始人在 CEO 这个岗位上失败。」
Ben Horowitz · Ben Horowitz On What Makes a Great Founder

Garry Tan 把它讲成需要持续做的一件事：

"I view YC as this tree of prosperity. Every organization in the world, no matter how dominant, could use a little bit of pruning. This is a garden, and in order to actually have a good bloom, a good crop coming out of your fruit trees, you've got to do all this pruning months or sometimes in the years in advance of that. It's really hard to say no to things and to people."
「我把 YC 看成一棵繁荣之树。世界上每个组织——不管多强势——都需要一些修剪。这是个园子，要让果树开花结果，你必须提前几个月甚至几年去修剪。对事情和对人说'不'真的很难。」
Garry Tan · Running Y Combinator Like a Founder

第二点：创始人必须用某种方式保持"贴地"——具体的、能审计的、和产品相关的接触

Chesky 的版本最极端：

"Founder mode is about not apologizing for being in the details. During the 2010s, I fell into the trap of over-delegating to professional managers, which led to a bloated, unrecognizable bureaucracy where I felt like I was in a car without a steering wheel."
「Founder mode 的关键是——深入细节不需要道歉。2010 年代我陷入了过度授权给职业经理人的陷阱，结果是一个臃肿的、面目全非的官僚体系，感觉像坐在一辆没有方向盘的车里。」
Brian Chesky · Brian Chesky - AI Founder Mode
"You manage people through the work. You don't manage the people, you manage the work."
「你通过工作来管理人。你不是在管人，你在管工作。」
Brian Chesky · Brian Chesky - AI Founder Mode

Foroughi 的版本看起来相反但落点一样——他从未让自己离开过决策一线：

"I made 10 decisions a day and most of them aren't strategic. If you plan, if you process, if you over communicate, you start over analyzing and slow down. You just plow through decisions, and if you have the capacity to execute on more, it's just more probability that something's going to work out."
「我一天做 10 个决定，大多数都不是战略性的。一旦你做计划、走流程、过度沟通，你就开始过度分析、慢下来。你就只管一路把决定推过去——如果你能多做几个，至少有一个跑通的概率就更大。」
Adam Foroughi · 20VC: AppLovin's Adam Foroughi

第三点：早期招聘必须创始人亲自下场

Chesky 给出了一个相当夸张的具体数字：

"My number one job is hiring, and I spend two to three hours a day on it… I act as the co-hiring manager for the top 200 people in the company. If you can hire someone without my help, we aren't reaching far enough."
「我的头号工作是招聘，每天花 2 到 3 小时。我作为公司前 200 个岗位的共同招聘经理。如果你不需要我介入就能招到这个人——说明我们的目标还不够高。」
Brian Chesky · Brian Chesky - AI Founder Mode

Foroughi 的版本同方向但角度不同——招聘是创始人自身销售能力的测试：

"If you can't sell your idea, your vision and close at that moment, the idea is wrong. The only job early on is to surround yourself with talent because the idea frankly doesn't matter."
「如果你不能推销你的想法、你的愿景、并在那一刻把人 close 下来——那你的想法就有问题。早期唯一的工作就是把自己包围在人才里——因为想法本身其实不重要。」
Adam Foroughi · 20VC: AppLovin's Adam Foroughi

Horowitz 从另一面强调了同一点——招到错的高管会反噬整家公司：

"References really matter. Don't just look for someone enthusiastic; ensure they qualify you. A good sales leader should assess whether the company has the money, the need, and a real shot at success. Also, consider who they are bringing with them, and verify those individuals are committed to joining. Any great sales leader has a following."
「推荐信非常重要。不要只找那些热情的人——要确保对方在评估你。一个好的销售负责人应该评估你的公司有没有钱、有没有真实需求、有没有真正的成功机会。还要看他会带哪些人来——并核实那些人是否真的承诺加入。任何优秀的销售领导者都有一群追随者。」
Ben Horowitz · Ben Horowitz On What Makes a Great Founder

第四点：坏消息要传得快——围绕真相做事，不围绕情绪做事

"I think really good companies, the very, very, very, very best companies tend to have founders and CEOs who ask pretty aggressive questions."
「我认为真正好的公司——非常、非常、非常、非常好的那些——往往有那种问着非常尖锐问题的创始人和 CEO。」
Ben Horowitz · Ben Horowitz On What Makes a Great Founder
"If you're running away from the truth to preserve feelings, that's a very dangerous thing in the tech company."
「如果你为了照顾情绪而逃避真相——那在科技公司里是非常危险的。」
Ben Horowitz · Ben Horowitz On What Makes a Great Founder

Chesky 用一种很温柔的方式说了完全一样的话：

"When I tell somebody it's not good enough, I'm not saying you're not good enough. I'm saying I see potential you don't see in yourself."
「当我告诉一个人'这还不够好'——我不是在说你不够好。我是在说我看到了你自己没看到的潜力。」
Brian Chesky · Brian Chesky - AI Founder Mode

Foroughi 把同一个原则推到了组织运行机制层面：

"We try to hire people who are inherently entrepreneurial and don't have that fear. I'm probably wrong eight times out of 10, but it doesn't matter because I take so many chances. You don't talk about the wrong. It just doesn't matter. It's just fluid and you're moving forward."
「我们尽量雇那种天生有创业精神、没有那种恐惧的人。我大概 10 次有 8 次是错的——但没关系，因为我尝试得太多。你不会聊错的那些。它就是不重要。整个公司就这么流动着往前走。」
Adam Foroughi · 20VC: AppLovin's Adam Foroughi

分歧在哪

跨过共同诊断之后，"怎么治这个病"出现了至少四种截然不同的处方。

阵营 A · "握紧 + 管工作"——Chesky 的纵深穿透派

Chesky 的核心动作不是"信任团队"——而是自己下沉到工作本身：

"You should start hands-on and give ground grudgingly, rather than letting go early and having to spend years unwinding the empire they built in the wrong direction."
「你应该从亲力亲为开始，极不情愿地一点点放权——而不是过早放手、然后花好几年去拆别人朝错误方向建起来的帝国。」
Brian Chesky · Brian Chesky - AI Founder Mode

具体落地是 Project Hawaii：

"The philosophy is to make the problem as small as possible. We took a lean, elite team of 10–12 people and gave them a single, focused problem—like improving guest conversion. We followed a 'crawl, walk, run, fly' system: first fix the bugs, then develop features, then reframe the journey, and finally reinvent the flow… It's the Navy SEAL approach: keep the team small, keep the problem small, and stay on the ground."
「核心理念是把问题尽量做小。我们挑 10 到 12 人的精干小队，丢给他们一个单一聚焦的问题——比如提升 guest conversion。我们按'爬、走、跑、飞'四步走：先修 bug、再做功能、再重构 journey、最后重新发明流程……这是海豹突击队那套：小队、小问题、扎根一线。」
Brian Chesky · Brian Chesky - AI Founder Mode

AI 时代 Chesky 把这套推得更狠——把整个组织变成 manager-IC：

"AI Founder Mode will be even more intense than traditional Founder Mode because you have near-instant access to information. Traditional management relies on a meeting-based culture—I used to do 35 hours of meetings a week just to get information. In the age of AI, I believe we will move toward asynchronous work and significantly fewer layers of management. Every person in the company, including managers, must be a 'manager-IC' (individual contributor) who stays close to the work. If you are a design leader, you must actually design; if you are an engineer, you must code. People who act as pure therapists or meeting-facilitators will not survive this shift because they lack the necessary context to make decisions."
「AI Founder Mode 会比传统 Founder Mode 更激烈——因为你能近乎即时拿到信息。传统管理依赖会议文化——我以前每周开 35 小时会就为了拿信息。AI 时代我相信我们会转向异步工作、显著减少管理层级。公司里每一个人，包括管理者，都必须是'manager-IC'——管理者同时也是个人贡献者，时刻贴近一线。设计负责人必须真的做设计，工程师必须真的写代码。那些只做心理治疗师或会议主持的人会活不过这次转变——因为他们缺少决策需要的上下文。」
Brian Chesky · Brian Chesky - AI Founder Mode

需要注意的是 Chesky 的口吻明显带着自己经历的代价：

"Adulation is like a cup with a hole at the bottom. You keep filling it in, thinking it's love, except it just keeps coming out the bottom."
「外部赞美像是一个底部有洞的杯子。你不断往里倒，以为那是爱——但它一直从底下漏走。」
Brian Chesky · Brian Chesky - AI Founder Mode

阵营 B · "从未握紧 + 信任 + 异步"——Foroughi 的水平扁平派

Foroughi 的路径在表象上跟 Chesky 完全相反——他从来没有经历过"放权→后悔→重新收回"的循环，所以他的"founder mode"看起来是放松管制：

"When I see a similar mindset to me in someone... I'll trust them wholeheartedly. And then at that moment, I don't pay attention to anything they're doing. I hand it off, it's theirs, and I move on to the next."
「当我在某个人身上看到跟我相似的思维方式——我会全心全意地信任他。从那一刻起，我就不再关注他在做什么。我把它交出去，那就是他的了，我去做下一件。」
Adam Foroughi · 20VC: AppLovin's Adam Foroughi
"I don't really operate like that. We don't hold meetings, we don't debate things. If we want to do something, we just go do it. And that helps us move quickly, helps us not overthink things."
「我不太是那样运作的。我们不开会、不辩论。如果我们想做一件事，就直接去做。这帮我们跑得快，也让我们不过度思考。」
Adam Foroughi · 20VC: AppLovin's Adam Foroughi

Foroughi 治理 1000+ 人组织的具体方法是把"管理"的成本压到极低、把"信任"和"决策频率"作为变量推高。他对游戏开发商的整合方式是这个哲学的具体注脚：

"Game developers want to create something great that's used by a lot of people. So if we combine these two parts, we really created a great equation for success by not requiring micromanagement, not requiring integration, but we just take care of growth."
「游戏开发者想做出被很多人用的优秀产品。所以我们把这两块结合起来——不要求 micromanagement、不要求整合、我们只负责增长——就拼出了一个很好的成功方程式。」
Adam Foroughi · 20VC: AppLovin's Adam Foroughi

但 Foroughi 自己在访谈末尾给出了一个很多人忽略的反话——他自己希望终有一天能离开 weeds：

"I would love if I wasn't in the weeds and I just trusted everything to the folks around me and I could go up a level and say, my job is to teach... It's success within those weeds, but it's not a bigger thing. And so eventually, I hope to get back to that."
「我多希望自己不用陷在细节里、可以完全信任周围的人、可以升一个层次说'我的工作是教导'……在细节里你能赢，但那不是更大的事情。所以最终，我希望能回到那种状态。」
Adam Foroughi · 20VC: AppLovin's Adam Foroughi

这一句把"Foroughi 是 founder mode 的反例"这个简单读法复杂化了——他也想脱身，只是经济结构和团队结构允许他先不脱。

阵营 C · "邪教版本警示"——Garry Tan 的清醒派

Tan 看到的是被模仿出来的、走样的 founder mode：

"The cargo cult around great founders is, oh, let me listen less. Actually, you have to listen 10 times more than the other people who don't listen."
「围绕'伟大创始人'的那种 cargo cult（邪教式模仿）是——'哦，那我少听点别人的意见'。恰恰相反——你必须比那些不听的人多听十倍。」
Garry Tan · Running Y Combinator Like a Founder

他把 founder mode 当成一种周期性的整体修剪而不是某一个时刻的英雄主义：

"Culture is everything. When you really get into the weeds on that, it's like fire/hire/promote decisions."
「文化就是一切。真深入到细节里，它就是开人 / 招人 / 提拔的决策。」
Garry Tan · Running Y Combinator Like a Founder
"What's really valuable is to have people around who have seen enough and are intuitive enough about ideas and technology and teams, that they can give you even spiky advice, sometimes like the worst advice. But good founders are like, oh, that's bad advice. I'm just not going to do that."
「真正有价值的是身边有那种见过足够多、对想法/技术/团队足够有直觉的人——他们能给你尖锐的建议，有时是糟糕的建议。但好的创始人会说：'哦，那是个糟糕的建议，我不会那么做。'」
Garry Tan · Running Y Combinator Like a Founder

Tan 把 Chesky 自己描述的事情起了一个名字——"founder mode energy"——但他强调这是特定时刻的状态，不是常态：

"Brian came to our board meetings with Founder Mode Energy before we even knew that it was called Founder Mode Energy, because he had gone through that moment in Airbnb's time where a lot of young founders go through that. Not everyone gets that moment, like COVID was his moment to get it back and reset the Airbnb culture."
「Brian 来 YC 董事会会议时带着'Founder Mode Energy'——那时我们还不知道有这个名字。因为他经历过 Airbnb 的某个时刻——很多年轻创始人都会经历的那个时刻。不是每个人都能撞上那个时刻——COVID 就是他把它要回来、重置 Airbnb 文化的时刻。」
Garry Tan · Running Y Combinator Like a Founder

阵营 D · "认知谦逊是前提"——Chip Conley 的内部观察派

Conley 五十多岁加入 Airbnb 做 Chesky 的导师。他对 founder mode 最有价值的贡献是从内部正面描述了 Chesky 的代价：

"I would never have gone to work for Airbnb if I didn't believe in Brian. I believed in Brian because he showed up with a curiosity and an appetite for learning. What was hard with Brian is that he assumed everybody else was going to work at the same pace and duration as him. His point of view was like, 'Hey, we're having a meeting in the office tonight at 10 o'clock. Be there.' Also, Brian admired Steve Jobs so much that he thought he knew better than anybody else, coming from the product world. The third thing was adding a zero in terms of expectations, or thinking that setting an unreasonable deadline was necessary."
「如果我不相信 Brian，我永远不会去 Airbnb 工作。我相信 Brian，是因为他带着好奇心和学习的胃口。和 Brian 共事困难的地方在于——他假设其他人都能跟他一样的节奏和时长工作。他的视角是'嘿，我们今晚十点在办公室开会，必须到'。另外，Brian 太崇拜 Steve Jobs 了，觉得自己什么都比别人懂——因为他来自产品圈。第三件难事是把期望加个零，或者认为设个不合理的截止日期是必要的。」
Chip Conley · Brian Chesky's secret mentor Chip Conley

Conley 自己作为"导师而非下属"的应对策略是认知谦逊：

"I had to be both wise and curious and often the dumbest person in the room. The key for me to work in that environment and make it work was to not pretend to know things I didn't know."
「我必须同时是有智慧的、好奇的，而且经常是房间里最笨的那个。我能在那种环境里立足并把事情做成——关键就是不假装懂我不懂的东西。」
Chip Conley · Brian Chesky's secret mentor Chip Conley

Conley 对那些被卷进 founder mode 的人给的实操建议异常具体：

"Before the meeting, clarify the intention of the iteration you're doing on the product, what defines success, and what you want to get accomplished in the meeting. Start the meeting with that because it helps ensure alignment. If there's no alignment, you might as well not have the meeting and spend the rest of the meeting talking about alignment. Also, be careful about being overly reliant upon PowerPoint because a founder may take you off path such that your deck in its current order will make no sense at all."
「开会前明确：你这次在产品上的迭代意图是什么、成功的定义是什么、你希望在这场会议里完成什么。会议一开始就讲这些——它有助于对齐目标。如果没对齐，你不如不开会、把剩下的时间用来讨论怎么对齐。另外，不要过度依赖 PPT——因为 founder 可能把你带偏到一个地步，让你的 deck 按原顺序完全说不通。」
Chip Conley · Brian Chesky's secret mentor Chip Conley

都没说透的

我的看法

判断（不是事实）：Chesky 的"管工作不管人"和 Foroughi 的"信任 + 异步"看起来是两种 founder mode，实质是两种不同的组织拓扑，不能直接互相借鉴——Chesky 是垂直纵深（创始人靠 Hawaii 小队一路扎到底审计具体工作），Foroughi 是水平扁平（去掉中层、决策推到一线、创始人只做高频小决策）。Chesky 的打法套到 Foroughi 形状的组织上是 micromanagement；Foroughi 的打法套到 Chesky 形状的组织上是放任。这两套都有效，但前提条件完全不同——Chesky 模型的前提是创始人愿意付 100 小时/周的代价；Foroughi 模型的前提是从未让中层堆积起来。真正危险的是 Garry Tan 警告的第三种——读了 Chesky 文章就开始"少听别人"的模仿者，没有付代价、也没有低层结构，只继承了表面行为。

把握程度：中等偏高。这条判断的最强支撑是 Conley 从内部观察到的 Chesky 代价、加上 Foroughi 自己承认"想脱身"——这两条独立佐证了"founder mode 不是一种永恒理想态、而是依赖具体条件的策略"。最弱的环节是"Tan 的 cargo cult 警告"——它只是 Tan 一个人的观察，但他在 YC 看 800 家公司的位置让这条 anecdotal 但很有信息量。

还想知道什么

取材

继续读
线索
阅读线索：agent 不到 9 nines 可靠，那应该卖给企业还是消费者
线索
阅读线索：如果 coding 是 commodity，AI startup 的护城河长在哪儿
线索
阅读线索：推理成本一年跌 10×，AI app 该按什么定价
线索
摘报
综述
AI Agent 安全与治理
综述
Agent 代购时代：广告与商业的重写