如果所有人都说你错了，有可能这才是你离正确最接近的时候。陆奇在美国的三十年与在中国的三年经历对比，看起来证明了这个道理。
陆奇的职业生涯开始得比多数人都晚，在1998年前往美国卡内基梅隆大学（CMU）进修博士学位之前，他甚至拿不出45美元的入学申请费。他37岁才第一次进入企业工作，以一位普通工程师的身份加入雅虎，用十年做到了雅虎高级执行副总裁，此后加入微软并坐上全球执行副总裁的位子，成为了美国科技行业担任最高管理职位的华人。
三年前陆奇回到中国，他遇到了一系列新问题。一方面，他得到了人们的热爱和广泛赞誉，但另一面，他回国后的两次职业选择都出现了突然的中断或者变更。出于职业道德，他在采访中不愿谈及百度。但那段经历就像一个黑箱子，几乎没人知道真正发生了什么，但人们能感觉到，他遭遇到了挫折。
在一些接触过陆奇的人看来，陆奇似乎有些不一样。比如他会默认一个人说的、想的和做的是一致的；比如他引用YC创始人Paul Graham的一篇文章说，好人和讲道德的人成功概率更大。
陆奇是一个寻求最优解的人，任何一个问题他都有系统的固定方式去解决——就是把它变成一个非常理性化、可以拆解成任务的方程。这些在成熟的商业社会，是行之有效的。但是在中国，在某些方面则不那么有效。
所以陆奇来中国三年，他说自己还在适应、还在理解、还在学习。但考虑到YC中国仅一年，奇绩创坛才刚刚开始，YC美国本身经过了七年才慢慢被大众接受，陆奇还有很长的路要走。
依然有非常多的人看好陆奇，他们希望他成功，但同时，他们又担心他不会成功。这两种想法是矛盾的，又是统一的。理解这一点，恐怕也是陆奇必修的中国课程的一部分。
一件值得去做、长期有价值的事，不被人理解是必然的。
《晚点》：加入YC一年就脱离并成立了新基金奇绩创坛，这次独立是被迫的吗？
陆奇：不是，是YC现在的领导人针对目前大环境做了他认为最理性、最好的判断，我们相互之间也有了最诚意的沟通。这是一个双向选择。
《晚点》：所以你在YC的职业生涯提前结束是因为遇到两个突发事件，一是YC全球CEO的更替，二是国际形势的变化。
陆奇：主要是大的环境。我个人很感激YC，我一定要强调这一点，他们送我们到了他们能送到的最远的地方。
《晚点》：相比于雅虎和微软十年和八年的职业历程，百度和YC这两段经历都是意外开始、仓促结束，你觉得是为什么？
陆奇：YC没有结束，只不过是做了一个调整。现在确实不是我最开始想走的路。当初我想的是，80%精力做YC中国，20%做YC research。现在暂时是100%做投资。
《晚点》：你对人生会做详细的短期、长期规划，但这些规划并没有让你看到中美关系和国际局势的变化？
陆奇：没看到。
《晚点》：如果真正想把一件事做长远，你能完全掌控它是最好的。最开始没有选择自己独立去做，是否意味着你没把它当作长远甚至终身的事业？
陆奇：没有一开始就选择独立是因为更看重YC的全球校友资源，以及，YC可以直接注资过来，我没想到融资这么难、这么重要。但原来也是想做15年，奇绩创坛也一样， 想做15年。
《晚点》：终身不是15年，一个人想得多远要看他是否想到死后的事情，而不是退休后。
陆奇：我认为建立一流企业第一阶段往往需要15年，我可能会在70多岁退休，但之后一定会有更好的人才接班。
对我来说我的人生是分阶段的，我个人的使命是know more、do more、be more，第一阶段是know more，去做学术研究，第二阶段是做得更多，第三阶段体验更多。
这个世纪创造财富最大的机会是创业，但20年30年后，我认为新形式的科研是最能创造财富的，比方说世界需要的二氧化碳固化。创业者越来越多，水平越来越高，但可商业化的技术会很不够；同时，早期创业的本质是用产品满足没有被服务好的市场，但有些市场是倒过来，市场需求在，它需要的是科技，比如老年痴呆症的治疗。
未来会有一种新的组织结构，今天是公司，以后是一个新物种，它既有公司（研发，营销等）、又有投资和科研。这三者组合起来，才可以懂得更多、做得更多、体验更多。
我希望把奇绩创坛先做好，基础夯实，将来再逐步拓宽到科研。我过去没有去创业，因为这是我要探索的一部分，至于我什么时候创业，做什么创业，我还在探索过程中。
《晚点》：VC的规模越来越大，大的创新机会越来越少。奇绩创坛的机会在哪里？
陆奇：对创业者来说钱不是最重要的，能投钱给他们的人很多。重要的是你能给创业者带来的能力提升和有价值的资源。我们的核心价值是手把手在早期帮助创业公司提升价值。我们和辅导一篇博士论文的导师做的工作本质上是一模一样。
《晚点》：李开复说，创业导师这个词，是彼此互斥的两个词，因为真正好的创业者，他自认为不需要导师。
陆奇：有可能是对「创业导师」的定义有不同的想法。我们是手把手一起与创业者解决他们的核心问题。你说做博士论文需要导师吗？麻省理工有存在的价值吗？做博士论文和创业早期本质上是一样的，都要独立解决一个没有被解决过的问题。
《晚点》：过去有哪一位优秀创业者是以你说的这个路径成功的吗？
陆奇：YC在过去15年的成功经验验证了这一点。有很多被YC加速的企业，如AirBnB、Dropbox、Stripe。中国创业生态过去做得很不错，以前基本上不需要基于技术的原创新，但中国的现在和未来需要。抄国外成功的创业公司或改商业模式是无法成为长期驱动力的。为什么世界上需要MIT，清华大学？需要有博导？如果论文都可以自己做的话，世界会很不一样。
对于这个问题我做了严谨的思考，不是没想清楚就去做。希望大家能多关注创业营22家企业的创业者，他们是对这个问题最有发言权的。
《晚点》：博导一次只带一位或几位学生，但你一期就有22家企业，带得过来吗？
陆奇：YC美国两个合伙人加两个兼职合伙人，一期可以带50个人。大家对YC不了解，都以为上上课，不是的，我们手把手一起帮他们解决问题，我们本质上在创业营期间是他们的co-founder——中国的创业生态里没人在做我们做的事。
《晚点》：一些投资人和创业者说，我们尊敬并喜爱陆奇，但我们并不看好他做的事。因为中国已经过了创业周期的红利期了。
陆奇：当初我决定做YC中国，我所有的中国朋友几乎都反对，没有一个人支持。这对我来讲一点不奇怪，某种意义上反而是我的论证。
一件值得去做、长期有价值的事，不被人理解是必然的。如果被所有人理解，你肯定做不大，因为所有人都想做一样的事情，那他们将全是你的竞争对手。
选择一个大平台，我通过它产生的影响会比创业更大
《晚点》：过去有没有一个打动你，让你想放弃一切去创业的机会？
陆奇：2008年从雅虎离开，我想过回中国创业。最后是鲍尔默（微软前首席执行官）说服我加入了微软。如果我去创业，有很多从零到一的工作要做。如果我选择一个大平台，我通过它产生的影响会更大。
《晚点》：在你的三次职业选择中（雅虎、微软、百度），为什么每次选择的都不是当时最新、最有潜力的公司，而是最成熟稳定的大企业？
陆奇：在职业选择上，我主要考虑我是否能在技术驱动创新的前沿工作，学习、提高的同时我也有能力做出有价值的贡献。比如雅虎在早期互联网开拓了很多创新业务，微软在大规模计算平台创新上一直有很大潜力。
《晚点》：看着雅虎一步步走向衰落，为什么没有更早离开？
陆奇：因为我答应杨致远做10年。当时雅虎进入了一个危机，杨致远找我说，中国有一个传统，朋友有难的时候朋友不应该离开。我说那我就不离开了。
《晚点》：听说张一鸣在头条C轮时去美国找过你，希望你能加入头条，而你没有见他。
陆奇：我不知道这件事情，我今天听到了，我很感激。
《晚点》：如果当时见了，会考虑离开微软去头条吗？
陆奇：那个时候可能不会，那时我对头条的认知远没有像今天这样。我和张一鸣有些接触，今天我认为一鸣和头条的机会很大，就像打败搜索的不是搜索引擎，打败Facebook的也不是和Facebook一样的，我认为可能是抖音。
《晚点》：17年你从微软辞职，最主要原因是如外界所说骑车把腿摔断了吗？
陆奇：主要原因确实是这个。另一个因素是想回中国。我很感激比尔盖茨一直挽留，他说，「百度能给你什么，我都给你。」我说，你不能给我「中国」。
《晚点》：据说你在郭台铭家中，郭台铭和李开复劝你，你去创业做什么都行，他们提供资金。他们可以给你中国，你为什么没有选择？
陆奇：因为我认为国民经济总的驱动杠杆是一个大的底层基础和平台。所以我首选是希望找到一个可以让中国的崛起，把一个个企业全部带起来的大平台加入。
《晚点》：回头来看，加入百度是一个错误的选择吗？
陆奇：不好意思，我真的不想多讲百度。
《晚点》：你可以谈微软，但是避免谈百度，为什么？
陆奇：是职业道德。我不想为它带来任何Distraction（干扰）。
比起一个产品的成败，帮助企业建造长期的生命力才是最重要的。
《晚点》：在微软做到了当时华人在美国科技圈最高位置，你觉得为什么是你？
陆奇：我比较幸运，我职业生涯遇到的很多人，他们都信任我、愿意帮我。
《晚点》：是什么让这些人都愿意相信你、帮你？
陆奇：可能是Integrity（正直），每次沟通我都讲真话，如果我答应你什么，我会尽我最大的努力去做，以及有愿意帮助人的心吧。
《晚点》：你如何评价自己的产品和商业能力？比如Bing搜索是你在微软做的主要产品，但很多人会认为Bing并不算一个成功的产品。
陆奇：Bing的价值是为微软建立了一个技术上的能力，把微软创新的基因扩充。过去微软互联网服务基因和机器学习的基因相对来说很弱，有了搜索引擎后，为微软今天的AI开发能力、云的底层能力打了基础。Azure（微软基于云计算的操作系统）核心团队不少都是Bing移过去的。
比起一个产品的成败，帮助企业建造长期的创新的生命力才是最重要的。
《晚点》：你在微软还接手过Skype，Skype如何错失IM这个巨大的市场，你是否有过反思？
陆奇：Skype确实给我一个惨痛教训——千万不要以为产品有很高市占率你就安全了，说不定哪天会被一锅端。
团队当时的文化和目标，过分重视语音交互，认为短信技术上太简单了。我后来复盘的时候才意识到，短信其实是人类历史上认知提高非常重要的一部分。
人类的通信工具，要么是非实时，信息量很多，像书信、邮箱；要么是实时，像电话、Skype。但我没有认知到的一点是——Skype实时的用户协同代价太大。用Skype我们必须事先约好，你几点有空，我几点有空。为了达成一个通话协议，我们可能要送好多其他信息来协同，。因为协同cost是经济cost里最大的因素之一。
短信不需要协同的，你一旦回，我就知道你在线，我们可以马上联系。短信不光可以送文字、图像，还可以送钱、送任何东西。人类历史上从来没有这样一个通信工具，既可以即时，又可以不用付协同代价。这也是我看好微信的原因。当我想明白这一点，再拼命去推，已经来不及了。
《晚点》：协同代价才是Skype失去竞争力的核心原因——你是如何总结出这一点的？
陆奇：跟很多人谈、去思考、看了很多历史。我问客户为什么选择WhatsApp、Zoom而不是Skype。他们会告诉你很多表面原因，谈得多了，你才能抽象出答案。
《晚点》：WhatsApp的开发者并不是因为意识到语音交互的协同代价才去做的短信，他们从一开始就选择了后面这条路，这是运气吗？
陆奇：WhatsApp体验上功夫下得很深，弱网状态下的流畅度做得很好。Skype在这方面忽视了，在弱网状态Skype是很难用的。任何一件事情做得好，既要有运气也要有努力。
《晚点》：微软经历了从传统IT到互联网到移动，从微软的几次跨越中你学到什么？
陆奇：战略的清晰度加执行的纪律性是微软能转型到云和服务的核心。这是过去几年我能观察到、最值得其他公司学习的。
但微软的成功是云转型，主要是商业模式的成功，它还没有做出新的主流产品。这意味着现在的成功是给它带来更多的时间，它将来的成功取决于新的主流产品创新。
《晚点》：回头来看，20年职业生涯中遇到最大的挫折是什么？
陆奇：微软从某种意义上移动生态是全军覆没的，我们没有抓到。
当初微软买诺基亚之前，我推的一条思路是，我认为APP生态不合理，我想推助手，即手机里没有任何APP，只有服务和助手，我们当时尝试和摩托罗拉合作，用他们的硬件＋微软的软件去做，没做成。最后微软还是买了诺基亚。
《晚点》：当你有个强烈的想法但没有被采纳，难道不想离职自己去实现它吗？
陆奇：当时没有这个考虑。不过这件事对我触发很多，我对移动生态总体考虑了不少——获客成本让大部分2C创业机构无法存活。所以微信小程序是可行的，张小龙是我非常敬仰的人。中国移动生态走得很远，很大因素是微信，微信可能是当代做得最出色的一个产品。
如果你想真正大规模改变世界，那你必须是这个企业的创始人，否则你永远受限于你的雇主
《晚点》：以下是一些关于公司战略和组织管理的问题。你被认为是一名高效的管理者，如何激活一家迷茫的公司？你的思考是什么。
陆奇：首先要清晰判断公司方向。方向是三个圆圈组成，技术、需求、市场，永远看三个交集。需求是最难判断的，有哪些被满足，满足之后又有哪些新的机会。比方zoom，做视频交互的公司很多，但zoom对需求观察很深，在美国打电话会议平均延时14分钟，另外视频的质量、体验、复杂度这些需求是不是被真正满足，这是很重要的。
贝索斯去年在股东信中说，客户永远不满足。一个真正好的创业团队要对已经被满足的市场持续观察，技术永远是驱动因素，你要对技术把控，技术会给大家已经被满足的市场又带来新的机会。
《晚点》：公司进行变革，先调整业务还是先调整组织、人？
陆奇：第一步要在战略上做判断，进哪些赛道？出哪些赛道？大部分公司的通病是该出的没出来，然后进的太早。有了进出的决定后你才知道，组织上应该怎么去调整。
但一切前提是这个CEO有足够的能力，以及他的团队足够强。做进和出的决定之前，你要确定是不是有足够好的人来帮你。如果没有，那你首先要把这个人找到。
《晚点》：是单点变革还是全面变革？
陆奇：管理上必须全公司、全方位变革。多数公司第二条增长曲线起不来不是产品问题、不是技术问题，往往是管理问题、内耗问题，原来的功臣都在第一条产品线上，所以CEO必须参与，不然军阀会把第二条曲线杀死。
《晚点》：新的核心方向和旧业务冲突，倾向于激烈替换还是温和改良？
陆奇：不能激烈替换，这是一个全方位的、综合的平衡解决方案。
《晚点》：在大公司做管理，如何做到穿透管理，真正了解一线的情况？
陆奇：我以前要求自己做到的是，我手下两层以下的人，他们每个人的工作我都能做。这样整个公司里，没有人敢欺骗你、忽悠你。这个非常重要，大部分企业中层都在忽悠高层，资源其实是浪费掉的。
《晚点》：一个部门leader，能力极强，价值观极弱，是否还会重用？
陆奇：如果他的价值观和企业价值观完全不吻合，你必须让他走。
《晚点》：很多管理者总是在抱怨人不行，你的建议是什么？
陆奇：需要倒过来想，一件事情如果人跟不上，你要解决的是，第一把自己的认知程度打开，第二把你的操作范围缩小，把权力放出去，让更行的人进来。
《晚点》：你供职过的几任公司CEO性格各异，如何与老板相处，心得是什么？
陆奇：主要是在高压并意见不一致的情况下能有效地沟通和协同。比如我跟盖茨先生花时间很多，他有时比较Demanding（苛刻）,他要做的事情必须马上有一个结果。所以我跟他制定了一个制度——如果我们意见不一致，我们继续讨论下去只有三种可能：
一种你对我错，我会把证据找到证明你是对的我是错的，那我照着你的去做；一种我对你错，我找到证据后我们按照我的方式做；最后一种，是我们对这件事的目标不一致，那这件事情应该在董事会上讨论。
如果和老板之间局面紧张，任何一个问题我都用一个系统性的方法去解决——就是把它变成一个非常理性化的、可以拆解成任务的方式。
《晚点》：上述方法有不灵的时候吗？
陆奇：对我来讲一个很大的挑战是，如果一个人他对你讲的话，和他想的、做的不一致，这样就没法真正共处了。
《晚点》：如何与下属相处？你百度的前同事说，你从来不和他们一起吃饭，你总是一个人在办公室或者酒店房间里吃蔬菜沙拉。
陆奇：形式不重要。我对团队的沟通上，一是保持人人平等的心态，我是你老板只不过我们两个人的岗位不一样。二是当领导第一使命是服务，你要为你手底下的员工服务，因为他们把自己的职业生涯交给你了。
《晚点》：如何找到好的副手，你的标准是什么？
陆奇：有两种做法，第一强的人才，即使没有适合的岗位也要招，我自己以前犯的通病是——等位置和组织设计好再招好的人，但好的人他不会等。
第二，你需要一个好的人才基础，如果找不到人，你就想尽办法把第一个牛人找到。我在雅虎建立研究院不是为了业务，是为了招人，你要用一个机制，它可以是一个技术项目也可以是研究院，吸引第一个牛人，然后就像钓大闸蟹一样，一串就来了。
世界上真正好的人是关心他做什么、他跟谁合作。你想雇一个好的人你至少要花2到3年时间说服他，我叫courtship（求偶）。
《晚点》：一些CEO说，他们苦恼如何打造一个强战斗力的团队，你有什么建议？
陆奇：我认为打不赢胜仗，核心原因往往在产品上、在企业的文化土壤上。以前我在微软做移动产品很难很难，以为把苹果的人搞来就能创新，但苹果的人来了一样做不出来，所以管理者更应该关注土壤环境的问题。
强战斗力我们叫high performance，但这个和high adaptivity（强适应力）是相悖的。强战斗力在固定市场可以冲得很凶，但市场一变他们就不知道怎么变。所以做到适合的平衡才难，战斗力本身不难。
《晚点》：你心中优秀的CEO是什么样子？你的榜样是谁？
陆奇：最强的人愿意为你而来，这是我觉得最好的CEO。在中国我觉得马云做得不错。
《晚点》：你观察到优秀CEO的成长分为哪几个阶段，分别有什么标志事件？
陆奇：Reid Hoffman有一本书叫《Blitzscaling》（闪电式扩充），里面提到，CEO一开始是做家长，公司早期是一个家庭，第二阶段是部落首领，第三阶段是村长，第四阶段是一个市长，第五阶段是一个国家的领袖。
公司小你可以用产品、用技术来控制，大公司只能用价值观、用使命来控制。当然，这几个阶段是可以跨越的。
《晚点》：你的人生中还有难题吗？看上去所有问题你都有方法面对。
陆奇：这取决于你的目标，如果你的目标是想当上帝可能很难，如果你的目标是像我们这种凡人，那所有问题都可以有解。
《晚点》：作为一名成功的职业经理人，对比那些你认为不错的企业家，你认为自己和他们最大的区别是什么？
陆奇：职业经理人和企业家的区别是什么？我想过很多。如果你想真正大规模改变世界，那你必须是这个企业的创始人，否则你永远受限于你的雇主。当这个平台是你亲手建造，那你可以发挥的能量和范围是最充分的。
物可以流、人可以流，不要只盯着高速不受限去做
《晚点》：作为拼多多的技术顾问，你如何理解分布式AI技术发展的前景？
陆奇：我个人对拼多多、对黄铮是非常看好的。我们的起点是一个比较哲学化的讨论——创造大规模价值，是集中式机会多还是分布式的机会更多？我和他在哲理上都更倾向后者。
分布式体系的生态更平稳，集中式的体系危险系数更高。搜索引擎就是一个很典型集中式的体系。但人类社会本质上是分布式，很多高度智慧在某一个大学、某一个企业里。
《晚点》：黄铮曾说，媒体的价值大于拼多多，而拼多多的价值大于人工智能公司。他认为1万个人工智能公司都比不上让五环外的人买到好的手纸重要，你是否理解并认可他的观点？
陆奇：我对黄铮所说的解读是，他所担心的是贫富差距越来越大会造成越来越多的不稳定。因为人工智能公司可能会加速贫富更大的不均匀。
在硅谷，美国很少百分比的人控制了大量的财富，因为他们控制了资本和技术。当我们有了1万家人工智能的公司，这个差距会可能越来越大。
《晚点》：随着人工智能在各个产业上深入渗透，将来人类社会大部分事情机器都可以做，人做什么？
陆奇：我们在思考如何重新设计一个人类社会的体系。YC研究院有一个研究项目就是在basic income下，每一个人对工作态度会发生什么变化。
《晚点》：AI技术逐渐成熟后和传统产业结合，你会选哪一个产业入手？
陆奇：我们总结了一个初步的人工智能的创业方法论：你要有一个很宽的行业场景闭环；必须以数据的支持为核心；必须是开放的、长期要平台化的；数据是多维度的，如果有交互应该以自然交互为主；往往需要软硬件都做。
《晚点》：你最看好的三个AI落地的场景是什么？
陆奇：目前有两个，智能场所、新移动。智能场所，是指任何一个空间（零售店、医院等）都有智能的能力；新移动是机器人的自动驾驶，自动、能走的体系。
《晚点》：以无人驾驶为例，你更看好通用这样的传统汽车公司收购创业公司（Cruise）来做无人驾驶，还是更看好有互联网基因的创业公司（比如Tesla、Waymo）？
陆奇：我认为更有可能成功的是以新一代技术基因为主的创业团队引进汽车工业的基因，而不是倒过来汽车工业引进新的基因。
《晚点》：Waymo、Cruise，特斯拉、Uber,，还有做开放平台的百度、阿里，在无人驾驶的诸多路径中，你最看好哪一种？
陆奇：我不想明确评价。但我可以讲一个纬度——真正L4L5这样的无人驾驶不是一个纯技术问题。
1，真正商业化落地需要基础建设（路网、法规的建设）。2，从工程研发角度，我们还没有发明一个工程开发方法可以可靠地达到车规级的安全要求，车规级的安全要求是故障率在10的8次方——非常非常低，飞机是10的9次方。我认为真正符合车规矩的安全要求，同时在商业化完全落地会是一个漫长过程。
《晚点》：你所观察到现在创业者和投资人对无人驾驶最大的误解是什么？
陆奇：很多人把无人驾驶的商业化形式盯在一辆车可以自己走——大家需要想清楚的是无人驾驶首先是一种能力——建立自主体系。我们应该把无人驾驶作为一个生态、一个技术种类来看待，这意味着它会有很多不同的可商业化形态。
我自己的看法是——商业化的路径上早期低速或者受限是核心，基于这个结论你可以有很多的想像空间。高速、且不受限是目前的挑战，因为它对法规、对技术的要求太高了。物可以流、人可以流，不要只盯着高速不受限去做。
《晚点》：它可以不是车，不是四个轮子。
陆奇：长期，车未必是人类最有效的交通工具，因为你要大规模造路、要基础建设。
如果你去和卡耐基梅隆大学做机器人研究的教授和学生交流，有不少人会跟你讲，在自然界蛇是在能量很低的情况下，在任何的地形都可以走，它靠骨头运动，比有脚的动物要好。这和中国以前想像出来的龙相似。
也许有人在未来可以做一个机器龙一样的东西，将来新一代的运物工具可能既不是车也不是飞机，而是一种新的物。不要把无人驾驶仅仅看成就是要造车，而应该把它看成一种新的能力、一种人类历史上没有过的能力，我们可以建一个自己可以动的体系。
把职业想象成放电影，电影放完你还可以回家、还会有明天、明天还可以再开始
《晚点》：你曾多次提到最喜欢的小说是罗曼罗兰的《约翰・克里斯托夫》，它对你意味着什么？
陆奇：我喜欢这本书喜欢到还手抄过这本书两次。对我来说它是描述人性的一个史诗般的著作。它对我最大的触动是——如果你不去努力、不去克服、不去追求，不去执着地追求，真正好的人性是没法体现出来的。
《晚点》：对你来说什么是「真正好的人性」？
陆奇：人性的美有多个层面，对我这样的人来说可能是去做一个系统。这本书主人翁的雏形是贝多芬，他代表人类克服困难比较好的例子，因为他一生走的路是很艰难的。
我第一次听贝多芬是高中，当时差点没办法考大学，因为这里（喉咙）开刀了，那时我最害怕的是大学不能考了，差点崩溃掉。那时候贝多芬的交响曲是我精神唯一的支持，9个交响曲的主旋律我基本都能背，当时每天靠在自己脑里放贝多芬的音乐活下来的。
英语说「overcome and triumph」（克服困难达到胜利），贝多芬对我而言是人类最好的典型，他的胜利不是某一点的成功，他的一生为人类带来了真正有价值的体验，他的音乐是全人类的财富。
《晚点》：现在焦虑时还听贝多芬吗？
陆奇：现在焦虑基本上就是洗个澡，俗语说退一步海阔天空，这句话对我很有用。
我有一个导师Fred Kofman，他写过两本书，《Conscious Business》（有良知的生意）和《The meaning revolution》（意义的革命）。为什么我们每天觉得焦虑太多，职业上都是障碍，其实是意义太少，我们缺乏meaning。
Fred Kofman提到每个人都要有一个safety harness（安全带），在安全带里可以找到重新启动的方式。你可以把职业想象成放电影，电影放完你还可以回家、还会有明天、明天还可以再开始。
《晚点》：你的签名「Do more, know more, be more」，是否是你人生的meaning？
陆奇：是。任何一个智能生态的信息架构永远是这三部分组成。人追求的永远是想懂得更多、想做得更多、体验更多，三者螺旋上升、永无止境。同时你需要减熵，知识可以帮你减熵，减熵的核心是为了建立更美好的结构。否则世界会混沌一片。
Sam Altman曾经分析YC2000家公司里成功的公司都具备哪些因素，分析结果是创始团队的个人动机非常重要。Airbnb的两个创始人是典型，他们不是为了赚钱，而是想成为最好的自我。
《晚点》：甘地说，be the change you want to see in the world。成为更好的自我和建立一个更好的世界，后者难道不会推动人走得更远吗？
陆奇：自我才会让人更永久。赚很多钱、建设一个成功公司，甚至建造一个国家、世界，都是外在目标。如果一个目标是外在的，你永远会在达到目标之后变得一片空虚。
《晚点》：有一些商人，他们的目标就是赢，他们乐此不疲，似乎并不空虚。
陆奇：你的追求是建立在别人输的基础之上，为什么世界上一定要有人输你才觉得你的生命是有价值的？我认为人的目标是你自己而不是任何外在的因素。
《晚点》：你把世界抽象成一个程序是为了更方便理解还是为了推倒并重新建构？
陆奇：我重视数学，数学是永恒的真理，它可以用符号、结构来描述世界上所有的信息。
我世界观的核心是——世界上所有东西都由能量和信息构成。任何一个企业、一个人都是函数，即把信息从一个状态变成另外一个状态。你明确了它是哪种函数你就知道如何改变它。
《晚点》：微软、百度、YC分别是什么函数？
陆奇：大部分VC是估值函数，计算所投企业价值多少，投进去回报多少。YC不需要估值，只要过了及格线我们就投，所以YC是个计算及格不及格的函数，它格外简单。
微软是一个很复杂的函数，它做了很大的产品矩阵，创造了很复杂的价值。这跟盖茨先生创建公司的思路也有关系，他比较考虑复杂性。搜索公司的核心函数是通过大规模计算，用大规模的数据找到商业模式。
《晚点》：如果万物皆数学，你又要投人，如何把人用函数来理解？
陆奇：我在百度有一个挺好的同事，我们谈论过，如何用量化的方式选人、找团队。他有一个很好的，对人的起步理论从《易经》开始，比如一个团队5个人是不是匹配？金木水火土，你是哪一个生的。其实古代人与人之间的认知很可能比今天的人高，因为今天我们每天都被干扰，很难静下心观察人，而古人总结的可能是对人更深层次的观察。
这是我们的起点，我认为只要累积的数据多了，就可以把选人量化。比如YC投人不看BP（商业计划书），我们只问你18个问题，看你1分钟的视频，我们有打分体系，这也是量化，只是非常粗糙的量化。
《晚点》：奇绩创坛训练营的问卷有一个问题，你有哪些观念是别人不同意但你很坚持的？这个问题你会如何回答？
陆奇：我很喜欢编程，我认为人类之间的任何关系，甚至包括爱与感情，都有它的信息与数学结构「the mathematic of love」但我从没把我的view跟别人讲，因为会有争议。
人需要吃饱，需要休息和睡眠，所以我们有了宾馆、有了家庭。这些我们今天都有有效的办法来解决。
我认为爱也有它自己的数学结构体系，它依然是能量＋信息。假定我们了解爱的信息结构是什么，什么情况下会有爱？程度如何？因为我现在不够理解，所以爱对我来讲大部分时候是随机的。但如果我们能回答这些问题，我们就有机会创造更多的爱。
「我们获取新的认知，起点是问问题，有时候问对问题，你基本上可以根据这个问题走大半辈子了。」
《晚点》：你做过最了不起的事情是什么？
陆奇：我没有做过了不起的事。如果一定要提一个，那其实很不好意思说。我有一个专利，2000年被选上MIT Technology Review 的patent of the year，是一个很简单的idea。
《晚点》：如果你现在刚毕业，or20多岁，你会选择投入怎样的行业，做怎样的事情？
陆奇：还是会选择技术，建立一个底层系统。我对我的女儿说，大部分人在25岁之前，对于人生想要做什么，其实只有模糊的感觉。
最理想的是你加入谷歌5年或者Facebook4年的时候，因为这样的公司给你学习的机会是最多的，你在大公司做了5年之后，基本上财务也独立了。接下来你可以寻找更好的地方，让你可以自由去追求你有热情的东西。
《晚点》：对20、30、40岁的年轻人各有什么建议？
陆奇：20岁需要做让你可以走得很快的事情，快速学、快速失败。30岁你要让自己可以走得远，建立一个核心支撑体系能让你走得很远。这个体系包括你的身体、你厚实的家庭基础和几个志同道合你可以信任的朋友。
一个人到了35岁，到了打造产品的黄金时段，我已前很关注这个年龄阶段的人才，因为他/她已经犯过不少错，他/她最需要做一个好产品让他/她的职业生涯有一个本质提升。
40岁后，理想情况是找到一个可以让你淋漓尽致去发挥的舞台，一个人的才华和一个公司的才华只有在真正被释放的情况下才能实现它的价值。如果这个舞台是你自己的最好。
《晚点》：回头看过去十年，有什么事你会觉得做得太多，有什么事你会觉得做得还不够多？
陆奇：我喜欢写程序，思路有可能会是有一点固化。我做任何事都会想像成一个系统里有界面、有什么操作。好处是让思考更结构化。不好是会简单粗暴把一些细节忽略，特别是我把人的微妙关系简单抽象成一个信息系统，这个未必是对的。
《晚点》：坚持了十几年早上4点起床跑步、工作，就没有想过早上干点别的吗？
陆奇：从没想过。我以前基本是4点起来先弄Email，弄完了以后跑步，跑30分钟，洗个澡，去办公室。跑步第一对呼吸系统是很好的锻炼，第二可以出一身汗，第三是我可以边跑边听书不浪费时间。有时候我会把PPT放在跑步机上翻着看。
我现在不行了，因为骑反向自行车把腿摔伤，腿走平地可以，上下、跑步根本不行。所以目前我对自己的生活效率很不满意，我还没找到一个让我感觉每天都很顺畅的方式。
《晚点》：如何看待冒险？
陆奇：我会有一个框架去冒险试错，基本是把机会成本的8%用来冒险，但8%冒什么险，我会很系统、很理性地来决定。一旦发现方向不对，我会立刻掉头甚至放弃。
《晚点》：你为了练习反向骑自行车把腿摔伤导致不能跑步，这是一个因为冒险带来的不可承受的后果吗？
陆奇：这不是冒险，是我自己不好。因为我觉得自己骑自行车应该没问题，我每天跑步身体挺好的。但是有的时候就是摔得不巧。
《晚点》：你是否经常会太过相信自己的力量？
陆奇：会。我太太一直觉得我太过于自信了。她老觉得我自以为是。
《晚点》：你被认为是一个非常勤奋的人，对你而言，最有效的学习方法是什么？
陆奇：任何智能体系要真正学会知识，这三者都分不开——感知（观察系统）、思考（智力系统）、行动（活动系统）。人、一个企业、一个国家都是这样。
我知道自己去管理企业、去做产品在本性上并不是很适合，我喜欢做系统，写代码做底层系统，永远想关起门来做。但我知道你不去做产品，不去做商业，你不端到端去做，你就无法真正提升自己的知识。
《晚点》：你曾经说，体验是最好的学习方法，所以你会去倒骑自行车。
陆奇：我们获取新的认知，起点是问问题，Sam经常会说下个五年我们还没有问的大问题是什么？有时候问对问题，你基本上可以根据这个问题走大半辈子了。
自行车是体验式学习是最典型的例子，你想学自行车不可能看书学会，要跌、要摔，但神奇的一点是你学会了永远不会忘的。我以前看了很多书，为什么自己去倒骑自行车，就是想把我自己的大脑去重新排一排。
《晚点》：回中国这段时间，有遇到什么和你过去价值体系相悖的事情吗？
陆奇：很多。但我是这样一个人，在接受一个新工作之前，我不设任何预期，没有预期你就没有惊讶了。你就会努力去了解现实、努力去做，我觉得只有这样才行。
《晚点》：你和拼多多创始人黄峥沟通很频繁，他被认为是最懂中国的创业者之一，他对你有什么建议吗？
陆奇：他第一次跟我见面聊天，聊了两三个小时，两个人对于每个问题观点都不一致，但我学了很多，愿意继续谈下去。
《晚点》：你们对中国的理解有什么不一样？
陆奇：很不一样，黄峥说，「Qi，你很美国，你一点都不中国」，他觉得我在很多地方会吃亏，他讲得很对。
和人相处，这三个点很重要——你说什么、你想什么、你做什么，有些文化这三点很近，有些文化这三点很远，中国目前我觉得有点远。如果别人都很远你自己很近，那你就无法适应。
《晚点》：如何调整？是选择不再相信这些人还是选择自己也远一些？
陆奇：要看情况做调整。我要调整自己的认知，调整自己的想法，做事方法要更接地气。我觉得我学习速度还不够快，我获取知识的渠道还不够多，对于中国的创新生态我还有很多地方需要了解。
(by Song Wei, Wu Yang and Yinghui, a willow in the palace)
If everyone says you're wrong, maybe this is the closest you can get to being right. Lu Qi's 30 years in the United States and his three years in China seem to prove this truth.
Lu Qi started his career later than most, and could not even afford an application fee of $45 before he went to study for his doctorate at Carnegie Mellon University in 1998. He joined the company for the first time at the age of 37, joined Yahoo as an ordinary engineer and became a senior executive vice president of Yahoo in 10 years. After that, he joined Microsoft Corp and took the position of global executive vice president. He became a Chinese holding the top management position in the US technology industry.
When Lu Qi returned to China three years ago, he encountered a series of new problems. On the one hand, he was loved and widely praised, but on the other hand, there were sudden interruptions or changes in his two career choices after he returned home. He didn't want to talk about Baidu, Inc. in the interview because of his professional ethics. But the experience was like a black box, and few people knew what really happened, but people could feel that he had encountered a setback.
To some people who have come into contact with Lu Qi, Lu Qi seems a little different. For example, he will acquiesce that what a person says, thinks and does is the same; for example, he quotes an article by Paul Graham, founder of YC, that good people and moral people are more likely to succeed.
Lu Qi is a person who seeks the optimal solution, and he has a systematic and fixed way to solve any problem-that is, to turn it into a very rational equation that can be broken down into tasks. These are effective in a mature business society. But in China, it is not so effective in some ways.
So when Lu Qi came to China for three years, he said he was still adapting, understanding and learning. But considering that YC China is only one year old and the feat has just begun, it took seven years for YC America itself to be accepted by the public. Lu Qi still has a long way to go.
There are still a lot of people optimistic about Lu Qi, they want him to succeed, but at the same time, they are worried that he will not succeed. These two ideas are contradictory and unified. Understanding this is probably part of Lu Qi's compulsory Chinese course.
It is inevitable that something worth doing and valuable for a long time will not be understood.
"later": one year after joining the YC, he broke away from the new fund and set up a new fund. Is this independence forced?
Lu Qi:No, it was the current leader of YC who made what he thought was the most rational and best judgment in the current environment, and we also had the most sincere communication with each other. This is a two-way choice.
"later": so your career at YC ends early because of two emergencies, one is the replacement of YC global CEO, and the other is the change of the international situation.
Lu Qi:Mainly the big environment. Personally, I am very grateful to YC, and I must emphasize this point. They sent us as far as they could.
"later": compared with the ten-and eight-year career experiences of Yahoo and Microsoft Corp, Baidu, Inc. and YC both started by accident and ended hastily. Why do you think?
Lu Qi:YC is not over, it's just an adjustment. This is really not the way I wanted to go at first. At first, I thought that 80% of my energy would be YC China and 20% would be YC research. Now it is 100% investment for the time being.
"later": you will make detailed short-term and long-term plans for your life, but these plans do not show you the changes in Sino-US relations and the international situation?
Lu Qi:Doesn't see it.
"later": if you really want to do something in the long run, it's best if you have complete control over it. Does it mean that you don't regard it as a long-term or even lifelong career if you don't choose to do it on your own at first?
Lu Qi:I didn't choose to be independent in the first place because I paid more attention to YC's global alumni resources, and YC could inject capital directly. I didn't expect financing to be so difficult and important. But it turns out that I also want to do it for 15 years, and I want to do it for 15 years.
"later": life is not 15 years. How far a person thinks depends on whether he thinks about what happens after death, not after retirement.
Lu Qi:I think the first stage of building a first-class enterprise often takes 15 years. I may retire in my 70s, but then there will be better people to take over.
For me, my life is divided into stages, my personal mission is know more, do more, be more, the first stage is know more, to do academic research, the second stage is to do more, the third stage is to experience more.
The greatest opportunity to create wealth in this century is entrepreneurship, but 20 and 30 years later, I think new forms of scientific research will create the most wealth, such as the carbon dioxide solidification the world needs. There are more and more entrepreneurs, the level is getting higher and higher, but the commerciable technology will not be enough; at the same time, the essence of early entrepreneurship is to satisfy the unserved market with products, but some markets are upside down, and the market demand is there. what it needs is technology, such as the treatment of Alzheimer's disease.
There will be a new organizational structure in the future, today is a company, and in the future is a new species, which has both companies (research and development, marketing, etc.), investment and scientific research. Only when these three are combined can we know more, do more, and experience more.
I hope to do a good job in making great achievements, lay a solid foundation, and then gradually expand it to scientific research in the future. I didn't start a business in the past, because it's part of my exploration. As for when and what I start a business, I'm still in the process of exploring.
"later": the scale of VC is getting bigger and bigger, and there are fewer and fewer opportunities for big innovation. Where are the opportunities for miracles to achieve success?
Lu Qi:Money is not the most important thing for entrepreneurs, and there are many people who can invest money to them. What matters is the ability improvement and valuable resources you can bring to entrepreneurs. Our core value is to help startups increase value at an early stage. We do essentially the same job as the mentor who tutored a doctoral thesis.
"later": Kai-fu Lee said that the word entrepreneurial mentor is two mutually exclusive words, because a really good entrepreneur, he thinks he doesn't need a mentor.
Lu Qi:It may be that there are different ideas about the definition of "entrepreneurial mentor". We work hand in hand with entrepreneurs to solve their core problems. Do you need a mentor to do a doctoral thesis? Does MIT have any value in existence? Doing a doctoral thesis is essentially the same as in the early stages of starting a business, both of which have to solve an unsolved problem independently.
"later": has there been any good entrepreneur who has succeeded in this path in the past?
Lu Qi:YC's successful experience over the past 15 years has confirmed this. There are many enterprises accelerated by YC, such as AirBnB, Dropbox, Stripe. China's entrepreneurial ecology has done very well in the past. In the past, there was basically no need for original technology-based innovation, but China needs it now and in the future. Copying successful foreign startups or changing business models cannot be a long-term driving force. Why does the world need MIT, Tsinghua University? Need a doctoral advisor? If the papers could be done by themselves, the world would be very different.
I have done serious thinking about this problem, not to do it without thinking clearly. I hope you can pay more attention to the entrepreneurs of 22 enterprises in the start-up camp, who have the most say in this issue.
"later": the doctoral instructor only brings one or more students at a time, but you have 22 enterprises in the first period. can you bring them here?
Lu Qi:YC has two partners in the United States plus two part-time partners, and can bring 50 people in the first phase. People don't know anything about YC, so they all think that they go to class. No, we help them solve problems hand in hand. In essence, during the start-up camp, no one was doing what we did in their co-founder-- China entrepreneurial ecology.
"later": some investors and entrepreneurs say that we respect and love Lu Qi, but we are not optimistic about what he does. Because China has passed the dividend period of the entrepreneurial cycle.
Lu Qi:When I decided to be YC China, almost all my Chinese friends opposed it and none of them supported it. This is not surprising to me, but in a sense it is my argument.
It is inevitable that something worth doing and valuable for a long time will not be understood. If you are understood by everyone, you can't do much, because everyone wants to do the same thing, then they will all be your competitors.
Choose a large platform, through which I will have a greater impact than starting a business.
"later": has there been one that moved you in the past and made you want to give up all the opportunities to start a business?
Lu Qi:When I left Yahoo in 2008, I thought about going back to China to start a business. Finally, it was Ballmer (former CEO of Microsoft Corp) who persuaded me to join Microsoft Corp. If I start a business, I have a lot of zero-to-one work to do. If I choose a large platform, I will have a greater impact through it.
"later": in your three career choices (Yahoo, Microsoft Corp, Baidu, Inc.), why did you choose the most mature and stable large enterprise instead of the newest and most potential company at that time?
Lu Qi:In terms of career choice, I mainly consider whether I can work at the forefront of technology-driven innovation, and I also have the ability to make valuable contributions while learning and improving. For example, Yahoo developed a lot of innovative business in the early Internet, Microsoft Corp has always had great potential in large-scale computing platform innovation.
"later": watching Yahoo decline step by step, why didn't you leave earlier?
Lu Qi:Because I promised Yang Zhiyuan to do it for 10 years. At that time, Yahoo entered a crisis. Yang Zhiyuan came to me and said that there is a tradition in China that friends should not leave when they are in trouble. I said, then I won't leave.
"later": I heard that Zhang Yiming went to the United States to see you in the headline C round, hoping you could join the headlines, but you didn't see him.
Lu Qi:I don't know about it. I heard it today, and I appreciate it.
"later": if you met at that time, would you consider leaving Microsoft Corp to go to Toutiao?
Lu Qi:Maybe not at that time, when I didn't know the headlines as much as I do today. I have some contact with Zhang Yiming. Today, I think Yiming and Toutiao have a good chance, just like it is not the search engine that defeats the search engine, and the one who defeats Facebook Inc is not the same as Facebook Inc. I think it may be Douyin.
"late": did you resign from Microsoft Corp in 17 years because you broke your leg on your bike, as the outside world said?
Lu Qi:The main reason is indeed this. Another factor is the desire to go back to China. I am very grateful that Bill Gates has been asking to stay. He said, "I will give you whatever Baidu, Inc. can give you." I said, you can't give me 'China'.
"later": it is said that when you are at the home of Terry Gou, Terry Gou and Kaifu Lee advise you to start a business and do whatever you want. They provide funds. They can give you China, why don't you choose?
Lu Qi:Because I think the overall driving lever of the national economy is a big underlying foundation and platform. So my first choice is to find a big platform that can let China rise and bring all the enterprises together to join.
"later": looking back, was it a wrong choice to join Baidu, Inc.?
Lu Qi:I'm sorry, I really don't want to talk about Baidu, Inc..
Later: you can talk about Microsoft Corp, but avoid talking about Baidu, Inc.. Why?
Lu Qi:It's work ethic. I don't want to bring any Distraction to it.
Compared with the success or failure of a product, helping enterprises to build long-term vitality is the most important.
"later": after Microsoft Corp achieved the highest position of Chinese in the American technology circle at that time, why do you think it was you?
Lu Qi:I am lucky to meet a lot of people in my career who trust me and are willing to help me.
Later: what makes these people willing to trust you and help you?
Lu Qi:May be Integrity (integrity), every communication I tell the truth, if I promise you anything, I will try my best to do, and be willing to help others.
"later": how do you evaluate your products and business capabilities? For example, Bing search is the main product you do in Microsoft Corp, but many people will think that Bing is not a successful product.
Lu Qi:The value of Bing is to build a technical ability for Microsoft Corp to expand Microsoft Corp's innovative genes. In the past, Microsoft Corp's Internet service gene and machine learning gene were relatively weak. With a search engine, it laid the foundation for Microsoft Corp's AI development ability and cloud's underlying ability. Many of the core teams of Azure (Microsoft Corp's cloud-based operating system) were moved by Bing.
Compared with the success or failure of a product, it is most important to help enterprises build the vitality of long-term innovation.
"later": you also took over Skype,Skype in Microsoft Corp. Have you ever reflected on how to miss the huge market of IM?
Lu Qi:Skype did teach me a painful lesson-don't think that your product will be safe because it has a high market share, and you may end up in a pot one day.
The team's culture and goals at that time attached too much importance to voice interaction and thought that SMS was technically too simple. When I looked back later, I realized that texting is actually a very important part of cognitive improvement in human history.
Human communication tools are either non-real-time, with a lot of information, such as letters and mailboxes, or real-time, such as telephone and Skype. But what I don't realize is that Skype's real-time user collaboration is too expensive. We must make an appointment in advance to use Skype. What time are you free and what time am I free? In order to reach a call agreement, we may have to send a lot of other information to coordinate. Because collaborative cost is one of the biggest factors in economic cost.
Text messages do not need to cooperate, as soon as you reply, I will know that you are online, we can contact immediately. Text messages can send not only text and images, but also money and anything. There has never been such a communication tool in human history, which can be immediate without paying the cost of collaboration. That's why I'm bullish on Wechat. When I figured this out and pushed it as hard as I could, it was too late.
"later": synergy cost is the core reason why Skype loses its competitiveness-how did you sum up this?
Lu Qi:Talk to a lot of people, think about it, and read a lot of history. I asked customers why they chose WhatsApp and Zoom over Skype. They will tell you a lot of superficial reasons, and only when you talk too much can you abstract the answer.
Later: WhatsApp developers didn't do text messages because they realized the collaborative cost of voice interaction. Was it luck that they chose the latter path in the first place?
Lu Qi:The WhatsApp experience is very hard, and the fluency in the weak network is very good. Skype ignores the fact that Skype is difficult to use in a weak network. It takes both luck and effort to do anything well.
"later": Microsoft Corp has experienced from traditional IT to Internet to mobile. What did you learn from Microsoft Corp's several leaps?
Lu Qi:The clarity of the strategy and the discipline of implementation are at the core of Microsoft Corp's transition to cloud and services. This is what I have observed in the past few years and is the best thing for other companies to learn from.
But Microsoft Corp's success is the cloud transformation, mainly the success of the business model, it has not yet made a new mainstream product. This means that the current success is to bring it more time, and its future success depends on new mainstream product innovation.
Later: looking back, what was the biggest setback in your 20-year career?
Lu Qi:In a sense, Microsoft Corp's mobile ecology was completely wiped out, and we didn't catch it.
Before Microsoft Corp bought Nokia, one of the ideas I pushed was that I thought APP was ecologically unreasonable. I wanted to push an assistant, that is, there was no APP in the phone, only services and assistants. We tried to cooperate with Motorola and use their hardware + Microsoft Corp's software to do it, but didn't do it. In the end, Microsoft Corp bought Nokia.
"later": when you have a strong idea but are not adopted, don't you want to leave your job to achieve it yourself?
Lu Qi:There was no such consideration at that time. But it triggered a lot for me, and I thought a lot about the mobile ecology as a whole-the cost of getting customers made it impossible for most 2C startups to survive. So WeChat Mini Programs is feasible. Zhang Xiaolong is a person I admire very much. China Mobile Limited ecology has gone far, a big factor is Wechat, Wechat is probably one of the best products of the contemporary era.
If you want to really change the world on a large scale, you must be the founder of this business, or you will always be limited by your employer.
"later": here are some questions about the company's strategy and organizational management. You are considered to be an efficient manager, how to activate a confused company? What are you thinking.
Lu Qi:First of all, we should clearly judge the direction of the company. The direction is composed of three circles, technology, demand, market, always look at the intersection of three. Needs are the most difficult to judge, which are met and what new opportunities there are after they are met. For example, zoom, there are many companies that do video interaction, but zoom has a deep observation on the needs, with an average delay of 14 minutes for phone calls in the United States. in addition, it is important whether the requirements such as video quality, experience and complexity are really met.
'customers are never satisfied, 'Mr. Bezos said in a shareholder letter last year. A really good entrepreneurial team should continue to observe the satisfied market, technology is always the driving factor, you have to control the technology, technology will bring new opportunities to the market that has been satisfied.
"later": to change the company, first adjust the business or adjust the organization and people first?
Lu Qi:The first step is to make a strategic judgment. Which tracks should you enter? Which tracks are there? The common problem of most companies is that they should not come out, and then enter too early. Only after you have made the decision to go in and out will you know how to adjust the organization.
But only if the CEO has enough ability and his team is strong enough. Before you make decisions about going in and out, you need to make sure that there are people good enough to help you. If not, you have to find this person first.
"later": is it a single point of change or a comprehensive change?
Lu Qi:Management must be reformed in the whole company and in all directions. The failure of the second growth curve of most companies is not a product problem, not a technical problem, but a management problem and internal friction problem. The original contributors are all on the first product line, so CEO must participate, otherwise the warlords will kill the second curve.
"late": the new core direction conflicts with the old business, which tends to be replaced violently or moderately improved?
Lu Qi:Cannot be drastically replaced, this is an omni-directional, comprehensive and balanced solution.
"later": in the management of a large company, how to penetrate the management and really understand the front-line situation?
Lu Qi:What I used to ask myself to do is that I can do the work of each of the people under two floors under me. In this way, no one in the whole company dares to deceive you or fool you. This is very important, most of the middle level of enterprises are bluffing, resources are actually wasted.
"later": a department leader, the ability is extremely strong, the values are extremely weak, will it still be reused?
Lu Qi:If his values are completely inconsistent with corporate values, you have to let him go.
"later": many managers are always complaining that people can't do it. What's your advice?
Lu Qi:You need to think about it the other way around. If people can't keep up with you, what you need to solve is, first, to open up your awareness, second, to narrow your scope of operation, to let out your power, and to let better people in.
"later": the CEO of several companies you have worked for have different personalities. How do you get along with your boss? what do you learn?
Lu Qi:Mainly in the high pressure and disagreement in the case of effective communication and coordination. For example, Mr. Gates and I spend a lot of time, he is sometimes more Demanding (harsh), what he has to do must have an immediate result. So I worked out a system with him-- if we don't agree, there are only three possibilities for us to continue the discussion:
One is that you are right and I am wrong, and I will find the evidence to prove that you are right and I am wrong, then I will do it your way; one is that I am right and you are wrong, and when I find the evidence, we will do it my way; and finally, if we do not agree on the goal of this matter, then the matter should be discussed at the board of directors.
If there is tension with my boss, I will solve any problem in a systematic way-that is, turn it into a very rational way that can be broken down into tasks.
"later": is there any time when the above methods don't work?
Lu Qi:A big challenge for me is that if what a person says to you is not consistent with what he thinks and does, he can't really live together.
"later": how to get along with subordinates? Your former colleague of Baidu, Inc. said, you never eat with them. You always eat Vegetables Salad alone in your office or hotel room.
Lu Qi:Form doesn't matter. I communicate with the team, one is to maintain the mentality of equality, I am your boss, but the positions of the two of us are different. Second, when the first mission of leadership is to serve, you should serve the employees under your hands, because they give you their career.
"later": how to find a good deputy, what is your standard?
Lu Qi:There are two ways to recruit the top talent, even if there is no suitable position. The common fault I made in the past is to wait for the position and organization to design and then recruit good people, but he won't wait for the good ones.
Second, you need a good talent base. If you can't find someone, you will try your best to find the first awesome person. I set up a research institute at Yahoo not for business, but to recruit people. You have to use a mechanism, which can be a technical project or a research institute, attracting the first awesome people, and then, like fishing for hairy crabs, a bunch of them will come.
The really good person in the world is concerned about what he does and who he cooperates with. If you want to hire a good person, it will take you at least two to three years to convince him. My name is courtship.
"later": some CEO say they are worried about how to build a strong fighting team. do you have any suggestions?
Lu Qi:I think the core reason for not winning the battle often lies in the products and in the cultural soil of the enterprise. In the past, it was very difficult for me to make mobile products in Microsoft Corp. I thought that if I got Apple Inc's people, I could innovate, but Apple Inc's people couldn't do it, so managers should pay more attention to the soil and environment.
Strong combat effectiveness we call high performance, but this is contrary to high adaptivity (strong adaptability). Strong combat effectiveness can be fierce in a fixed market, but when the market changes, they don't know how to change. Therefore, it is difficult to achieve the right balance, and the combat effectiveness itself is not difficult.
"later": what is the excellent CEO like in your heart? Who is your role model?
Lu Qi:The strongest person is willing to come for you, this is the best CEO in my opinion. I think Jack Ma did a good job in China.
"later": what stages have you observed in the growth of an excellent CEO, and what are the hallmarks?
Lu Qi:Reid Hoffman has a book called Blitzscaling (Lightning expansion), which mentions that CEO started out as a parent, the early days of the company as a family, the second stage as a tribal leader, the third stage as a village head, the fourth stage as a mayor, and the fifth stage as a national leader.
Small companies can be controlled by products and technology, while large companies can only be controlled by values and missions. Of course, these stages can be crossed.
"later": are there any problems in your life? It seems that you have a way to deal with all the problems.
Lu Qi:It depends on your goal, if your goal is to be God, it may be difficult, if your goal is to be mortal like us, then all problems can be solved.
"later": as a successful professional manager, what do you think is the biggest difference between yourself and the entrepreneurs you think are good?
Lu Qi:What is the difference between a professional manager and an entrepreneur? I've thought about it a lot. If you want to really change the world on a large scale, you must be the founder of this business, or you will always be limited by your employer. When the platform is built with your own hands, the energy and scope you can exert is the fullest.
Things can flow, people can flow, don't just stare at the highway and do it without restriction.
"later": as a technical consultant, how do you understand the prospects for the development of distributed AI technology?
Lu Qi:Personally, I am very optimistic about Pinduoduo and Huang Zheng. Our starting point is a more philosophical discussion-is there more centralized or distributed opportunities to create large-scale value? He and I both prefer the latter philosophically.
The ecology of the distributed system is more stable, and the risk coefficient of the centralized system is higher. Search engine is a typical centralized system. But human society is distributed in nature, with a lot of high wisdom in a university or an enterprise.
"later": Huang Zheng once said that the value of media is greater than that of Pinduoduo, while the value of Pinduoduo is greater than that of artificial intelligence companies. He thinks that 10,000 artificial intelligence companies are not as important as allowing people outside the Fifth Ring Road to buy good toilet paper. Do you understand and agree with him?
Lu Qi:My interpretation of what Huang Zheng said is that he is worried that the growing gap between the rich and the poor will cause more and more instability. Because artificial intelligence companies may accelerate greater inequality between the rich and the poor.
In Silicon Valley, a small percentage of Americans control a lot of wealth because they control capital and technology. When we have 10,000 artificial intelligence companies, the gap is likely to get bigger and bigger.
"later": with the in-depth penetration of artificial intelligence in various industries, machines can do most things in human society in the future. What can people do?
Lu Qi:We are thinking about how to redesign the system of human society. The YC Research Institute has a research project on what happens to everyone's attitude towards work under basic income.
"later": when AI technology is gradually mature and combined with traditional industries, which industry will you choose to start?
Lu Qi:We summarize a preliminary entrepreneurial methodology of artificial intelligence: if you want to have a wide closed loop of industry scene, you must take the support of data as the core, it must be open and long-term platform, and the data is multi-dimensional. if there is interaction should be based on natural interaction; often need to do both hardware and software.
"later": what are your favorite three AI landing scenes?
Lu Qi:At present, there are two, smart places and new mobile. Intelligent place, refers to any space (retail stores, hospitals, etc.) have intelligent capabilities; the new mobile is the robot's autopilot, automatic, walking system.
"later": take self-driving as an example, do you prefer traditional car companies like GM to buy start-ups (Cruise) for self-driving or start-ups with Internet genes (such as Tesla, Waymo)?
Lu Qi:I think it is more likely that the entrepreneurial team based on a new generation of technological genes will introduce genes from the automobile industry, rather than introducing new genes from the auto industry in reverse.
"later": Waymo, Cruise, Tesla, Inc., Uber, and Baidu, Inc. and BABA, who are working on open platforms, among the many self-driving paths, which one do you like best?
Lu Qi:I don't want to judge clearly. But I can talk about one latitude-self-driving like the real L4L5 is not a purely technical problem.
1, the real commercial landing needs infrastructure (road network, laws and regulations construction). 2, from the point of view of engineering research and development, we have not invented an engineering development method that can reliably meet the safety requirements of the vehicle gauge level, the safety requirement of the vehicle gauge level is that the failure rate is 10 to the 8th power-- very low, and the aircraft is 10 to the 9th power. I think it will be a long process to really meet the safety requirements of the car rules and complete commercial landing at the same time.
"later": what are the biggest misunderstandings about self-driving among entrepreneurs and investors?
Lu Qi:Many people focus on the commercial form of self-driving that a car can walk on its own-- what you need to think about is that self-driving is first and foremost an ability-- to establish an autonomous system. We should treat self-driving as an ecology and a type of technology, which means that it will have many different commercialized forms.
My own view is that early low speed or limitation is the core of the path of commercialization, and based on this conclusion, you can have a lot of room for imagination. High-speed and unrestricted is the current challenge, because it has too high requirements for regulations and technology. Things can flow, people can flow, do not just stare at the highway unrestricted to do.
"later": it can not be a car, not four wheels.
Lu Qi:In the long run, cars may not be the most effective means of transportation for human beings, because you have to build roads and infrastructure on a large scale.
If you talk to professors and students of robotics at Carnegie Mellon University, many people will tell you that in nature, snakes can walk on any terrain with very low energy, and they move on bones, which is better than animals with feet. This is similar to the dragon imagined by China before.
Maybe someone can make something like a robot dragon in the future, and in the future, the new generation of tools may be neither a car nor an airplane, but a new thing. Don't think of self-driving just as building a car, but as a new ability, an ability that has never been seen in human history, and we can build a system that we can move.
Think of your career as showing a movie. You can go home after the movie, and you can start again tomorrow.
Later: you have mentioned many times that your favorite novel is Roman Roland's John Christopher. What does it mean to you?
Lu Qi:I like this book so much that I copied it twice. To me, it is an epic work describing human nature. Its biggest touch to me is that if you do not work hard, do not overcome, do not pursue, do not persistently pursue, the really good human nature can not be reflected.
"later": what is "really good human nature" for you?
Lu Qi:The beauty of human nature has many levels, and for people like me, it may be to make a system. The embryonic form of the hero of this book is Beethoven, who represents a better example of human beings in overcoming difficulties, because the road he has taken in his life is very difficult.
The first time I heard that Beethoven was in high school, I almost couldn't go to college because there was an operation here. At that time, what I was most afraid of was that I couldn't take the college exam and almost collapsed. At that time, Beethoven's symphonies were my only spiritual support. I could basically recite the main theme of 9 symphonies. I survived by playing Beethoven's music in my head every day.
Say "overcome and triumph" in English. Beethoven is the best example of mankind to me. His victory is not a success at some point. His life has brought a truly valuable experience to mankind, and his music is the wealth of all mankind.
Later: do you still listen to Beethoven when you are anxious?
Lu Qi:Now anxiety is basically taking a bath. As the saying goes, it is very useful to me to take a step back.
I have a mentor, Fred Kofman, who has written two books, Conscious Business (Business with conscience) and The meaning revolution (Revolution of meaning). Why do we feel anxious too much every day, professional obstacles, in fact, too little meaning, we lack meaning.
Fred Kofman mentioned that everyone should have a safety harness (seat belt), in which you can find a way to restart. You can think of your career as showing a movie, and you can go home after the movie, and you can start again tomorrow.
"later": is your signature "Do more, know more, be more" the meaning of your life?
Lu Qi:Yes. The information architecture of any intelligent ecology is always made up of these three parts. What people pursue is always want to know more, want to do more, experience more, the three spiral rise, never-ending. At the same time, you need to reduce entropy, knowledge can help you reduce entropy, the core of entropy reduction is to build a better structure. Or the world will be in chaos.
Sam Altman has analyzed the factors of successful companies in YC2000 companies, and it is important that the analysis results are the personal motivation of the founding team. The two founders of Airbnb are typical, not to make money, but to be the best of themselves.
"later": Gandhi said, be the change you want to see in the world. Wouldn't the latter push people to go further to become a better self and build a better world?
Lu Qi:The self makes people more permanent. Making a lot of money, building a successful company, and even building a country and the world are all external goals. If a goal is external, you will always become empty after you reach it.
"later": there are some businessmen whose goal is to win. They enjoy it and don't seem to be empty.
Lu Qi:Your pursuit is based on the fact that others lose. Why does someone in the world have to lose before you think your life is valuable? I think the goal of a person is yourself, not any external factor.
Later: do you abstract the world into a program to make it easier to understand or to knock it down and reconstruct it?
Lu Qi:I attach importance to mathematics, mathematics is an eternal truth, it can use symbols and structures to describe all the information in the world.
The core of my worldview is that everything in the world is made up of energy and information. Any enterprise or person is a function, that is, changing information from one state to another. You know what kind of function it is and you know how to change it.
"later": what functions are Microsoft Corp, Baidu, Inc. and YC respectively?
Lu Qi:Most VC is a valuation function, which calculates the value of the invested enterprise and the return on the investment. YC does not need a valuation, as long as we pass the pass line, we will vote, so YC is a function to calculate whether to pass or not, it is very simple.
Microsoft Corp is a very complex function, it makes a large product matrix and creates a very complex value. It also has something to do with the way Mr. Gates started the company. He is more concerned about complexity. The core function of search companies is to find business models with large-scale data through large-scale computing.
"later": if everything is mathematical and you want to vote for people, how do you understand people in terms of functions?
Lu Qi:I have a good colleague in Baidu, Inc.. We talked about how to select people and find teams in a quantitative way. He has a good starting theory on people, starting with the Book of changes, such as whether five people in a team match? Gold, wood, water, fire and earth, which one were you born? In fact, the understanding between people in ancient times is probably higher than that of people today, because today we are disturbed every day, it is very difficult to calm down and observe people, and the ancients may sum up a deeper observation of people.
This is our starting point. I think as long as we accumulate more data, we can quantify the selection of people. For example, YC does not watch BP (business plan), we only ask you 18 questions, watch your one-minute video, we have a scoring system, this is also quantification, just a very rough quantification.
"later": there is a question in the questionnaire of Qiji Chuangtan training Camp. What ideas do you have that others disagree with but which you insist on? How would you answer this question?
Lu Qi:I like programming very much, and I think any human relationship, even love and affection, has its information and mathematical structure "the mathematic of love", but I never tell anyone my view because it will be controversial.
People need to have enough to eat, rest and sleep, so we have hotels and families. We have effective ways to solve all these problems today.
I think love also has its own mathematical structure, it is still energy + information. Suppose we understand what the information structure of love is, and under what circumstances will there be love? To what extent? Because I don't understand it now, love is random for me most of the time. But if we can answer these questions, we have a chance to create more love.
"when we gain new knowledge, the starting point is to ask questions, and sometimes when we ask the right questions, you can basically walk based on this question for most of your life. "
Later: what's the most amazing thing you've ever done?
Lu Qi:I haven't done anything great. If I have to mention one, I'm actually very embarrassed to say it. I have a patent, which was selected as MIT Technology Review's patent of the year in 2000, which is a very simple idea.
"later": if you just graduated now, how old is or20, what kind of industry will you choose to join and what kind of things will you do?
Lu Qi:Or will choose the technology to build an underlying system. I told my daughter that most people have only a vague sense of what they want to do in life before the age of 25.
Ideally, when you join Alphabet Inc-CL C for 5 years or Facebook4 years, because such a company gives you the most opportunities to learn, after 5 years in a big company, you are basically financially independent. Then you can find a better place, so that you are free to pursue what you are passionate about.
"later": what advice do you have for young people aged 20, 30 and 40?
Lu Qi:At the age of 20, you need to do something that allows you to walk fast, learn quickly and fail quickly. At the age of 30, you need to allow yourself to go far, and establishing a core support system will allow you to go far. The system includes your body, your solid family foundation and a few like-minded friends you can trust.
When a person reaches the age of 35, when it comes to the prime time of building a product, I have paid close attention to the talent at this age, because he / she has already made a lot of mistakes. He / she most needs to make a good product to make an essential improvement in his / her career.
After the age of 40, the ideal situation is to find a stage where you can make the most of it. A person's talent and a company's talent can realize its value only if it is really released. If this stage is your own best.
"later": looking back on the past ten years, what do you think you have done too much, and what do you think you haven't done enough?
Lu Qi:I like to write programs, and my thinking may be a little fixed. I imagine everything I do as an interface and operation in a system. The advantage is to make thinking more structured. It is not good to simply and rudely ignore some details, especially when I simply abstract the delicate relationship of people into an information system, which is not necessarily true.
"later": after getting up at 4: 00 in the morning for more than ten years to run and work, have you ever thought of doing something else in the morning?
Lu Qi:Never thought about it. I used to get up at 4: 00 to do the Email, then run, run for 30 minutes, take a shower and go to the office. The first pair of running is good exercise for the respiratory system, the second is that I can sweat, and the third is that I can run and listen to books without wasting time. Sometimes I would put the PPT on the treadmill and look at it.
I can't do it now, because I hurt my leg by riding a reverse bike. I can walk flat, but I can't go up and down or run at all. So at present, I am not satisfied with the efficiency of my life, and I haven't found a way to make me feel smooth every day.
"later": what do you think of adventure?
Lu Qi:I will have a framework to risk trial and error, basically 8% of the opportunity cost to risk, but 8% of the risk, I will decide systematically and rationally. Once I find that the direction is wrong, I will turn around and even give up immediately.
"later": is it an unbearable consequence of risk-taking that you can't run because you broke your leg to practice riding your bike in the opposite direction?
Lu Qi:It's not an adventure. It's my own fault. Because I think it should be no problem to ride a bike, I run every day in good health. But sometimes it's just a bad fall.
"later": do you often put too much faith in your own power?
Lu Qi:Yes. My wife always thought I was too confident. She always thinks I'm self-righteous.
"later": you are considered to be a very diligent person. What is the most effective way for you to learn?
Lu Qi:Any intelligent system to really learn knowledge, these three are inseparable-perception (observation system), thinking (intelligence system), action (activity system). This is true for people, an enterprise and a country.
I know that it is not naturally suitable for me to manage the enterprise and make products. I like to do systems, write code to do underlying systems, and always want to do it behind closed doors. But I know that if you don't make products, you don't do business, if you don't do it, you can't really improve your knowledge.
"later": you once said that experience is the best way to learn, so you will ride your bike backwards.
Lu Qi:When we gain new knowledge, we start by asking questions, and Sam often says, what are the big questions we haven't asked in the next five years? Sometimes when you ask the right question, you can basically walk on it for most of your life.
Bicycle is the most typical example of experiential learning. If you want to learn bicycles, you can't read and learn. You have to fall and fall, but the magic is that you will never forget what you have learned. I used to read a lot of books. Why do I ride my bike backwards? I just want to rearrange my brain.
"later": have you encountered anything contrary to your past value system during your return to China?
Lu Qi:A lot. But I am such a person that you won't be surprised if I don't set any expectations before accepting a new job. You will try to understand the reality and try to do it. I think it's the only way.
"later": you communicate frequently with Pinduoduo founder Huang Zheng, who is considered to be one of the entrepreneurs who understand China best. does he have any suggestions for you?
Lu Qi:He met and chatted with me for the first time for two or three hours, and the two people had different views on each issue, but I learned a lot and was willing to continue the conversation.
"later": what is your different understanding of China?
Lu Qi:Very different, Huang Zheng said, "Qi, you are very American, you are not Chinese at all." he thinks I will suffer in many places, and he is right.
These three points are very important to get along with people-what you say, what you think, what you do, some cultures are very close, and some cultures are very far away. I think China is a little far away at present. If everyone else is far away and you are close to yourself, you can't get used to it.
"later": how to adjust? Do you choose not to trust these people or to be farther away from yourself?
Lu Qi:It depends on the situation to make adjustments. I want to adjust my cognition, adjust my thoughts, and do things in a more approachable way. I don't think my learning speed is fast enough, I don't have enough channels to acquire knowledge, and I still need to know a lot about China's innovation ecology.
