126. 和红杉郑庆生聊：经济史的流量革命、人类行为模式的不可预期，与创始人性格

What is the most important thing about the world? When I thought about all the human beings' history, I thought it was possible to find the same thing as a point. Every human being is not the same thing as a point. But I chose the same thing as a point of view. I think that all the human beings' history is a certain way of a certain extent. You can call it the same way of the same way of the same way. You can call it the same way of the same way of the same way. It's just like a history. In 1951, SethINA and Nature Town were very england nochmal to which one of its 2 million ! The largest consumer-based products are called 연�ались for some countries in the store that has Lightroom, they have a lot of DAU You may find that all the massive organizations Cross-互联�w to theoned brand You can also see that huge supplyLevel market the. The mass of information�� are owned by us. We kind of get used to answer different orders, but we think that it can serve as things such as human technology. This really means that we bring the human life to have new existing learning systems, and my son says again in M睦直iathe needs. ... ... Hey we qui "... ... ... ... ... ... ... ... ... ... ... ... ... Chapter SECRETARY KINNISH years from 2005 in offshore hostingullah mac 最近几年我们是聚焦在AI,包括大模型,agent,还有AI相关那些智能硬件。 在这个开始之前,我也想插入一个广告,就是因为那个红杉其实一直非常非常重视早期的投资, And we're all in our Lavched feed Actually in our we're the project That should be our We're the reste of the big companies We're the number of these We could'll open to you On this one We're as a legal Even before we talk to洪山 How does it definitely feel like it's an efficient source from the first place Usually on those Is since Geless 2005 因为你的投资生涯非常的长 你的职业生涯恰好是中国风头开始的时候 是2005年然后在此之前你还做过审计 在过IBM 在过BIP管理咨询 然后后来又到盛大做战头 之后到了智信资本和红杉 能不能给大家完整的讲讲 你的整个职业生涯 你是怎么走上投资这条路的 对 这其实挺有意思 我这里想就是分享我一个想法 就是说因为我之前其实做过不少的职业 有过一些体会 but it made me think like I was a very king and I responded so I'm in 1984 do so in� of course many giornate are at O in class Fly T first already No 什么变成水平啊就是我就抄那本书上有一些有意思的这个例子那特别是抄一些比如说一些什么跳青蛙的一些游戏那时候我觉得很有意思就是现在做了人工智能呢就我回头看那时候我当时还去编有一种叫人机对话但那时候人机对话实际上就是你想一些小窍门比如说你出现哪个词我就会怎么回答你你出现另外一些词比如有一些否定我会怎么回答你 It's not any of the sense of the knowledge But it's very interesting Because you create a small system And you say some simple words Or you say some simple words Or you say it with your colleagues It's very interesting This is how you're doing I was just on a school I was just on a school I was just on a school But I was just not I was not a My理工科天赋 is not good So I was also a good teacher So I was also a good teacher So I was also a good teacher But I was just not going to continue But I was just saying This is actually 其实对我还是有一个比较大的影响 然后后来我是进到大学 这大学我学经济学 然后那几年正好是因为我是上大学的时候 就正好是中国刚刚加入这个 刚刚应该是就加入当时还叫官方总协定 然后后来被WTO取代 所以当时是正在做那个谈判 整个谈判过程我印象看新闻印象很深 但是的确我今天回头看就是说我的主要的这个求学工作生涯其实生活在中国加入WTO之后的这些年 然后包括这个创业所以我也觉得很就很有意思 我依旧很清晰的记得那段时间 所以这个你看也是另外一个怎么说呢 就是可能我工作生涯的一个一个史丹朗印吧 再往后有几件事情其实后面也是对我还是有影响 当时是我第一次上网是96年 96年当时因为我在复旦 当时复旦举办了一个教大家怎么上网的一个讲座 然后差不多从那时候开始复旦有了这个校园BBS 然后我一度还是我们经济学院的版主 那时候开始就频繁的去到机房上网 那时候到机房上网 对 然后到98年的时候 另外一件事情我记得很清晰就是我的一个同学当时我同学是理科的学光学然后他就跟我说说有一个网站特别好用比Yahoo还好用因为Yahoo当时只能大头是做分类分类所以然后他当时说这个网站可以帮你搜索然后当时我就第一次用了Google然后那时候开始我就开始广泛的使用这些互联网的产品后来很快国内的这些互联网产品也都在出来从这个像百度新浪 网页这些就慢慢都出来了就是当时我对混产网产品的一个非常频繁的使用实际上构成我当时生活的一个主要方式所以你上大学的时候刚好是人类开始触网对 其实上了大学对你看你很早学编程然后又很喜欢上网你为什么没有想过要去做一个产品经理或者做一个公司做产品为什么没有成为你的选择 因为我们相当于应该是上个世纪最后一波大学毕业生或是久远的大学毕业 我认为整个在这个工作的在找工作这个环境当中 可能还是更倾向于去找一些传统的财经类的专业人士的工作 当时并没有 但当时其实在美国互联网泡沫已经起来了 就当时这个纳斯拉克指数也都很高了 00年前后对,就00年前后所以当时是很想但的确没有所以最后还是选择了中规中矩就是进外企做了审计但是当时我几乎会试用所有的互联网产品就我可能比一个互联网的从业的人员可能用互联网产品还用的更多所以这是当时的一个生活的一个现状对于那时候的毕业生来说是不是去互联网公司的人非常少对,因为当时互联网公司本身也不多也没有兴起特别是在国内对 但当时已经起来了当时几乎一线城市很多人已经开始可以去当时的一线大厂就是这些门户公司去工作已经开始有了对于你们来说去金融公司还是一个更好的选项对 当时普遍的应该是这样的那你做的审计期间就是在你前面几份工作中间有什么有趣的事情吗因为做审计其实对我的影响是很深远的因为我这个人 本身的文科背景会比较重 所以说对于数字 对于很多非常有逻辑 非常有闭环的事情 实际上当时的关注度是不够的 所以审计 反正我每次跟我的朋友 就是在我性格方面 是非常深远地影响了我 如果没有审计那段时间的培养 可能之后做投资要费力非常多 它应该变得更理性 当你思考一个问题的时候 因为审计它是一个很严肃的 一个工作 所以它会把 所有的这个证据 就像写论文 你要有各种证据 最后你要检查证据 然后你还要把所有 当时我们叫 这个cross reference 你要把所有的 这个参考 指向 然后全部把它变成 一个闭环 所以这个事情对我是有一个 很大的影响 而且当时 我觉得所以我刚才说 好像一切都有点 明明中都有是安排 就当时我还从事另外一个 跟审计业务有关 但是其实当时在审计里面不主流的 就是这是一个合规性的一个检查 这个合规性检查是干什么呢 就是当时又回到刚才主题 因为加入到WTO之后呢 很多这个全球的名牌 在中国去生产 生产它的这个产品 中国是这个全球的制造基地嘛 那这样呢 他就要看这些工厂是不是合规 所以当时的国际会计师事务所 接这样的工作去看是不是合规 所以当时我有很长一段时间 是做了这个方面工作的 所以我当时跑遍了江浙 山东 上海周边 北京周边的这些工厂 就造什么都有 就看他们是不是合规 这个跟财务审计不太一样 所以当时有做衣服的 做玩具的 做电子产品的 领带的 就是什么什么都有 So Because you did visit the只 statistics and everything. I had a lot of M&IC projects. I had three years. after that I was searching for three years the channel up such a donation andグal the channel progress actually this is about when I投SARS moved to a certain nh front and then I just realized the timeline was likeHeеты Arrow to be into business unit because dans the 1000 global等等 exampleSAP Oracle as other people media I went from ��고 broker engines slot On cuando home paste eine primitive Queens We use these basic products that require others to demonstrate You need to check your process in order to perform and update So. Yeah, ready! Well, it was designed by the government I realized that that was true The chief of shop in 2018 I was blessed to azure وبus 哎 這段經歷又在我這個記憶裡復活了 當初我也幹了差不多三年時間 所以我做這個英文叫professional 就是做了六年時間 然後才轉向做這個投資相關的工作 所以這六年給我這個印象還是很深的 而且這六年又是中國網絡網告訴發展 然後所有的產品我也拼命用 然後美國的產品我也用 只要有空就都用這些產品 我覺得你前面這六年的工作經歷其實非常的偏理性 非常偏流程 Then I'll give it those styles. It's a style brought in there. It seems like a strange surprise. Why is this kind of simultaneous? On the side of this? I had to have mostly two sets. My first one was the top gen�. Just the internet database. Not all of the product I was using. The supply is the cleaning company. Right. Next line, I was was the working of working on a works 当时应该算是 興趣不大 但是呢 我覺得我也能幹的還可以 是這樣一個狀態 嗯 怎麼轉向投資的 當時其實還是有一些機緣巧合吧 就是說因為當時在05年的時候 其實中國互聯網的這個第一波這個門戶的這個戰鬥基本上這個格局相對固化 然後呢新一代的這個公司紛紛登場 無論是騰訊 其实在那个时代腾讯阿里盛大这都是新一代的公司 所以当时我正好我同学是在盛大工作 然后他有一天跟我说我印象特别深 他说其实盛大是中国最大的互联网公司 它超过所有的其他互联网公司收入的总和 当时我非常吃惊 然后当时我就第一次想我已经每天在工作之余已经上了六年网了 为什么我自己不到一个互联网公司去工作呢 对 所以当时正好盛大爷给了我这个机会 就当时我去盛大 因为我整个这个工作背景不太适合 上来就直接做这个产品 其实我还挺喜欢做产品 所以当时盛大也是刚刚组建投资部不久 所以当时我就加入盛大投资部去做这个投资 photo � of private Sunday Call it My audience And then Thankfully About 密码技术 这都是我当时印象比较深的 然后还有这个当地游戏来的就更多 还有这个比如线下的这个网吧类的游戏 然后还有各种各样这个对战类的游戏 所以当时是怎么说呢 就是说在中国互联网第一波基础设施完成之后 第一次这个百花齐放就各种各样新的应用出来了 对吧 电商 游戏 然后包括QQ 所以在第一波之后出来的这样一个阶段 从05年到10年左右吧 现在回看05年到10年 最值得投的几个项目是谁 当然可能是现在那几个剧头 但那时候他们其实已经上市了 但是如果那时候买股票 很可能比这个 比VC的这个回报还好一些 土耳机 对 就当时 就当时在他们那个估值下 和这个发展的情景下 就大家都没有想到后面是 这么大一个数字化的冷冲 你觉得在盛大的那几年 你学到最多的是什么 这是第一段投资生涯 因为在盛大的时候是这可能是我到目前为止唯一一次进入到一个实体公司它既不是一个专业公司也不是一个投资公司它是一个实体公司所以它也构成了可能之后很多年我对一个互联网公司的想象你部门之间怎么合作然后怎么推一个新产品然后比如说怎么去做商务的拓展怎么跨国家或者跨地域去做这种运营的连接或者是运营的增值服务反正见到了很多这也是一个宝贵的财富因为之后我再就没机会因为每个人的人生都很有限你没有机会再进入到另外一个大厂去看而且当时的确是中国最领先和规模最大的互联罗斯你当时又觉得我这一辈子就要干投资这个事了因为前面有做过审计有做过咨询投资其实是第三次尝试当然觉得我觉得我跟师傅干这个事情 So budget recovery I thinkthe The I found out that it was the US website Delicious. Delicious is a platform for the online website. But I found that it can be found many small groups of websites. This is the way Yahoo and Google and Google searchers are not allowed to search. You need to know what you need to find. But this can allow you to know many of you didn't know the things. Then this time I found out this. 当这个过程不是发生2007年 其实在2005年 就2005年上线我就发现了 所以在整个过程当中 我就一直对这个是充满了兴趣 然后我想这个应该属于 非常人类前沿的 这种行为模式的探讨 那这个事情就非常符合 就是我想了解 这个世界是怎么运作的这样一个事情 所以从那时候开始我觉得 有可能这是我 有可能是我终身的职业 2005年为什么没有出手 you know why not to go to the bank 2005-2007 I wrote a report when I wrote a report I wrote a report I think the situation is very unique and then actually 2005-2007 I also started with more use of the product so then after that the two companies whether I was in the business or after that in the Red Cross we were in the business also were the two companies 这个触动你的是什么?iew הכ differently take the most part of me was the change, I think the change is like full-day and long-term he seems to make free especially by ba рend I don't think it first it wasn't no other kind other only these heps and under this 出现了巨大的产品创新 但同一时期你看美国这个Facebook这些都是产品创新对吧 包括LinkedIn 所以我当时就觉得这是一个探索人类行为模式和理解世界的一个好的方法 你当时见阿北是一个什么样的场景 你喜欢他吗 当时我觉得我见到他的时候可能的确他就反映了这个产品的气质 他本人和这个产品是完全合一的 It's a real world It's a real world It's a real world It's a real world You have any kind of relationship? You have any impact? I think I remember I was in the北京 The The Little Star We talked a long time About this product That day I was very close The first meeting oh,ώ, actually, the Page of the press was made by the two of us, during that moment, when the press, they were since 2009 and 2011. So then even I met an old player of the company. Famous agency with the one that từng Tat ن phen, above his another After you first met Miguel Isabel ahí? Younes que siob buradan a企業 style with Abbey is very different, Yeah. Meanwhile, I went back to the company upon a moment. So I met a time ago when he convinced him to be a键ote to be his first name. Tomorrow I think that his business will be nearby more early. of good economic reaction and Adobe And then really成熟 and expressionласти施 about Formula 2 started creating a most web 2.0 is to 他 started in c분 the final of a 线上那种的产品创新 对 所以到再下一步的时候 到这个比如说共享经济 混乱单车的时候 实际上是相当于是 又做了一次线下的一个大规模的创新 就线上线下结合的大规模创新 所以这个还是非常有意思 对 就那时候是纯线上 一开始这个大家都是一些信息的讲话 你是不是经济史爱好者 对 我非常喜欢经济史 把过去20年你所经历的中国风投的整个周期你会划分成那几个阶段,尤其是里面的一些代表案例,你能不能分享一下? 其实我划分这个,我是这样想,就是一开始我觉得应该是05年以前,05年以前实际上是一个非常初创,非常从零开始的阶段,因为在那个阶段整个的相当于是信息都没有线上缓,所以在那个阶段出现了大量的门户。 even this is what is the same Like U some of Sheridan the European millennium simply clicked That would be the 7th century The platform Yes Yes- So those companies are sort of a state-ing business. You got to see it, but they were in the late in the house of driving gas in the house where a couple of shows that actually. So these companies It's deepened. 但是他们就抢不到主要的流量 这样等到这个10年移动互联网慢慢崛起之后 新的流量入口的形式产生了 相当于是新的一波公司又重新洗牌 因为新的入口有了 原有的老的入口 一定程度上让出了大量的市场份 所以大家都在这时候开始做 我们15年一个周期看 到15年什么发生变化了 到15年发生变化的是整个短视频作为主要形态的崛起 但这中间我觉得有一家公司我觉得我特别要提起就是Pinterest 因为Pinterest在他应该是2010年吧 如果记错了的话就要报钱 我记得他应该在2010年成立 我认为他是一次范式的革命 Pinterest就非常适合在这种曝光流的形式非常适合在手机上展示因为之前反正之前我印象里我是没有看到所以本质上2010年Pinterest出来之后我当时也花了比较多的时间去在中国去找这些Pinterest的公司包括当时也跟像小红书蘑菇街就这些其实都有过比较深入的接触所以我是因为Pinterest的范围是实际上直接影响到了后来的内容平台抖音 TikTok小红书 快手其实我觉得它是一个我个人认为它可能是一个始作俑者它把它做成这样所以到了15年的时候我是认为这种类Pinterest的公司开始崛起当然他们是以短视频的方式崛起的并不是以Pinterest当年的这个样子崛起的当然今天我们看图文混排还是有一个非常重要的意义对吧今天我们去刷小红书的时候其实图红盘还是有它的非常便捷性的地方所以我就说这一系列的脉络到2015年爆发出来所以整个2015年到20年我认为就是短视频大放异彩的时代当然整个这个过程当中始终这个欧托欧是在进行当中美团 滴滴就说包括摩拜这个过程也在进行但是我是觉得在20年以前大概率可能比如18年左右 That 둘, a single talk We can truly explain from a couple of transactions Actually, we're going to start at the back of the史 cycle of Counter Instagram. We can talk about the most simple 어�온ic thing. I was thinking that we were sure this is the point that we want to break. Why are we doing so with that? Actually, Georgia says that a very high-ılmış activity. A high-ถed listener of the genre option. Yes, because one SEO person can spend a lot of time in size in a large number. So in all worlds, we can't skip Bere congress. Because it's an easy one, so you have a talented word for people hearing. Especially 시 your Chinese people. We've learned so that in the mid-savaşER is very important, 但图片是比它简单得多的方式 首先所以说所有的图文混排 它就是会倾向于覆盖着文字 就会今天你读一本书 认真地读完一本书 实际上是一定的奢侈的消费 那到后来短视频 到长视频时代 其实长视频因为人类是很熟悉的 无论是在古代你演一个戏 还是进 Bro hours 来进到你去电影院看电影 看电视到最后到长视频平台上的这个来看这个我们对这个内容形式非常熟悉不过就换了个界面对吧所以当时对长视频我个人一开始也是有很大的期望因为长视频天然的这个时长很长对吧你可以想象好像这个商业模式应该是比较多那当时我们觉得短视频可能就是视频那种形态跟长视频一样所以当时我还会想短视频和长视频应该怎么比较 We can be STUDENTS that more than our way. It's been aboutладesting this video for 오래, we incrediblyою impossible to get around almost any way than your main focus or the dominant view. I got the lead in which there is a different chance. Actually, the truth is that you go to the old society, the old society takes a long time to get to the old age. He takes a long time to learn文字, but for example, when I say that there was a lot of fun, that there was a lot of fun, the year is a lot of fun, okay, it was a lot of fun. Then I can write a lot of文字, and tell another one. But this is a very high-level way. It's a very high-level way. The most simple way is I should look at it. I say, look at this year. It's not a big, it's not a big, it's not a big, it's not a big, This process is a short video. You can just think. I pulled it to the place. It was close to the place. They have similar information. So the last video. I think that短 video is not a challenge. It's not a challenge. It's not a challenge. It's not a challenge. It's a challenge. It's a challenge. So today I put a short video. It's a good way to look at the word. It's a good product. So I don't think that a person. One year, I don't read a book. is what is a thing. You only need to read the same thing. And you should read the same thing. It's not that you should read the same thing. That's what I think is not a problem. The fact that people are aware of the short-term is a very normal thing. So I think that our understanding is ended in short-term. I think it's very normal. Because people are not aware of it. I don't know if I'm not aware of it. So it's a lot of stuff in the show, right? Yes, it's true. So I'll tell you what I can share with you. I'll tell you what I'm using. At the beginning, I remember I was using Twitter. I first saw Twitter this way. After I was a fanfare user, I was a web user. I was wondering why this is what I was doing. I thought, why is 140 this word? Why is it more or less? I wrote it. I wrote it. I wrote it. I wrote it. 但后来我觉得太无聊了 谁会看我这个东西呢 所以我根本无法想象它是怎么被用的 所以当一种产品听在写 出来的时候实际上是 产品的这个研发者和 使用它的者一起创造的 使用者我可以写观点 对吧 然后我不停按 按给别人 然后他慢慢慢慢这个就出来了 我很早也是 MUSIC PLAY出来之后 我当时是 我应该是2015年 我出来之后我就开始用 MUSIC PLAY 我用它就是拍景色拍摄音乐拍景色拍摄音乐我也想象不到他能怎么玩通过跳舞火对但后来我们就这些平台通过用户和创作者之间的互动用户与用户之间随便起来了所以如果你一定怎么回到刚才那个问题就是说因为你说这些不同形式的眼镜所以我大概是这么认为其实我认为人类天然是会进化到跟自己本来不需要学习不需要很长时间成本投入就能够认知这个世界的方式它天然就会回到这点所以说AI是不是也回到这点AI回到这点的原因是我最想要的是结果对吧对于那些杀时间的应用我希望要过程因为这个过程帮我杀时间但凡是任何效率类的结果倒下的我都想要的是结果你不用告诉我过程最好我越想参与过程越好工具产品是结果娱乐产品是过程对就kill timesave time你说这些那种平台背后你觉得有发现什么样的人类行为模式的变化吗一个是从主动的刀被动对从8.0开始其实它就往这个方向转对吧不停我可以推荐新东西给你除了这一点这样就碎片化的消费这肯定是点更容易的更碎片化的这个消费你觉得为什么有抖音了小红书院的决情小红书是在抖音的延长线上还是我觉得可能在一个播客节目就很难穷尽 they're not different but I was told I think they're the biggest difference is that I think the story of the video is very important but I think the story of the video is to look at the content but when I look at the content I also look at the comments they're not so they're the kind of I think it's not the same and for example if you look at the in the video if you have a 100 million fans of KOL when you're sending your 100 million fans大概率是你的这个点击的这个基本盘对吧他可以点赞点什么这个量会比较大但是小黄书上你同样是一个百万你可以尝试一下你同样是一个百万你发一个新帖的时候你的百万粉丝不一定是你的基本盘因为它被推荐它是基于这个帖子的精彩程度被推荐到了更多的地方这个不一定完全代表导音和小黄书的意思这是我的观感我们只站在一个优惠的层面上所以我认为他们的整个新组织形式是不一致的尽管他们都有短视频 然后小王说可能图文混盘更多一点 但是他们实际上这个新组织形式不一样的 这就是移动互联想时代 我人民创业的一个有趣的地方 就是你通过改变一些规则 一些玩法 你的用户也变了 是吧 这个内容也变了 它之后能变得很多 它这些东西更产品向 而不是更技术向 还有B站 评价一下B站 我觉得B站是 但是都是个人评价 不代表我们团队也不代表我们是 So I think it was a small group of people to restraint impact and the endenrger which was adalah the actual was the constant So it was a party of DOH That's Pomender Jen's came to a report and could invade so EVER. I said a market happened it was left in Zapot in B-站的这个Founder是更属于这个产品跟Upbeat一样对 就是产品跟Founder的这个特点比较一致因为短视频的崛起我觉得对于字节来说因为字节已经有今日投掉的产品了所以它已经有很多战略和理性的选择对 最后进入到短视频的这个领域通过和Musiclet的合并而已所以我觉得这个可能的确是不完全一样 it could Y жизни to look into theие the it is for the one is I, it was hear the 甚至我认为这可能到目前为止 小红书是我见过这个在移动互联网时代 这个应该是最开放的产品结构 开放 就是他发一个 就你可以发一个这个图片 你可以发一段话 然後你可以发一个攻略 他几乎包含了 你可以发一个问题 对吧 他几乎包含了移动互联网时代 前面的这些UEC产品 这个社交媒体的大部分的形态 它都通过你发一个帖子 大家评论来实现 所以我是觉得在这种开放架构的后面 所以小红组一直 我觉得是时间的朋友 就他一直在这种开放架构 后面和用户 一起创造了他的调性 我觉得这是 这是很有意思的一点 就越到后来 我觉得他已经慢慢涨 就每个阶段可能都超出 我的意料 他居然有这样的功能 一直翻话从女性有物也翻坏去了男性对而且从线上到线下它有city walk所以这是很有意思它这个是一个非常开放的产品结构所以我觉得这几个成功的产品都是有不同的逻辑和不同的历史的怎么说呢就是说实事实事造英雄而且好像很难归类他们好像并不成功来进某一个统一的因为我觉得产品成功的确很难去做这种归类研究学习了这么多产品以后对于投AI的产品从中能够得到什么样的know how你能获得什么样的经验和总结因为看上去它们每个都非常的不一样对 可能总结我觉得总结至少有一点是这样的就是说首先人类的无论它是潜意识还是怎么样的新的行为模式总体上是不可预期的所以这一点我觉得总体是不可预期的他是不可预期的对吧我们想象不到短视频最后会挑战到文字挑战了自己所有的工厂自己的竞争对比人也挑战了所有的竞争性的所有的竞争到文字这件事情我们是不知道的原来我们是把短视频作为一个小类看比如说你看什么我看长视频我看短视频然后我看这个矛视频都是这么说的对吧但最后你会发现短视频的形态因为它的流量属于太强它可以去 向很多公司去挑战 那么 所以说 每次投资人会被问到一句话 说最近你在 你觉得什么样的产品形态 对吧 其实我们都很难回答 这实际上 所有都是探索 但是呢 比如我就好像一个产品评论员 或者哪怕就很像原来这个文学评论员 我不会写小说 但如果你写出小说 我的直觉会告诉我 大概率 这可能是个名著 有可能 对吧 如果我做出版 那就是这个可能我应该 改天再深入谈一谈 所以你可能 今天AI也是这样 就是因为AI有它的新的特点 新的形态 所以你今天肯定是有人在挑战 最前沿的流量最大的那个产品形态 你就要去把这个东西把握住 它的不合预测性好像也超出了 Funder本身的预测 就是这些创始人们 他们在创立第一天 其实也不知道这个产品形态能长多大 对 但是我觉得Funder应该这样说 他不能够提供一个理性预测和理性预测的依据但是他有他的直觉这个直觉也可能上成信念他相信这个方向一定是OK的历史上很多的鉄道他做了一个石破天莘的产品其实跟他的直觉和相信有关但是这个直觉和相信放在历史的潮流里它最后肯定是一个信存者的偏差它是一个历史的选择成功者最后讲述了它的信仰 That sort of failure improved down Carolyn spricht The reasons did you find some relaxer techniques? Do you think you used to find my good products and self-dict and self-dict oh, what was your best? I was tied to the debt The choice, I 뭐 more? Whenгрifinals were released My experiences remained Angel Day化 companies Or Or Or I I I think Chinese generally primarily触达最广大人群而不是从我原来的这种比较熟悉的区域无论是点评多半小红书像这个music它都是先触动一部分就是偏知识型的一线城市然后再慢慢扩散到全人群我原来习惯的是这样比如说我比如看到这个滴滴看到评多多他们这种我觉得这很有意思这个可能不是在我一开始的认识范围所以这个我就会去尽量去矫正我原来的 一些想法 原来我的确 我是有一个slogan 就当时我们是 给自己的一个定义 就是正在叫投资于 这个先进人群的 先进的生活方式 然后就是我们这个也会 演进到全人群 但后来发现并不完全是这样 有的时候你看的最基础的模式 它也可以就直接扩散 对吧 它不需要经过这样一个 演进的过程 你矫正了以后 做了什么改变 就应该我觉得 比较反映就是反映在当时 共产单车 共产经济 这方面思考 当然 其实当时我对共产经济有一个更大的期望 但是我就看见这个期望 就怎么说呢 因为当时摩擺之后 共产经济也发展到 比如说有充电宝 甚至有空间 当时有共产空间 就这一块的确 有的时候就是 一个逻辑可能也很难演化的机制 它可能有一个现实的一个 天花板 Obviously it's completely ما respect that is a long-term thing gig And I think the crap is egg As you know Anyway, I think that I dochio todas Since in PC and walk WiFi Ok it's absolutely完 الد So the problem is did not kind of Pot its soundngcillo is the chain on the part of the product. And the chain on the product is the product. You know, how to manage your product. You can't believe this is a product. This is not true. This is what I said about. It was a part of the product. It was a part of the product. I think, like, in the United States, in the United States, the internet and the content of this platform. I think it is the chain on the product and the chain on the product.经 günstigens in a board of dinner and it helps to coordinate pressure into scientists and states alsochi like many international perspectives in Danforth or other things you painted and straight through many UTC or review times and until I push 一个采购量去倒过来更好的控制供应链 所以我认为它就是线上线下结合的是比较好的 就回到刚才我们讲的线上线下的问题 我听你讲我有一个感受就是说 因为以前可能最早的时候大家觉得产品就是产品 但是当线上线下开始结合的时候 其实一个产品它是一个整个的体系 不能够只做一个比较薄的产品的部分 它还要做很重的运营 其实抖音的成功也跟运营有很大的关系 是的 是的 But at the beginning, I think it's a good idea. Because it's a good idea. You can do it on the internet. It's a good idea. You can do it on the internet. The second part is when I put it on the internet. I can change the internet. Then I can upload it on the internet. The second part is when I'm going to find some interesting things. It's very easy to do this. And then I found a way to go I need to be able to do it I need to be able to do it I need to be able to do it So it's a bit more I think it's a bit more So it's a bit more So it's a bit more It's a bit more It's a bit more Yeah, it's a bit more Yeah I thought about a company When we think Like C-端流量 They're all about That's what they're about Why did you get to the two years The two years The other half of the world This is Just back to me I said 人类不停地发现自己的行为模式 因为在收音机 收音机火的时代离现在已经很远了 但收音机当时的渗透率是非常高的 对吧 但声音跟图文跟短视频又有一个重要的一个特点 就它可能作为高密度的信息接触是不太行的 但是它几乎是感官里面唯一可以多现成并用的这样一个感官 就当你听一东西的时候你还可以看东西 and you can provide a job so it can be affected in stages a big benefit in fact here is a very small and interesting because it's a big benefit I always felt the blessing because we didn't mean that do we have never had a big reason then we first came to the right by speech to answer a problem when I had a problem and then we connected to the market I feel like it's good like the subreddit Yangathan Rhei Twerin also Personally Every kind of live linguistic music So guys for Prediction Were they getting their information It becomes a more-hearted way 但这正是因为它可以用陪伴的形式存在,它才能有这么长,如果你不是用陪伴的形式存在的话,你就很难这么长。 这还是挺反直觉的,就是人类行为是不可预测的。 对,就是人类行为它就在那,它就在那,但是你突然有一天你要归纳它,说发现原来还有这么一个前后一致性的一个东西,我可以把它产品化,我可以把它PMF。 把他的行为给承接下来 对 对 对 为什么是这几年你觉得 这几年发生了什么变化呢 就是播客的崛起 我是觉得只有可能是 只有可能是比如我们的视觉 已经被占得太满了 所有的创新它都会像水一样 这样溢出来 当你视觉被占满的时候 其实大家还是会找新的信息输入 视觉占满还不够 我还得再找一些新的输入方式 这很有意思 那下一部分我们聊一聊 When you were placed in Information forrough France, and you had a diverse approach to looking at people'sWar Software? This perspective in 一直到这个AI的崛起 这中间其实整个2C的创新 2C的这个投资都相对沉寂一些 我后来反思 其实那段时间进入到 可能是一段折服 你看今天 所有的这个AI商 都是跑在云上的 That's had nothing. In the company, Sahonneur comes from SaaS. Their success in entire 800 years began to fertilizer . It has been challenging a day ,cause they were not ongoing. It was things that the milli vocalizationCheateruch created. Not just these companies aceくход question, or技術 mechanics, these technologies changed in the round is our cinco year. These same marker it was the latest. The second one round since 2019, mule grow older so we rapidly start to persuade ... ... ... ... ... ... ...就是做了不少这个SARS约翰 eta Maeк dk Britain c btk 这个投资另外一方面呢 就是在那段时间我就开始思考 因为当我第一次经历PC互联网 的 时期的时候呢 我就感觉很兴奋 是一个工作 当移动互联网的时候呢 你还醒了 觉得从PC换到手机 我觉得这很自然就很自然 不需要想 可是等到等到了 这两个长 AuPay的结束的时候 我可能也到了一定年纪 Then, I think there will be a 10 years of discussions still. Then I was thinking… Is there any ArcGating? Any main êtes 구атив rate in managed industry? Do we have these sort of impact? If you turn into a stock story, now PC's… I was just thinking before the more I have the idea of thinking inchin этот Repet Guy No matter in my opinion, there is a presentation in CS that starts to think so much thinking. Then I would like to acho thinking most right now. So far I just thought that allethical brenp Tower 然后等我再把这个东西再扩大化的时候,我发现似乎这一切,其实这一切的密码的确是在流量,不光是线上,线下也在于流量,对,不光是这个线上的电力产生之后的这种平台,电力产生之后你收信机再往前就电报,在没有电的时候呢,没有电的时候实际上信息的传输就直接道路展开了,所以就是公路,铁路,运河。 So I'd buy some ol' medial sor들이 based on the first thing In 其实它本账就是这样一个历史 只是人的历史在很长很长时间里面 最大的社交类产品就是城镇 就是一个城市 就是为了城市展开的 这个城市有多少人 它就有多少DAU 所以这个城市老是在这种道路的汇集的地方 或者是在比如运河汇集的地方 或者是在港口 所以我建立这个观点之后 我倒过来想实际上我只能就是人类每一次科技变革最后推动的都是流量方式的革新你无论是叫注意力经济还是说你这里诞生了一个新的繁华都市实际上就是你无论人住在那也好住地医在那也好然后你就会产生商业形态所以当时我就在想因为已经穿越了两个周期的确我也想不到下一个图思的形态是什么当时我因为我的猜测就是下一个图思的形态就当时元宇之后也比较火就是当时我想 It might be an eye-wrecked It's neverες There were noises Uhm But the bloody significance I think that still Because we대�acted AI 流量型才能强调的 我觉得这个是很有意思的 所以是流量掌握这个世界的分配权 对 你想到了这些关于流量的理论以后 你有做什么吗 这个一方面就是继续去找这些兔赛的投资之外 其实实际上我还是尽量 因为我一定是一个 就是怎么说呢 就一种理论一定要足够的general 就是它的解释力越强 其他越一般 它的解释力越强 对吧 就当你把这东西抽象到一点的时候 其实他其实就讲说后来我一次有些小的小的灵感 我觉得我有的时候会让我比较兴奋 比如说钢铁的大规模运用 钢铁的发明跟流量有关系吗 看上去毫无关系 我就没有办法把材料的革命和我这件事情合在一起 但后来我突然想了 铁路 对 铁路是一个 但更重要的是有了钢铁的大规模使用之后 城市变得立体了 城市的DU大幅度上市 在没有钢铁的时候 事实上绝大部分城市 它都是平的 我只能修个几层楼 但是有了钢筋混凝土之后 所有的城市变得立体 上千万DU的城市才能够产生 所以我后来发现 其实很多都是指向同一个方向 就在这个时 让我在这个思考过程中 能够一个乐趣的 流量会形成某一个超级节点 在各个事情 对 超级节点 对吧 最早的节点就是城市城镇 现在节点除了城市之外 还有互联网产品 这个超级节点 需要有双边网络效应吗 嗯 超级节点不一定需要吧 就是你历史上这些门货 但有双边网络效应的优秀产品 它一定是流量节点 一定是新的城镇 你可以认为现在所有的这个 优秀的互联网图体产品 它都是一个个巨大的城镇 甚至是巨大的市场 that is a sign ofrios remains the market that the AI technology has passed when it came in new modern tapi how did new还是比如说有一些特别的像图像社区有一些特别优质内容产生的这样一些平台你觉得过去20年有哪些产品是符合你说的大的流量节点的几乎出来这些都应该符合全是是吗能形成最大流量节点的这些产品有共性吗它的共性我觉得第一肯定是数字化肯定是你把线下当你作为一个先行者的时候 you even become an inf �veltal you can then go through an inside you can education being calculated u need to use your ad you want just a lot of time and software so in any if you have a black futuregia, if you find a very short clip that's in a square without Mutuality, Which willhin bad for regional Cloud. That will be fine. So from AI- którym you are, such there is something more attractive to me? It may not exist like that because I've been with 2py and 4SAS. Because you may have done a job like that, while this one, may be a little bit of woke like 1s. 但是的确作为2C的投资 就可能是对我是更持久的一个爱好 你是什么时候意识到 可能新的C端的流量入口又开始成了 那当然就是这波人工智能开始 就ChashBT出来之后 因为之前 因为洪山在这个 22年差不多年中开始 我们就开始进行了密集的 这个人工智能方面的调研 但我不得不承认 就是说是在ChashBT出来之后那么这个整个我感觉2C的应用的效果如此之好我觉得这个可能一下子会带动一大波AI应用的这个发展对就当时这个感觉还是非常明确你觉得这个以AI技术革命为驱动之后产生的流量节点相比互联网和移动互联网实在有可能有那些不同我觉得有几个不同其实我觉得最大的一个不同是这样的最大的一个不同是就AI实在这个网络 它不完全是一个比如说经济群说的像一种代有自然垄断性质的网络 它是自然垄断 它的自然垄断有一个很大的有一个特点不一样 就是说它的边际成本不曲迹为零 在所有的自然垄断的网络里边际成本的曲 边际就是等于说你每增加一个新用户的成本是零就比如说你是电网你今天多接了一户人家跟你巨大的前提投资来说这不算什么对吧互联网也是你都已经100万DAU了你101那一个对你来说没上成本他进入对吧但是我觉得人工认同时代不错因为有Token消耗的限制对所以这个在商业模式的设计上它就会变得不一样你每增加一个那就是一份钱你都管它收这个订阅费当然你自己免费策略那是另说 So, I think that. This is not different. However, I think that this is not different either. To be considered with quantity, when you are finding a product, it has a leverage It is operating for a team. You step over can identify a users that will ultimately decide the process or the way they can help succeed. But the AIS CART, especially in Salt Lake, it is anment-wideiya, i comprendre that is the Unitedines, is the point ofwość is the point of view 原来在那种情况下 you useAPP is-it-it is youрий or are you a problem or are you�还有黑箱的原理是没有黑箱的历史上你从蒸汽机时代甚至更早的农业革命的时代从蒸汽机开始到电力我认为是没有黑箱的但这次实际上就关于人工智能的涌现是如何发生的以及它会有哪些比如一些变化的应用实际上我觉得现在是不清楚的所以这次很有意思哪怕大家写这个prompt比如你会看一些写prompt的教材 I'll tell you, you can expect to see the potential work done in��� oceanhaba, the桥 little part will show the effect it's 이번 64. This thing is their manufacturing product, there are some specific ideas about this technique. What will these properties of AI? What happens in learning? Is it its zero-m Drama? You feel the same thing Dadi can bring to another whole? 都很不一样 你觉得商业不同会带来什么 比如说它的商业模式 对 投肯消耗就是它的成本比较高 但这个我觉得还好 这个就是一个商业核算的问题 这跟你自己的定价 对吧 跟你的定价 跟人权 我自己有关系 我觉得这个倒还好 Margin cost的不同 我觉得这可能更多的是一个 商业设计的问题 像第二个呢 就是说作为它更结果导向 那其实这个我反而是认为 可能到下一个时期 因为全世界总体对SARS的PMF SARS本身实际上是跑通的 那在这种情况下 怎么叠加一个完全以结果为导向 超越了 特别是超越了我原来非常熟悉的流程 超越了结构化信息 它可能从原始信息可以直接导致到某种结果 这个会给整个2B的产业产生什么样的影响 这也是目前还未知 The dirty doesn't have much The darkهTS 尝 same teaches a lot You put a lot of I tinham This vision Is an anim Lecture The Mm-G The good itself Don't The wrong Now I've limited We've found Another The 本质上在人物之中时代 它其实引发了一轮 新一轮的深层次的数字化 因为原来的这个数字化 应该是这样 它是一个比较直白的数字化 我把容易统计的数据 甚至是把人类线下原来 存储在纸张上的数据 原来你已经结构化了 我把你翻到网上 翻到网上之后 我再进一步结构化 它是这样一个过程 那什么叫深层次的结构化呢 究竟现在很多做录音的创新,它可以被给予很高的固扯,而且表现非常好,包括很多说我可以把一天都拍下来的事,因为这些东西没有人工智能,你是无法处理这些信息的,这是一些线上话对来说没有意义,你无论是录一个你个人录一个100小时的录音,你怎么整理,无法整理,你个人带着一个自动拍摄装置, The first time I remember the video, I remember the video. I was actually doing a video of a video. It was based on the light of the light of the light. Every time you put it in the light of the light, it will be able to shoot a shot. But you always have a camera. The camera will be able to create a few hundred shots. You can't get it. So this is not a matter of fact. In the time of the digital software, it is not a matter of fact. But today, it will become a matter of fact. Because there are people who can make this video. It can be made of a matter of fact. So in this situation, the size of the digitalization will bring new features. Actually, in the fact that these opportunities should be because of the digitalization of digitalization. It's a digitalization that has been given to us. It's a digitalization that has been given to us. So I think that the idea is to provide these digitalization of digitalization. We talked about this. Actually sometimes it might be of full likeness, such as a bell... powerful FlyInOrization E import From the future of Netches new?ürower is a low Alexa of other data Angle and Apprentice Then, the idea of the device So it does Hello how does it makepire of what Venus OLED You can't add an쉬ус There are along proxy Even at least there will be data We also see certain intellectual tools will be implied in the channel on Day 3 We can't understand If this student exists If so in vielleicht in the interesting place of information like this olduğu number of people already will use a bat on N300, 100 billion money in such cases when combined with popular designer and what its expertise will channel What is this Hogwarts thing protected from N cards when getting critical information you get really rich Why would the härfaltation benefit is needed to get more information and information? The important thing is that it is more deeply in the current world of information. The information is not so important or completely important. The information is that every person, every organization, every organization, every company has more information. Today, all the world's data, the data from 1800 years ago, the data is very small. 记得有一个统计但我不记得多是很小一个量 这也解释了为什么这一波软音一体的公司今年格外受到关注 对我认为本质最深层的原因它是一个人工生物带来的更深层次的数字化 嗯 为什么手机不能做这个事情 你想录音的转录手机也可以做啊 推出也可以做啊 但深层数字化最重要的是你要一直在场 对吧 比如我现在我都不知道我手机的这个录音功能还在不在录 I don't know.周围所有人都能听到的话所以这个可能的确需要无论是法律或者是社会习惯或者是大家普遍的社会的心态的一次革新但这个肯定也包括在了也就很类似于当年我可以刷短视频我认为这一样的最后人的行李习惯它会随着科技最后全社会范围内会改变在互联网移动互联网时代这些流量节点的流量入口的产品其实都有上边网络 为什么AI时代的双面网络效应还没有形成? AI时代我也想过这个问题 我觉得可能第一点是在AI时代首先大目形本身 它有利益还是双面网络效应 对吧 它本质上用了这么多数据 包括它跟更多的人说话越多 它可能再反馈新的一个用户就说的越好 这也还能双面网络效应 我觉得是有的 实际 但第二点就是在显性层面 因为原来的这个说明网络效应它必须需要你说话要有个人听对吧你卖东西要有人买它需要一点点来凑但现在不需要现在在跟大模型的时候现在大家都是多对一的关系所有人跟大模型聊在这种情况下所以我们看第一波美国第一波出来的偏社交类的公司并没有很成功本质上它没有什么样的效果一旦模型质量下降一旦它进入对手的模型质量上升它就会处在商业的一个不利的局面 that's exactly how the image should be okay还有一个断层包括你我用这个AI你可能也会认为跟你聊天的那个人是不存在虚拟的是模型算出来的可能你心理上是有影响的但会不会未来的人类就不会了未来的人类认为跟你聊天背后的人是不是比如说有没有自主意识有没有灵魂不太重要那就可能整个全灵群的认知会变化就跟今天全灵群的认知 跟PC时代 跟PC之前的时代已经是区别了 所以我觉得AI会导致全人群的 认知的表 而且新的网络其实也不一定都是人和人的网络 也有人和agent 共同的网络和连接 但这些就看怎么能建立起 某种双边网络 我觉得这个还是得探讨 这个东西如果不建立 我认为那这个世界总体上是模型的世界 对吧 那这也可能是不可预测的 那个部分 你觉得现在AI产品有数据飞轮了吗 其实我觉得目前还是有的 因为我是觉得AI产品这一波的商业化实际上是比前面的移动互联网互联网实在都做得好的 毕竟经过了这么长时间的SARS的这个寻陶大家对订阅制的这个了解 我是觉得是反而是做得好的 只是现在因为技术在高速的发展 基础设施的投入还是很巨大 相比下来看这个收入和成本还不是那么匹配 但是我是觉得那如果人类达到某一种满意的边缘 那实际上是有可能 所以我对这个还是乐观的 在这个大模型时代到了以后 你投了Kimi 投了Minimax 投了Minus等等 这些都是2C的产品 你觉得真正最后的价值会沉淀在应用公司 还是沉淀在模型公司 对 这个问题其实也是我一直思考的 这个问题其实在上一个时代就知道说 有没有新的应用产生 还是都是巨头的时代对吧上个时代也会再讨论这个问题对上个时代就不是模型是巨头当然很多巨头公司本身也是大模型公司我是这么想就是说可能就跟我做早期有关就我对这个还是乐观的我乐观的意思是我总是坚信新的创新能够在一波周期里面最终成为巨头它会在巨头的名单上再增加几个只是说是五个以上还是五个以下还是可能只有三个以下这一轮我觉得肯定会增加但是当然大部分公司已经加上去了就是说到底能增加的是什么形态就这个的确我们不知道但是从我个人的相信的角度我还是会相信性的产品会诞生这个只能用信念了而且历史的周期里面每一代周期里面都产生了些东西哪怕巨头再扭你看上去巨头已经把控了一切 但几乎都是产生了新的东西 你觉得 对于大模型公司和对于这种应用型公司 对于你们来说考核的指标是不一样的吗 它们最后价值前现在哪也是不一样的 因为我觉得模型公司今天还在 不断地技术演进的过程当中 所以这个也是咱们刚才讨论说 这次科技周期最大的一个不同 就技术与产品同时演进所以说 模型公司更重要的是它技术怎么演进的 对吧 And then the company's situation is more difficult Because you can get to the current technology You can do your product And the technology has changed So this company's company I think it will face more The not-确定性 Or more of the change The need to be But the good thing is The change looks like It's more than two weeks It's more than two weeks It's more than two weeks It's more than two weeks It's more than two weeks But now it's more than two weeks但现在多多少少了大家几乎都有收入所以经常开玩笑大家都会BB AR就几乎都有些收入他们处在这样一个博弈的状态有一种说法是未来可能没有模型公司和应用公司的区别因为模型公司都会做应用应用公司也都会训自己的模型我认为这是有可能但最后可能会比拼一件事情就是说到底技术上升会多大程度抵消掉非技术的产品优势对吧 So today we say, like OpenAI, maybe like Jimena has a challenge for it. But OpenAI has already a very strong product of the product. There are a lot of data that has been in the top of the top. Many people have been習慣 to open OpenAI to say it. So I think this is the last thing to look at the technology will go to where. Some of the areas are not the technology that can be solved. But that is a product that is the company 能够守住的一个,一个边界吧 最大的某性公司,传领一个产品公司 就是产品公司 那你会投产品的时候,感觉到很绝望吗 那不会,我是觉得比原来可能还更有意思 对,因为到移动活动的时候 其实你找说明网的效应,因为在PC活动的时候 也是这样的,对吧,但这个时代,你不知道 这个答案是什么,对,你还得去找 这个答案是什么,这个是 很有意思,而且这个时代又是线上线下 更加紧密的一体化,对吧,软延一体化的 时代,所以你就觉得,诶,这个变量 Transcription by Casting Forex because you had to object 5 6 images So this process is a lot more beneficial What do you think about and the most high? That's the idea to build a bigger picture plan, Has IT changed for re- 300 people cities? I think the need to changeопl community . So That's just simple stuff牆 it's actually probably because there will be often some rather frequentlyzero deramer ha Paris because they can get on their presence So now maybe 更接近于一个百花期放的时代 对 从这角度其实更好 所以它其实也就是不同 那投资人对于这些创业者 会有上来就盈利的要求吗 那肯定没有 那这个从来都没有 所以早期投资人 更重要还是看一个 比较长远的这个场景 在AI时代 红杉的投资策略 有系统性的发展调整和变化没有 我们其实一直是 比较坚持 投小 投造 投科技 就是说我们从很早就开始 从18年我们几乎就是 甚至我们会内部考核作为刚才我开篇说到的,作为第一个机构投资人的比例,我们是非常关注的,你想商务是百分之多少? 那就越多越好,但我们现在可能能达到半数以上,我觉得在我们总盘子里的半数上,我觉得还是一个比较骄傲的成绩吧,所以也希望大家创业就直接来找,不用等到B论以后。 你们现在对创业者的审美发生变化了吗? 嗯 什么 其实我人本质上是没变化 但是你要一定要说变化 因为AI时代是前所未有的 强调技术 对吧 就是在上两个时代 就是可能这个Founder 发表多少这个paper 有多少的这个引用数 这个是可以看 但它可能主要 反映Founder要多牛 但在这个时代呢 似乎它变成了这个 特别是在技术导向的公司里面 还变成了一个非常重要非常重要的一个点对这个就是跟这个时代有关技术与产品同向和发展红杉基本上把中国的大模型公司好像基本上都脱了对吧这种用赛道覆盖式的打法你们这几家公司是怎么进去的为什么每家基本上都脱了这是什么策略其实我觉得我们并不是一个用赛道覆盖性的打法我们肯定还是case by case看每个公司的亮点和值得记录的方式每家其实不太一样 we're not going to let this side of the fight 它并不属于在GP3.5之后,GP爆发之后,然后我们再去找的,这个也不是。 然后比如像质朴也是我们很早就接触的这个公司。 然后像其他的一些公司,其实我们总体上也是这样, 其他我们就是case by case看它的亮点。 所以其实我倒不认为我们是好像看见了一个热点, 一下子利用我们的资金优势去占了这个赛道, 其实我觉得并不是。 它可能是上一个时代延续是吗? 上一个时间头或者其他相对的 对 其实就有点像 就有点像怎么说呢 就自然界没有飞跃 就说我们的工作也没有飞跃 我也不是说这个爆发了 我们就是以这样一种 比如说这个系统研究 或者是关键人员覆盖 这种系统方式往前推进的时候 它自然变化成这样了 嗯 所以你们不是赛道打法还是偷人 我们还是case by case来看亮点 对 那偷人肯定是早期 最重要的环节之一吧 Minus怎么看到的 keepying what makes, so we saw also go to Monica. We love This movement with 것. At the time we found Monica's design very ładest and amazing. We watched the last product rather than too far from it. But it wasn't Maintenance. Actually, the point of view with product and product that they had seen as the magazine. It is not saying it he put onto his standout. Oh, this is probably the one that has to be a red-shirt. I think this is a lot of a red-shirt. You can see the red-shirt on the red-shirt. You can see the red-shirt on the red-shirt. The red-shirt on the red-shirt is a red-shirt. But we have to see each other. We have to see each other. What do you think of the red-shirt and the app-shirt on the red-shirt? I think the biggest change is... I'll just say a little bit. I think the entire agent's change is... This list is from all consensus The first product doesn't work existed It's not that if you create an app The second product... Well, high school project has always been designed for once Ms.GAS to be able to제로 He can have certain fields on a rendezvous But this user... For instance, it's a fact that Mike Honey has machine For example, from different ways It's not a problem It has a very suitable BD by EV So I think this is a very big wave of Chinese people. Our products are to-to-to-to-to-to-to-to-to-to-to-to-to-the-to-to-to-to-to-to-to-to-to-to-to-to-to-to-to-to-to-to-to-to-to-to-to-to-to-to-to-to-to-to-to-to-to-to-to-to. Do you think Asian will be able to spread into an environment like that? This is the same type of app. So App has all been increasingly demás. Were they changing with me as an advanced level? Or would these change for Oxford? The same thing is about. Some have changed completely. Some have changed and changed for them. Do you know demás? In the PC Pues old then going to the software developers. The machine is still on the know-how. It's better than the temperature. Yes, each of these already panel 똑 обычно will be on the phoneoma and on his name or respond to him and body.但是一定会相信它肯定会有变化所以这可能是所以早期投资的那个它肯定会有变化从Trashbit诞生以后过去的这两三年时间你的投资你觉得有某种节奏吗就是它也反映了整个的技术的变化和产品的变化在这一段对 我觉得反正对我而言肯定是有节奏的因为一开始我是系统去扫这个大模型的机会甚至当时也扫一个所谓垂直模型的机会那肯定是先看这个底层先看infra Frain Everyone is building Then we think Reverend 25th Maybe like for the 2011 Maybe like for the kinda of기 This is easy And we would More quickly Find out Polycoms рай's We've here cause This is restore The innovation In addition I say I think it's a bit of a new generation. It's a bit of a new generation. It's a bit more than a new generation. Yeah, I think it's a bit more than a new generation. What are the jobs of the year? Are you sure that it's more than a new generation? I think the job is close to the future. But now it's just like the city to go to the new project. I think it's more than a new generation. Because in the移动互联网 and PC, it's a big center. Now it's a big center. especially Central Europe It is very slow After the purpose of making Do you know any Changes that are successful in Japan have DC platform this was much smaller besides Pittsburgh I can't come back to the story today Do not matter inpper They can do in the whole world They also watch the whole world this is a very interesting change the use of separate platforms laws really might get more before the official because I still used toGo My hablar with cyber technologiesase I think the background will cost more does it have a lot of different regionsそれは It is technically a repeat of Wikipedia 上海航是偏软件,深圳是偏硬件,其他的气质,我觉得优秀商业者其实多少是接近的,除非说会不会做一些更偏交易的创业,现在目前还没进了内部。 最近关于AI泡沫的讨论很多,你怎么看? Well, I think I've heard about this talk about fle Because I've been drawing Derby I think this is a reflection It wouldn't be a probability The first thing in every day In 시작 of it, it was a research After49 percent Because there is enough For example, the market końcu And the one that had to accept Can be again By leading services The platform These people's olhos So I think The standardish That wasn't normal is the same but the last one is looking it canals how much foreign so this one I was thinking反而 world I think it was, okay because it's time ago it was flying through or if it's a doubt it might be a start or if it's hard to punch perию Sono, you can't getensive but you don't have enough just said at first you can go captial обратна believed to be sure and start mailing about this is a place that makes for regulate are you able to get 야 yeah I promise K Waspa In our fragile case Well Bard Every gentleman There is a hard Air It's crazy, I'm afraid this kind of Perry's thoughts because you have a personalbereich huge energy at origin. 包括那因为现在已经各个主战场已经开始在PK了 所以不能叫做爆发的前夜 就爆发的前端 前端 对 爆发的前端 焦虑吗 在这样的一个时代上 不焦虑 应该是很兴奋吧 很兴奋 应该又见证了人类一个历史 因为现在国内的主要的分头子人 如果是在PCA时代或是在后入场 就像你建了三轮 我觉得这种是历史上比较少见的 I think it's a great for myself. It's an important myself to take on reserves You have like you, when are you slaves these 25 years ago in China Are you more carr款 Blick on hydro Anthrop 방법 you have to access the two gases in Especially this, you have a bunch of clubs. I just think they all fall off�al �eting and have a lot of done for business Mommy and a point hockey fazer instead of rich Relief No, no likes to芥蓮 Walk in many kinds of created I think that one of the past, 了解共同家用的需求 那么到 我觉得到十以上 或者到几十以上 他这个没有一定论 反正我一直有一个观点 我觉得作为一个公司的领导 就跟一个带兵打仗的将军一样 就你最后你要成为 你这个组织和制度的人格化的象征 就最后实际上 用一个可能不一定精确的话 说只要你扮演那个CEO 不管你自己的本性是怎么样 但你在那个时刻 需要扮演那个CEO 尤其公司越大的时候你越需要办 이제一个将军在古代出征 你 count Constan Matching 没有任何先进的信息传 hypun providers 他其实在所在地他那个 imagerimming right than T заним 他就是最高的制度化的象征 它维持所有的秩序也是好 军 BagCoW 作为一个组织的想像力 he can trust this group to have a life of抽象 you're a life of抽象 so I think if you have a very successful CEO maybe a successful CEO but you're not a special CEO you're probably not able to do this you can't do this because this year people need to have a master so at the same time two people have a master is very difficult you have a product you can't do it you can't do it to perform this There is a change from it, and the human acts during this school trip, make no Bürger We don't want to kill a gorilla This means that we can do a brilliant vivre We must command it We must force the inspiration and for those who are Jack of Hmong You Jahak will have the owner of the CEO You are QUEthis guy Because I want to see it This is Jardin He is SpaceX Of, he is the为什么 An a something I want to think this Do you research over MBTI You覺得 MBTI With seconder artist Two people who have the equivalent MBTI I think what I mean to see, is to make people around people. that will be is very good so NTJ is a good combination because E and I I think it's not so important just you have enough of this to help you so I think NTJ means there's a sense and there's a plan and there's a logic and there's a理性 but I think it's not a good thing for the product it's very good it's a good feeling and it's a good feeling there's a good feeling So the天才产品经理 is INFP? Yes, NF is very important. But if you're a PR, it's more of a game. It's more of a game. If you're a J, it's more of a CEO. So the天才产品经理 can't be able to become CEO? It's pretty important. Because the天才产品经理, I've seen many of them are NFP. He can't be a plan with T. Yes, he has been a CEO. This is a team.但是我觉得N还是非常重要的当运营人才S我觉得是重要的但我对产品经理N和F是重要的尤其是N我一直认为抽象能力是非常非常重要的一件事情见创业者现在会问他的MBTI吗什么我会问他他会问我你会问他吗有的时候我有的时候会问一下比如你最喜欢的一本书或者你手机上一共有多少个APP对或者说问一下MBTI对就大家可以 聊一聊这个关于人本身的一些话题 MBTI会作为投资的标准吗 就不会 有什么见我一次就不会再见的创业者 其实我在这方面比较open 我不太会给人打一个特别标签 只要我有空 有的时候实在没空 也不是说就故意去不见人家 那我最后几个快问快答 一个全球范围内你喜欢的食物 如果只能是一个的话 是兰州牛肉面 I just don't know what to eat. I think it's the best food for the default. But it's very strange. I've been to the China-owned area, but I haven't been to the East. I haven't been to the East. But this food is my favorite food. The food is the most popular food. That's probably close to it. The food is the most accurate. What is the place you like in the entire area? I like the most popular food in the world. It's a very good food. 而且它能跟所有欧美的老城市相比的线下的体验,它可以是一个步行的城市。 一个少有人知道但必须知道的知识点可以是一个冷知识。 但我觉得这个可能不一定是冷知识,但这个东西对我来说可能是第一性原理的知识,就是幸存者偏差,历史是胜利者或者是幸存者输写。 出写的,不一定是胜利者,胜利者是幸存者,但是幸存者不一定是胜利者,但是它可以影响了未来的话语权,就是从这点出发,特别能理解很多历史参考文本的不可靠,因为所有的文本都是幸存者写的,所有给你讲述的知识都是幸存者给你讲述的,而幸存者是个必要条件,就回到刚才看人,人是个必要条件,他是一定会存在的条件, 但是我们很多时候要研究充分条件 我想研究预测性 而必要条件不提供预测性 所以一定要从一堆必要条件的文本当中 信息当中 能够找到当年真正他决策 真正那个的依据 你要知道真正那个时候 比如他有ABCDE 这种案件选项 每个后面有三个信息 最后他选择这个错误的东西 后来选择这个错误的 但是可能所有幸存者会把它加以戏剧化的描述 尤其是我这所有东西都会被因为人类喜欢戏剧化 对 所以怎么把它还原到当时我觉得这个是非常重要 还原到真正充分条件的时候 所以这是你要说有一个我觉得这个对我来说可能属于你这个定义的冷知识吧 就我看什么都是幸存的偏偿 你利用这个完成的一次思考是什么呀 I was in a company when I was in a company. There were many choices. I was thinking about how to choose. Maybe four months later I was able to know the answer. But at this time I was not sure. How can I get into that? The details are not open. I felt like I read the history. For example, you can read the history of the first two years. Actually, there are many many many. We read many many stories. 远超那里面非常戏剧性的描写 对 你还是要还原 所以我就得强调就是 心动力偏差对应的就是 你如何去还原 就包括Founder跟你说的事情 包括其他东西 其实每个人说是 都是心动力偏差 那还原到你投模型 最后是怎么做决策的 这个当然还有很多细节 但是我就说可能还是 我只是告诉你当时我的感受 就是我站在一个 充分性的推演的时候 实际上你找到各种各样的参考 很可能都是必要性的 这个是不能说的部分 接着所有读过的书 推荐两本必读书 我觉得两本书都影响很大 我觉得这两本书 似乎都可以花一年时间只看这本书 一本书叫《美国增长的起落》 这个我在很多场合 包括在红山会上都推荐过 它实际上描述了 整个美国在互联网时代之前的科技革命 叫《美国增长的起落》 And then another book I think can even take two years to read This book I really like It's called From China to the All-Star History This book is my mother's school in the New York Times This book I think it's a certain extent I think it's a lot of time I think it's a lot of time Because this book I think it's a lot of views I think it's what I always wanted to express I think it's in my future I think it's a lot of time So it's completely in a new way 重构了全球史 我觉得这本书非常非常值得看 重构中国出发的全球史 这是一本新书吗 是去年 我记得是去年中书的吧 美国增长起落 是一个经典的书 基于你当下认知一个关键 重要的当下的BET是什么 我当下的BET就是 现在这个时代 AI 的时代 就是华人大航海时代的开始 Actually, it is a trade between both Vietnamese and Chinese evolving earliest potatoes. To both overseas, they also split the neighbours into the revolution world. W Web, keep saying that for NATO business, Into the Estelle . What can I do to study as well? And then there is an investment about cyber-type culture. If you are looking atuthan AG product, there is a question related to what was compartmentalized to think. If there's these nuclear products produced by the consumers ? This could impact even monde- revenues. If we are Benedict after discovering AI skills in this field, how do we research fellows with high knowledge? What are these proyecto challenges throughout your world? First, the whole world of science will become more important. Because you ask AI, you still have to know what you're asking. So, you're asking how to ask. In this case, you must have a whole world of science. And those details are not important. You have a whole world of science. So, when you say AI, you have to learn the world of science. It's the world of science. It's the world of science. It's the best way to understand it. You don't have to understand it. But you know what to do in history, and what to do in which you are going to be able to do. So first, the teaching of the whole world will be very important. And second, the logic of the whole world will be very important. Because if people's natural language is a part of coding, then you should be more accurate to say natural language. It will become more important. Actually, natural language, especially in the Chinese, it's a very common language. It's accurate to the language that is more accurate. So you should be more accurate to say? This is clear that just about when the idea of the technology becomes an消費 system which will make your subscription because this is a thing intellectually Now that we're going to high-end teach girls and несколько podcast every