Hello Interconnected Readers:

I’m trying something new starting this week. In addition to the six stories summary that you’ve been getting, I will open this weekly issue with a more raw, personal, and reflective letter to you about something I thought about last week that stuck with me.

It’s part of my ongoing effort to feel more comfortable learning and living in public. I also hope this change triggers more responses from you! I always enjoy and learn from your replies and feedback.

A few days ago, I stumbled upon the transcript of an extended interview that Charlie Munger did with the Wall Street Journal in May 2019. Now, I’ve read most of Munger’s speeches and annual meetings over the years (many multiple times), so I was somewhat disappointed in myself for missing this one. It’s unique because Munger was loquacious, which is unlike the curt reputation he developed from years of Berkshire’s annual meetings, where he would usually follow Warren’s answers with “I have nothing to add.”

Charlie Munger, at his home in Los Angeles, discussed his philosophies on business, investing and life.

Whether you agree or like Munger or not, he always tosses out thought-provoking observations every time he speaks. The one that stuck with me from this interview is this passage:

“Part of the reason I’ve been a little more successful than most people is I’m good at destroying my own best-loved ideas. I knew early in life that that would be a useful knack and I’ve honed it all these years, so I’m pleased when I can destroy an idea that I’ve worked very hard on over a long period of time. And most people aren’t.”

He was speaking in the context of his architectural obsession designing college dormitories -- an experience where he self-admittedly realized after a year and a half that “every student ought to have his own bedroom.” (He subsequently called himself dumb for not realizing it for so long.)

I am a brief beneficiary of this obsession of his, having lived in Stanford's graduate student dorm that was also designed by Munger, where I got not only my own bedroom but bathroom too! Sadly, I didn’t benefit from thinking about the power of “destroying your own best-loved ideas” until now.

Being able to creatively destroy our own idea is rare, because it’s uncomfortable, if not downright against the inertia of human nature. Confirmation bias, sunk cost fallacy, and sheer delusion is everywhere and the default. Of course, if you can train yourself to do it, then you will stand out and be asymmetrically rewarded.

Everyone wants asymmetric rewards, but the number of people who can destroy their own best-ideas and develop a habit to do so regularly is probably a rounding error. That’s also why the reward is so asymmetrical, I guess. If lots of people can do it, the reward eventually evens out.

What are some of your best ideas recently? Tell me, and let’s see if we should destroy them!


The summary below covers the period between October 19 - 25, 2020 of six news stories – three from English language sources, three from Chinese language sources. Disclaimer: all translated article titles are done by me, not official translations from the media outlets.

Before you go on, please check out last week's deep dive post: "iPhone: 'Made in China -> India?'"

Don’t be blind to China’s rise in a changing world” (English Source: Financial Times Op-Ed)

My Thoughts: This Ray Dalio op-ed has stirred up a lot of discussions, and will surely be translated into Chinese soon. It is unabashedly pro-China, or at least not anti-China, which is bound to piss off many people. Dalio’s business and personal connection with China goes back more than three decades and has profited handsomely from that long period of engagement. While I don’t agree with the entirety of his views, I do think his general mindset to see China for what it is, not what others want it to be, is important whenever we think about US-China relations. At the end of the day, Dalio is an investor and wealthy investors' perspectives on subjects outside of investing tend to have more influence than they should. An investor’s job is to raise money from rich clients and (hopefully) return more money than they raised. Investors should not be tasked with making foreign policy, and foreign policy wonks would usually make bad investors. As Charlie Munger often admits, and did so again in his interview I cited above, he should never be tasked with doing brain surgery.

Introducing the Microsoft Azure Modular Datacenter” (English Source: Microsoft Azure Blog)

My Thoughts: As long time readers of Interconnected know, I write a lot about the cloud industry and can be a bit of a data center geek, so much so that I’ve written posts tracking the data center locations of all the major cloud platforms. Needless to say, I got really excited reading this Azure announcement of a modular data center the size of a trailer, with the same compute, storage, and networking (via satellites) functionalities of a standard cloud data center. Currently, building a data center takes a lot of land, access to cheap power, and fast underground or undersea fiber optic cables. These physical constraints mean accessibility to cloud computing is still limited. In fact, when I wrote “Is the Cloud Recession-Proof?” a few months ago, I called the cloud “part utilities, part commercial real estate, part railroad, and part something new.” If this modular data center works, it’ll increase access to technology for some of the most remote and harsh environments in the world. I’m personally excited for that future.

The first AI model that translates 100 languages without relying on English data” (English Source: Facebook AI)

My Thoughts: This multilingual machine translation (MMT) model released by Facebook’s AI research team last week is super cool. If it works as advertised, it’d be a huge step forward in not just AI but cross-cultural communication generally. The current state-of-the-art machine translation models are almost all English-centric, e.g. to go from Chinese to Spanish, the model goes from Chinese to English to Spanish. That reliance introduces accuracy problems and inherent biases in the English language. This new model is “many to many”, directly from one language to another without English. And as with most AI models, it is open-sourced! The current state of this model is still in research and not yet productionized in Facebook’s own product; getting there will take some more time. But I’m hopeful that by year 5 of writing Interconnected, I can expand to more languages and reach more people who don’t speak English or Chinese with the model!

Lufax updates IPO prospectus for US, expects to land on NYSE on October 30, sprinting to become the biggest fintech IPO in US” (Chinese Source: iponews)

My thoughts: Another week, another Chinese company plans to go public in New York. This time it’s Lufax, the fintech arm of PingAn Insurance, a Fortune 500 financial services conglomerate. According to Lufax’s prospectus, PingAn owns 42.3% while some of the other major shareholders are separate investment vehicles controlled by PingAn executives. Lufax’s core business is micro-loans and credits targeting SMBs and individuals, with approximately 13.4 million users so far. This IPO is not only significant in that it counters the narrative of “delisting”, its size may also make it the largest fintech IPO in US history, with its high-end fundraising target at $2.36 billion USD. The most analogous IPO would be Square’s back in 2015, when it raised $243 million USD.

New Oriental Sprints to HK Stock Exchange, Ahead of TAL Education, Becomes First Education Company to Return to HK Exchange” (Chinese Source: Sina Financial News)

My thoughts: And another week, another second listing in Hong Kong by a New York-listed Chinese company. This time it’s New Oriental, one of the OG education companies and one of the first to go public in the US back in 2006. This is now one of many second listings, after Alibaba, JD, and Netease. As we discussed a couple of weeks ago, BiliBili plans to do the same next year. There’s also rumor that Baidu may do the same later this year. (For readers interested in keeping track of HK secondary listings, here’s the search result on the exchange’s official website.) The US-China tension is definitely a factor contributing to this trend, but having a backup capital market access, while raising more capital in a world full of cheap capital in the more friendly APAC region, is just solid business opportunism. New Oriental could certainly use the capital to fuel new growth directions -- its so-called OMO (online merge offline) strategy -- as education becomes increasingly virtual. One last interesting note is the term that Chinese media uses to frame these second listings as “return” or “reunification” (回归) to the motherland -- the same term that was used to describe Hong Kong and Macau, when their colonial rules expired.

Chip Madness: Institutionals fight for deals, valuations soar, whose bet will be China's future chip giant?” (Chinese Source: Interface News)

My thoughts: This trend of private capital, both large institutional investors and VCs, flowing to any startup in China that has the word “chip” or “semiconductor” in its pitch deck is unsurprising. As China’s lack of good domestic chip technologies becomes evident and with the central government fully committed to changing that, there’s a cascading effect of private capital falling in line (or in lane) with what the government wants to see happen. But as I have written in “RISC-V, China, Nightingales” and “China’s Semiconductor Future: What Can $1.4 Trillion Buy?” (paywall; The Wire China), money itself can’t buy much in the world of semiconductor, if spent on the wrong people, doing the wrong thing, with the wrong timeline. As this article revealed, a lot of the money piling in are from VCs who’ve made money from China’s consumer tech growth and barely have enough patience for even B2B enterprise tech startups, let alone chip startups. That is an example of the wrong people, doing the wrong thing, with the wrong timeline.

If you like what you've read, please SUBSCRIBE to the Interconnected email list. To read all previous posts, please check out the Archive section. New content will be delivered to your inbox (twice per week). Follow and interact with me on: Twitter, LinkedIn.

Hello 《互联》的读者们:


这也是我不断努力进步的一部分,让自己更习惯在公开环境下学习和生活。我也希望这个尝试能够引发读者更多的回应! 我非常喜欢读大家的各种回复和反馈,并在反馈中学到了很多。

几天前,我偶然发现了Charlie Munger在2019年5月接受《华尔街日报》的一次长篇采访的记录。过去我已经读过Munger多年来几乎所有的演讲和年会(很多都是多次阅读),但唯独错过了这次采访,所以对自己还是有点失望的。这次采访的独特之处在于Munger的滔滔不绝,这与他多年来在Berkshire年会上塑造的古板名声截然不同。一般在年会上,他通常只是在巴菲特回答完问题后说,"我没有什么要补充的。"

Charlie Munger, at his home in Los Angeles, discussed his philosophies on business, investing and life.


"我比大多数人更成功的部分原因 是我善于摧毁自己最喜欢的想法。我很早就知道这将是一个有用的技能,这些年我一直在磨练它,所以当我能摧毁一个我长期以来非常努力在想的想法时,我很开心。而大多数人却不是这样的。"

他说这段话时是在聊他痴迷于设计大学宿舍的建筑。他自认花了一年半的时间才意识到,"每个学生都要有自己的卧室 "。(后来他说自己真是个傻子,因为过了这么久都没有意识到这一点)。

我算是他建筑设计的短暂受益者,曾住过他为斯坦福大学设计的宿舍,不仅享受到了独立卧室,还有独立卫生间! 遗憾的是,直到现在我才从 "摧毁自己最喜欢的想法 "这一思考中受益。






还希望大家有空看看我上周最新的深度分析文章:《iPhone: "中国制造->制造印度?"

中国在全球变化中的崛起,不要视而不见”(英文来源: 英国金融时报 )

我的想法: Ray Dalio这篇专栏文章引起了很多人的讨论,也肯定会很快被翻译成中文。它的观点是不折不扣的“亲中”,或至少不是“反华”,这必然会惹恼很多人。Dalio与中国的商业和个人关系可以追溯到三十多年前,并且他在长期的交往中获利颇丰。虽然我并不同意他的全部观点,但我认为,无论何时,只要在思考中美关系,他的总体思路 - 即客观的看中国是什么,而不是他人希望它成为什么 - 还是很重要的。归根结底,Dalio是一个投资人,而有钱的投资人对投资以外话题的看法的影响力往往大于它们应有的影响力。一个投资人最终的工作就是向有钱的客户筹集资金,投资后(希望)还给客户更多的钱。投资人不应该制定外交政策,就像外交政策专家通常会成为很糟糕的投资人一样。正如Munger经常承认的那样,在我上面引用的采访中也再次承认的:他永远不应该给任何人做脑外科手术。

微软Azure模块化数据中心介绍”(英文来源: 微软Azure博客)

我的想法: 长期看《互联》的读者都知道,我写了很多关于云计算行业的文章,也可以说是个数据中心的geek,以至于写了很多跟踪各大云平台数据中心地理位置的文章。读到Azure这篇宣布模块化数据中心的时候,我真的很兴奋。这个数据中心只有房车那么大,但配有和标准云数据中心一样的计算、存储和网络(通过卫星)功能。目前,建造一个数据中心需要大量的土地,廉价的电力,以及快速的地下或海底光缆,这些限制意味着可以使用云计算的地方仍然有限。我在几个月前写的《云计算是否能抵御经济衰退?》这篇文章中,就把云计算称为 "既像水电,又像商业房地产,还像铁路,同时甚至是一个全新的事物"。如果这种模块化数据中心能成功,它将使世界上最偏远、最恶劣的一些地方得到云计算的资源。我个人对这种未来非常期待。

首个不依赖英文数据翻译100种语言的AI模型”(英文来源: Facebook AI)

我的想法:  Facebook的AI研究团队上周发布的这个多语言机器翻译(multilingual machine translation,MMT)模型非常酷。如果真的那么好用的话,那它不仅在人工智能方面,而且在整个跨文化交流方面都是一大进步。目前最先进的机器翻译模型几乎都是以英语为中心的,比如要从中文翻译到西班牙语,模型必须要从中文翻译到英文再到西班牙语。这种对英语的依赖不仅对准确性有影响,也会有英文固有的偏见。而这种新的模式是 "多对多"的,直接从一种语言翻译到另一种语言,中间不需要英语。和大多数人工智能模型一样,它是开源的! 目前这个模型还在研发状态中,没有在Facebook自己的产品中生产化。要达到生产化的状态还需要一段时间。但我希望在写《互联》的第5年时,能够用这种模型把写的东西扩展到更多的语言,给更多不懂英语和中文的读者!

陆金所更新赴美IPO招股书,预计10月30日登陆纽交所,冲击美股最大金融科技IPO”  (中文来源: 独角兽早知道)

我的想法: 新的一周,又有一家中国公司计划在纽约上市。这次是世界500强金融服务集团,平安保险,旗下的金融科技平台,陆金所。根据陆金所的招股书,平安拥有42.3%的股份,而其他一些主要股东则是由平安高管控制的其他投资体。陆金所的核心业务是针对中小企业和个人的小额信贷,至今拥有约1340万用户。此次IPO的意义不仅在于继续对"退市"的说法唱反调,陆金所的规模也可能成为美国历史上最大的金融科技IPO,高端募资目标为23.6亿美元最类似的IPO应该是Square早在2015年的上市,当时募资额是2.43亿美元

新东方冲刺港交所:抢先好未来 成最早回归港股的教育企业” (中文来源: 新浪财经)

我的想法:  新的一周,又有一家在纽约上市的中国公司在香港二次上市。这一次是新东方,老牌教育公司之一,也是最早在美国上市的中国公司之一(2006年)。这也是继阿里、京东、网易等之后,又一个二次上市的公司。正如我们几周前讨论的那样,B站也计划在明年这样做。还有传言说,百度可能在今年晚些也会去港股二次上市。(对于有兴趣跟踪香港二次上市的读者,这里是交易所官网对这类上市的搜索结果。)中美关系紧张必定是导致这一趋势的因素之一,但准备好后备的资本市场通道,同时在一个充满廉价资本的世界里,在更友好的亚太地区筹集更多的资金,也是应当争取的商机。随着线上教育越来越普遍,新东方绝对对资本有需求,来推动其新的增长方向--也就是所谓的OMO(线上合并线下)战略。最后值得一提的是,国内媒体将这些二次上市的公司定义为 "回归“ -- 就和当年形容香港和澳门的殖民统治到期时的说法一样。

疯狂的芯片:机构哄抢,估值暴涨,谁能赌到中国未来的芯片巨头?”  (中文来源: 界面新闻)

我的想法: 私有资本,包括大型机构投资商和风投机构,正在涌向国内任何一家在投稿中带有 "芯片 "或 "半导体 "字样的创业公司。这种趋势并不奇怪。随着中国缺乏优秀的国产芯片技术的现状变得日益明显,加上中央政府全力想改变这一状况,民间资本自然会受到连带效应,与政府的期望(或指定赛道)看齐。但正如我在《RISC-V、中国、夜莺》和《中国半导体的未来:1.4万亿美元能买到什么?》 (需在The Wire China网站付费看)这两篇文章中写到的,在半导体的世界里,如果钱花在了错误的人身上,用来做了错误的事,有着错误的时间线,再多也买不到什么。正如这篇文章所揭示的那样,很多涌进来的钱都来自于那些从中国to C科技发展中赚到钱的风投,它们连to B企业服务创业所需要的耐心都没有,更别说芯片创业了。这就是一个错误的人,做了错误的事,有着错误的时间线的例子。