Looking East

Silicon Valley does need to learn lessons from Chinese tech: just not the ones that Mike Moritz thinks are important

This last week Silicon Valley has been abuzz with an article penned by legendary Sequoia Capital VC Mike Moritz. Titled “Silicon Valley would be wise to follow China’s Lead,” Moritz details how a recent trip to China highlighted the differences between Mainland Chinese tech culture and Silicon Valley — and ultimately why the former contains elements for success that the latter would be wise to follow.

While in the general case I think this is true, the content of Mike Moritz’s editorial has ignited a firestorm of controversy. Moritz contends that the Chinese tech worker is more successful because of their superior “work ethic”, citing the following as evidence of a serious divide in dedication between engineers in the PRC and staff in Silicon Valley that lead to the former being better positioned for success than the latter:

• Chinese tech workers work longer (“Here, top managers show up for work at about 8am and frequently don’t leave until 10pm. Most of them will do this six days a week — and there are plenty of examples of people who do this for seven.”) to the point that they willingly accept “unhealthy” work environments and “work with a “disregard paid to physical fitness.” Mike particularly highlights the difference between Silicon Valley engineers and their counterparts in Shanghai and Beijing here: “[Chinese] Engineers have slightly different habits: they will appear about 10am and leave at midnight.”
• Chinese tech workers are willing to put their job and their company before their family. Workers do not question (or even seek) “the appropriate length of paternity leave or work-life balance,” and Moritz praises this dedication: “such high-flyers only see their children — who are often raised by a grandmother or nanny — for a few minutes a day.”
• Chinese tech workers are frugal and more focused. “You don’t see $700 office chairs or large flat panel computer screens at most of the leading technology companies,” he writes. “Instead, the furniture tends to be spartan and everyone works on laptops.” Mike also notes that Chinese engineers don’t need to worry about having “jam sessions,” and other luxuries that their American counterparts employ.

As a result of such contentions, Mike’s commentary has invited tsunami of repsonses across the Silicon Valley and SF tech community.

Unsurprisingly most of the response is negative. Many cite that Mike’s comments on paternity leave and seemingly slavish dedication run contrary to empirical evidence of the success of maintaining work life balance. Others note that Mike himself (an investor who has never actually worked in Silicon Valley as an operator and joined Sequoia after being a journalist) seems out of touch with the brutal realities of Silicon Valley’s already-imbalanced work live balance, and that his post is clearly just an attempt to “win points” with Chinese tech firms that Sequoia is courting.

This isn’t to say that all of the response to Mike Moritz’s contentions are negative. I’ve had a series of conversations (mostly with other VCs) who are supportive of at least Mike’s possible intentions with this post. Some Chinese-American entrepreneurs have also come out to defend Mike’s commentary, noting that the work ethic and dedication in Chinese tech is an admirable result of habits in Chinese millennials who are the sons and daughters of those who lived through the Cultural Revolution.

I’ve spent a good amount of time working with Chinese tech companies. As an associate at GGV Capital with a unique technical background, I had the opportunity to work very closely with Chinese companies such as YY and collaborated with my colleagues in Shanghai on evaluating investments in cloud infrastructure, enterprise tech, and gaming.

Because of the rise of enterprise and cloud infrastructure in China in the last decade, much of my operational career outside of VC has also involved working with Chinese high tech companies. As NetApp’s product manager for cryptography and compliance, I worked closely with NetApp co-founder James Lau and our head of corporate counsel on evaluating legal issues and compliance issues around selling our products with security features in China. At HashiCorp I also work on similar issues.

After Moritz’s article came out, I’ve had a series of conversations with friends and old colleagues of mine who have similar — and frequently better — experience working in and around Chinese tech society. After all of this (and a few nights sleeping on the issue) I feel confident in the following:

Mike is right that Silicon Valley can, and should, learn lessons from Shanghai and Beijing tech cultures. But none of those lessons have to do with the unhealthy dedication to work he praises in his article.

In fact, Mike Mortiz’s article instead highlights deep informational divides about engineering and product development between non-operator, non-technical VCs and the Silicon Valley engineers who work at companies funded by them.

Context is for Kings

Before we get into all of this, it’s important to set the context for this article. Mike Moritz’s article comes at a time when it feels like Western tech culture (particularly Western VCs) are just starting to appreciate the value of Chinese technology companies and startups.

Commentary about Chinese tech from Western investors has spiked over the last 18 months, and reflections on Chinese tech from investors like Moritz or Jason Lemkin’s commentary below are becoming more common.

This is not serendipitous. Most venture investors, really all of the good ones, use social media and personal branding as mechanisms to source and better competitively position themselves when competing against other VCs.

I did it too. As an associate at GGV I wrote articles in VentureBeat talking about my perspective as an ex-engineer in VC. As a principal at Amplify I wrote articles about my experiences in security and cryptography and ultimately uploaded my full investment thesis.

Content marketing isn’t nefarious. Cultivating a personal brand and strategically developing content is an important part of the job in modern venture capital. It helps entrepreneurs to get to know “you” and what you’re interested in, and improves the efficiency of sourcing, dilligencing, and ultimately funding tech startups on both sides of the table.

The rise of western VC content marketing on China is ostensibly a response to the extremely price and sourcing-competitive nature of Silicon Valley tech — and a time of uncertainty for the asset class of venture capital as a whole.

According to Cambridge Associates, US VC performance (a market dominated by Silicon Valley venture) in 2017 was “middling” compared to its other counterparts in private equity. US VC funds within the CA index in 2017Q2 returned only 4.7% YTD — almost half of what similar money invested in other areas of Private Equity would have made (7.6%) or even investment in the NASDAQ Constructed mPME.

Performance like this has anecdotally given pause to LPs and fund of funds who supply venture firms with the capital to invest. There is a concern that larger institutional LPs may revise their allocations in VC in the upcoming next few years. For VCs looking to raise money (or more in the case of Moritz, a largely non-investing partner at Sequoia who presumably is incentivized for the fund to just continue to succeed), it is very important to find ways to beat this “middling” average and return far north of the 2.6x CoC return norm.

The best way to do that has been to look outside of Silicon Valley for new investments. Areas like Seattle, New York, Austin, Tel Aviv, and the UK have sparked with venture investment from Silicon Valley firms looking to find less competitive markets with better arbitrage opportunities in local startup valuations.

China is one of those places. While startup enclaves like Beijing and Shanghai already have local VCs with deep pockets, the combination of a potentially less competitive market for valuations (Chinese guanxi business culture has much more wiggle room for valuation than Silicon Valley’s multiples-based approach given high NTM TEV/REV in US tech markets) with mature, capable technology has been very attractive for US investors who look to wield their own deep pocketeted-funds and attract startups with favorable content marketing and promises to help shepherd their entrance into US markets.

There’s just one problem with this recent Chinese enthusiasm for Western VCs new to China: China is not Silicon Valley, and not respecting the major economic and cultural differences of the two in comparing each’s tech cultures leads to erroneous assessments like those found in Mike Moritz’s article.

Comparing Apples to Mandarin Oranges

When I first started working with Chinese tech companies, I began hearing a phrase that has been repeated in a variety of different forms over the years:

China is not the United States.

You don’t need to speak Mandarin fluently to understand this, but I do think you need to visit China and spend time working with Chinese companies to appreciate this difference first hand. I’ve spent time writing about this in the context of why Silicon Valley tech companies struggle when entering China, but the same can be said for VC: China is simply a different world.

Major differences in population (New York has the population of a tier 2 Chinese city) and a very different political, social, and economic culture lend themselves to a very different components for success in China. Western failures to appreciate, and most importantly respect, these differences can and do lead to failure.

Mike Moritz’s article seems to tread a familiar path in not respecting these cultural differences. For example, while he praises the insane work culture of Chinese tech and notes it has to do with something culturally Chinese, he fails to appreciate that it is not an enviable part of Chinese culture and instead is a fatalistic holdover from aspects of the cultural revolution.

Many Chinese tech workers aren’t working these hours because they want to — they’re working them because they feel/actually have to in order to adequately support their families. As one of my friends from Shanghai noted, Chinese families who survived the Cultural Revolution have embraced the need to “eat bitterness” and sacrifice for their families to survive.

If you read Mike’s article, you might think that the Chinese software engineer willingly embraces the 70–100 hour work week and casually leaves their family behind. In reality, many Chinese engineers work these hours because these are simply the table stakes of their field and they need to work them in order to help their family prosper.

They are eating the bitterness of the environment, and trying to attribute their success by simply comparing hours worked to their American counterparts both fails to appreciate these cultural differences and is even somewhat insulting in its lack of respect for the major cultural differences and differences in history of industrialization in China and the United States.

For many of my Chinese friends who worked in China and had a visceral response to Mike Moritz’s article, this is what set them off the most. Sacrificing personal health and relationships with their children and spouses is by no means enviable, and Mike Moritz’s tone deaf praise of such behavior fails to appreciate that Chinese history and that Chinese tech is an accomplishment in spite of such adversity rather than a consequence of it.

Mens Sana, C[ode] Sana Est.

Furthermore, while Mike Moritz was willing to comment on how perceived cultural differences between Silicon Valley engineers and their more frugal and seemingly harder working counterparts in China led to success, Mike fails to review why this is the case.

Why, for example, when we’re using the same programming languages and frequently the same software engineering methodologies is Chinese tech superior to its Silicon Valley counterparts? What variables in Chinese software development are different — is it just the difference in hours? Are there differences in software design methodologies?

The reason why Mike doesn’t go into such detail in his writing here is simple: he doesn’t know.

While Mike Mortiz has had a legendary history in venture capital investing, he’s never been a software engineer or physically written and shipped production code. Prior to joining Sequoia, Mike Moritz was a writer for Time Magazine who covered figures like Steve Jobs.

For what it’s worth this isn’t a bad thing. There are serious contentions about whether or not modern VCs need to have operational experience. But as a VC with operational experience, I don’t think you do. I’ve met wonderfully successful VCs (even those who invest in heavily technical areas) who have never compiled “Hello World” and have been an elemental part of advancing high tech through their investments.

That being said, the moment those individuals start talking about engineering culture they wade into an area where they don’t have personal experience. At best, non-operational VCs without experience building and shipping code can highlight the experiences of others who do have operational engineering experience in citing their commentary.

But you run a dangerous line in trying to make strong contentions about the practice of software engineering without actually having worked in the field, and here again Mike Moritz’s rush to conform his limited exposure to generaization has led to error.

To make a very long story short: software engineering is weird. Despite the practice being nearly four decades old, we still struggle to find good ways to quantify and streamline the field — even from the inside.

But while we’re still figuring out how to operationalize and improve effectiveness software enigneering, we’re very aware as an industry and culture of ways to hurt our effectiveness. I’m reminded of a great blog post from an early Twitter engineer on his experiences watching the company deal with scaling and growth per-IPO:

I think a big part of the problem is that we — as an industry — are not very good about thinking about how to make engineers effective. For our software, especially back-end software, we can measure its goodness by the number queries per second it can handle, the number of incidents we experience, and the amount of hardware we have to buy to run it. Those things are easy to measure and even fairly easy to tie to financial implications for the business.

Engineers’ effectiveness, on the other hand, is hard to measure. We don’t even really know what makes people productive; thus we talk about 10x engineers as though that’s a thing when even the studies that lead to the notion of a 10x engineer pointed more strongly to the notion of a 10x office.

But we’d all agree, I think, that it is possible to affect engineers’ productivity. At the very least it is possible to harm it.

One area where we know effectiveness gets harmed is in overwork. Producing quality, production code is not simply a sack race, and forcing engineers to work longer hours with the idea that product quality and features scale linearly with time worked is a frequent misunderstanding from laymen who do not have experience working in tech.

There have been a number of studies into the ineffectiveness of overwork in software engineering. One area notorious for overwork leading to problems is in video game development, where the infamous “crunch” of a pre-release breakneck sprint of work can lead to major problems in game quality on release.

In 2006, a post from the spouse of an engineer at Electronic Arts sparked a controversy about the ineffectiveness (and unhealthiness) of crunch. This sparked a study by economists and computer science grad students at Stanford, which yielded strong empirical evidence of the dangers of “crunch” engineering.

The result of their analysis was simple: engineers working too long started to see deminishing returns on code quality and production. As the project discovered:

Thus, overworked employees may simply be substantially less productive at all hours of the work day, enough so that their average productivity decreases to the extent the additional hours they are working provide no benefit (and, in fact, are detrimental).

Further studies into labor economics and software engineering have revealed similar results into the long-run detriment of employing “crunch” development. According to a paper submitted to the IGDA analyzing Stanley Chapman’s work in labor economics to software engineering:

Chapman’s diagram of the work curve assumes that a working day of a given length is maintained over a considerable period of time. Thus it incorporates both simple and accumulated fatigue into its model. At first the declines in output per hour simply reflect the effects of fatigue on both quantity and quality of work performed toward the end of a given day. But eventually daily fatigue is compounded by cumulative fatigue. That is, any additional output produced during extended hours today will be more than offset by a decline in hourly productivity tomorrow and subsequent days.

Studies like these have provided empirical insight into why software engineering effectiveness does not linearly scale with hours worked. But experience programming also yields similar results.

Sometimes you run up against difficult problems that you can’t solve on a first pass. After a while, working more to try to solve that problem is counter-productive. The common response to these problems is simple: take a break. Step away from your computer and your copy of the CLRS, and do something to take your mind away from the problem so you can come at it fresh and rejuvinated.

The importance of small breaks and a positive work environment where engineers are not overstressed or overworked cannot be understated. This is why places like Google and Facebook are bastions of engineering innovation: they recognize why investment in engineer well-being is elemental to innovation and success, and are willing to invest millions in things like rooms for “jam sessions” because they will receive billions in return on new products and services built by their happier, more effective engineers.

Again, Mike Moritz’s lack of deep personal experience and generalizations about his exposure to the subject have led him astray. You can’t just force an engineer to work more hours to be effective.

So if that’s the case, why is China technology so successful? Why have companies like Tencent seen such roaring success both in China and abroad.

Again, the devil is in the details — details Mike Moritz unfortunately doesn’t go into in his article.

To the Stars, Through Adversity

During one of YY’s visits to the United States, I had an opportunity to spend time with their senior leadership team as they attended a series of roadshow meetings with industry analysts in preparation for their IPO.

I was astonished to learn how advanced YY’s technology was compared to the norm of Silicon Valley infrastructure tech at the time. While Silicon Valley was just being introduced to mainstream distributed computing, YY was using concepts in grid computing and graph-theoretic workload optimization to streamline synchronous communication between their users. We are just now starting to focus on gossip protocols and decentralized computing in Silicon Valley with the rising interest in cryptocurrencies and blockchains. But these were table stakes areas of tech for YY since the late 00’s.

When I asked how YY was able to pull off creating and integrating such advanced infrastructure technology so early, I remember YY’s CTO shrugging and simply noting that it was simple: “we have to.”

China’s advances modern consumer internet is ridden on the back of a surging explosion of domestic infrastructure tech. In a country where major portions of the population sit in areas with unreliable connectivity, Chinese tech companies have had to innovate radical solutions in order to serve what we consider normal levels of service in connected applications.

Even more impressively, the scale at which Chinese technology companies (even new startups) operate on is almost unfathomable to most US startups. In a country where there is easily more than a billion people than the total population of the United States, growth-stage Chinese startups regularly deal with user traffic on daily volume of YouTube. Maintaining autoscaling and performance optimization to deal with the deluge in traffic is critical — and normal — for Chinese engineering organizations.

Chinese technology firms’ history of innovation through adversity is not simply due to the challenging scale of users in China. The West has historically restricted access to advanced computing and networking technology sent to China, such as when Intel was precluded form selling high end Xeons to Chinese enterprise tech companies in 2015.

In response, many Chinese technology companies began to invest in distributed computing over GPUs. Suites like Nvidia’s CUDA gained acclaim in China for their general purpose compute properties. And as a result, China quickly became the home for most Bitcoin mining as the country already had a GPU/Cell-computing expertise due to their need to solve HPC problems that otherwise couldn’t be addressed with using US-banned high performance traditional processors.

As China and Austria recently celebrated the launch of the first quantum key distribution system in production (a poignant advance given the US’ history of restricting cryptographic technology outside of their partners in NATO) it is once again made clear that Chinese tech is built from a hacker culture of innovate or die.

TL;DR

There are a lot of things Silicon Valley can learn from Chinese tech culture. But in learning from Chinese tech we need to appreciate why Chinese tech is successful and innovative.

Chinese tech is successful because Chinese technology companies have been forced to innovate or die. In a world where “Facebook”-level scale issues are normal, Chinese engineering culture has had to develop radical solutions just to survive. As a result, successful Chinese technology companies harbor a laser focus on innovation and technical substance. Disengenuous hype with minimal technical substance, such as what we’ve seen in Silicon Valley around Bitcoin, is downplayed in importance or discarded.

But contrary to what Mike Moritz has contended in his review of Chinese tech, Chinese tech is not successful because of difficult work environments. As we have learned ourselves throughout the 40 year history of software engineering in the US, simply adding on the hours for engineers does not yield a better product.

Silicon Valley engineers are not lazy, spoiled, or entitled: they are simply different than their Chinese counterparts. If we want to learn from each other, we need to respect those differences and appreciate them.