Matt's Thoughts In Between - Issue #85

Forwarded this email? Subscribe here. If you like it, please forward it to a friend.

The remarkable transformation of China's elite

China’s economic transformation over the last three decades has been astonishing - and some of its implications are understudied. This fascinating new(ish) paper by three economists looks at the impact on the sociology of Chinese elites. As the authors note, the increase in GDP per capita that China experienced between 1988 and 2013 took the West 200 years to achieve, which allowed a long period of social adjustment. Imagine the whiplash of two centuries of economic change in just 25 years.

The big stories are the Chinese Communist Party’s (CCP) accommodation of capitalists and rocketing inequality. In 1988, the makeup of the Chinese elite (defined here as the richest 5% in urban areas) shifted from government officials to professionals and business owners. The paper suggests that the income premium associated with CCP membership has slowly declined - except among the owners of the largest businesses, where it's increased dramatically. This co-option has political consequences: new elites are less inclined to fight the system if they’re already part of it.

It’s interesting to read this alongside this new paper by Shruti Rajagopalan and Alex Tabarrok, which looks at the behaviour of Indian elites. Their thesis, summarised here, is that Indian elites often take their social/cultural cues from Western elites, which sometimes leaves the country with more regulation than it has capacity to enforce. The contrasting paths taken by India and China represent one of the most underrated threads in understanding the world today.

How "open source" chip designs might change the world

The Economist has a very good series of articles (here, here and here) on open source designs for microprocessors. The author has a good summary thread here. The open source software movement has had extraordinary impact, so the expansion of similar principles into the world of semi-conductors is fascinating.

As the Economist notes, there are two important implications - one commercial, one geopolitical. First, chip design is currently dominated by Intel and Arm, so open source designs like RISC-V could open up innovation and competition. There’s already movement: Alibaba announced its own chip earlier this summer and India is making bets in the space. Most interesting are the implications at the bottom of the market. While Arm's licensing fees are tiny relative to the cost of an iPhone, they’re large in the context of mass devices, so open instruction architectures could enable new products.

Arguably more consequentially, RISC-V decouples the global chip supply chain from Western-owned IP (i.e. potentially subject to export control). As discussed previously, China has come to see its lack of a domestic semiconductor industry as a national security risk - and we’ve seen in Huawei how vulnerable Chinese hardware companies are to US pressure (Relatedly, it’s worth reading Jeff Ding’s recent translation of a deck on 5G that has gone viral in the Chinese tech community; commentary here). It's going to be an important space to watch.

Silicon Valley sours on IPOs... and Softbank

Silicon Valley is turning sour on IPOs - historically the most sought after exit event - in favour of "direct listing". In a traditional IPO, a company pays fees (sometimes egregious - one of my all time favourite blogposts) to investment banks to underwrite the issue of new shares. In a direct listing, no new shares are issued; shareholders simply sell their stock directly to the public. The tide has shifted rapidly. VC extraordinaire Bill Gurley is vocally calling for direct listings (see this excellent podcast for his case) and Airbnb, one of the most valuable private companies, has indicated it will take this approach.

This is happening against a backdrop of a tough year for IPOs. A number of heralded startups, such as Slack, have had disappointing openings on the stock market, while WeWork’s disastrous attempt to go public (see here for previous coverage) resulted in its CEO stepping down and, according to some analysts, the beginning of the end for the company.

That’s not to say that direct listing would have saved WeWork - or any other underperforming stock. Price discovery works the same however you go public, as Fred Wilson notes. The most interesting victim of the public market’s ruthlessness may be Softbank’s Vision Fund (see previous coverage), which may struggle to raise a second fund with both WeWork and Uber underwater. But given their bizarre structure - they will draw down capital from investors to… pay those investors a dividend! - perhaps IPOs are the least of their problems.

Quick Links

Where the magic happened. Great Q&A thread on the most productive science labs of the 20th century.
As fast as possible, but no faster. Striking animation that illustrates the speed of light.
Judging covers by their books? Another great Q&A thread on "What 3 books should someone else read if they want to understand the real person inside you?"
Machine learning meets Aristotle. Tantalising story on a breakthrough in reading fragile scrolls without opening them.
The meta-content you crave. When did charts become popular?

Your feedback

I'm falling slightly behind on my (arbitrary) 1% growth a week target, so I'd love to get 30 new readers this week. If you like Thoughts in Between, please consider forwarding to a friend or two who might enjoy it too.

I'm always happy to get replies if you have comments or feel free say hi or follow me on Twitter.

Until next week,

Matt