When the world's top economic think tank was located in Redmond, Washington

diff.substack.com · by Byrne Hobart

In this issue:

  • When the world's top economic think tank was located in Redmond, Washington

  • Clubhouse and Clubhouses

  • Private Planes Go Public

  • More Decarbonization

  • Supply Chains as a Hostage Negotiation

  • Coasian Parliament

"Think tank" is a relatively modern formalization of something that's always existed: semi-organized efforts to think on behalf of powerful people. Usually, the more important your job is, the less time you have to carefully craft some written principles behind what you do. (There are exceptions, of course, but among politicians it seems rare but not unheard of for them to write their own memoirs.) But it's hard to scale leadership without some effort to systematically write things down, so leaders with good instincts leave a trail of ex post explanations written by other people in their wakes.

Since think tanks are a new term for an old phenomenon—the wikipedia article has examples from a millennium ago but dates the modern usage back to the 1980s—it's easy to miss some think tank-like entities that somehow escaped the characterization. One example: Microsoft in the 1990s. Sriram Krishnan's corporate memo collection is full of Microsoft memos

  • There's this pitch to Warren Buffett on why Microsoft was a buy in 1997, for example—a compelling look at why it's good to own a low-marginal-cost essential complement to a rapidly-growing commodity. It's also a good piece on how to bet given a narrow but significant information advantage. (Raikes notes that he had 90% of his net worth in Microsoft, because he understood the business and knew its advantages would persist for a while.)

  • There's this memo from May 1995 on why the web is an essential platform[1], with a strategy for Microsoft to dominate it. It's aged incredibly well: among other things, the piece argues that Microsoft should offer the same products in CD-ROM and downloadable form; part of the bull case for GameStop early in its run was that the massive file size of modern games made downloads inconvenient for people with poor connections, so some software businesses still need multiple distribution modes.

  • Also in 1995, Bill Gates marveled that it was easier to find information on the web than on Microsoft's internal network. This is an early look at a case where decentralized platforms can aggregate more effort than centralized ones, even though centralized ones are more focused on the end goal.

  • This Nathan Myhrvold essay from 1993 is mostly far-out futurism (including estimates for how much bandwidth it would take to transmit smell and touch), but it includes a very accurate assessment of what the Internet will do to various kinds of media—books will be fine, but newspapers, with their reliance on classified ads, will be decimated. (In 1993, there were around 5m Internet users, and the count was growing 30% annually. Even to people who had heard of the Internet, the long-term implications were not obvious.) It also features accurate predictions about finance ("Once you eliminate the exchange and notion of buying or selling through a broker, most of the other Wall Street jobs that remain will still be in high demand."), retail ("Sam Walton on steroids"), and video ("The whole notion of a channel simply disappears").

  • In 1994, Steven Sinofsky wrote a memo about how networked computing had completely changed Cornell in just six months. New forms of socializing often drive the fastest adoption, especially among younger people, so the piece gives a great deal of attention to email.

These memos are amazing cultural artifacts, because they articulate a very accurate vision of the future, very early. In retrospect, it's astonishing that Microsoft wasn't more dominant in the late 90s, since its senior executives were so aware of the most significant change happening in their industry at the time, and since the company had unmatched distribution capabilities. It would be surprising to see a trove of internal memos from the Union Pacific Railroad in 1900 accurately predicting how the car would change the world, or from Nokia in 2004 explaining exactly how the smartphone would alter human behavior. They're surprisingly broad-based; there wasn't a reason Microsoft employees should be especially aware of how stock trading works, or what kinds of ads dominate newspapers' revenue.

At many companies, there's internal resistance to talking seriously about small but rapidly-growing competitive threats. They generally look like toys, and they don't seem commercially viable. It's low status to worry about them. Somehow, Microsoft avoided much of this internal signaling problem, and applied plenty of brainpower to explaining how dangerous those toys were.

Why did this happen? Microsoft hired almost purely for intellectual firepower. Early Bill Gates interviews make references to hiring for intelligence and desire to learn, without much talk about being a team player. "The key for us, number one, has always been hiring very smart people. There is no way of getting around, that in terms of I.Q., you've got to be very elitist in picking the people who deserve to write software. Ninety-five percent of the people shouldn't write complex software." Hiring for experience can be a good model, but the younger an industry is, the shorter the half-life of experience—and in a winner-take-all business, the most valuable experience by far is the kind you get only if you're the winner. For a company like Microsoft, the growing complexity of their products also placed a premium on raw talent over accumulated knowledge: they had to hire people who could learn to navigate the Windows 95 codebase as it existed then, not people who could figure out the theoretically ideal way it should have been designed.

Hiring smart people over specialists means hiring people who can apply their talents to whatever they happen to be interested in, and hiring for openness over agreeableness means hiring people who will compete to be the first to know about some amazing new trend. Paying people in options means having an employee base that's very concerned with what could happen to their net worth if the company is blindsided (and diversification wouldn't necessarily help; a Microsoft employee who sold their stock to buy a house in Redmond was trading a direct long-MSFT position for a synthetic one).

The reason we have these memos is also the reason they aren't being produced any more: antitrust. Economic theory is all well and good, but antitrust law is not based on cutting-edge analyses of competitive advantage. It's based on precedents set the last time a major company got taken down. And modern antitrust cases, like other complex white collar criminal cases, rest heavily on determining mens rea from internal emails. Microsoft was an early adopter of email as a communications medium, which was a great way to coordinate a business whose key employees all seemed hyperlexic (Bill Gates would take stacks of books on short vacations). That meant it was a late adopter of modern corporate email norms, in which certain economic terminology is discouraged. Microsoft survived its antitrust scuffles well enough that many of the people involved could retire if they wanted to, but it also developed enough scar tissue to make freewheeling internal communication anathema. These memos are oddly similar to cuneiform tablets, in that they were meant to be temporary, got added to the historical record due to a conflagration painful for everyone involved, and now provide detailed documentation of an interesting time that's very much over.

It's a lucky set of circumstances. A combination of industry tailwinds, a distinct hiring strategy, the rise of email before the rise of attorneys who suggest policies around email, and a big antitrust case have given us a unique corpus of forward-looking thoughts on how the Internet affects everything. And because they were looking far ahead with respect to one technology, which hasn't been supplanted in the economy or popular imagination by any other similarly impactful technology, they're still relevant today.

[1] Intriguingly, the memo gets one detail almost right, worrying that $2,000 desktops will be replaced with a $500 stripped-down device that integrates with an existing, widely-deployed consumer electronics product. In 1995, the worry was that it would be a TV, but it turned out to be a phone.

This piece arose from some helpfully detailed reader feedback on this earlier issue on newspapers.


Clubhouse and Clubhouses

A longstanding financial tradition, slightly less classy than a SPAC, is to take a public company that doesn't have any operations, change its name, and then merge it with a private company in a more hyped sector. This is a way to take a small company public, albeit as a penny stock. (You can trace historical hype cycles by looking at a company whose SEC filings include a line like "Formerly New World Entertainment Corp.; Formerly Morningstar Industrial Holds Corp.; Formerly Katie Gold Corp"). One recent example of this is Tongji Healthcare Group, which recently acquired a group of TikTok "hype houses," and changed its name to Clubhouse Media Group.

Its timing was fortunate, since Clubhouse the company raised $100m and then hosted an interview with Elon Musk, in which Musk interviewed the CEO of Robinhood. It's a well-hyped company, and is, of course, private. Clubhouse Media Group is public, and is up 457% year-to-date.

Just like most of us still call Alphabet "Google," even when we're referring to the parent company, I'm going to continue calling Clubhouse Media Group "Tongji." The interesting thing about Tongji is not just that it's a mis-hyped company, but that the hype it's trading in response to is bad for Tongji's business, such as it is. Tongji pivoted to being a TikTok proxy (and somehow got a writeup in the New York Times). The scheme was: retail traders want to invest in TikTok, but there's no easy way to make this bet, so they'll supply something TikTok-adjacent. But they wanted to invest in TikTok because it's the most hyped social network—and now Clubhouse is a viable competitor for TikTok's trader mindshare.

(There is, interestingly, a company that is not named after Clubhouse, but does sell it services. Agora Inc., ticker API, which is merely up 85% this year.)

Private Planes Go Public

Last week I wrote about the recovery in private planes, which have gained share from business-class commercial. In a confluence of other recent trends, this is now an investable theme, as a private jet charter company called Wheels Up plans to go public through a SPAC ($, FT). Demand for private jets is lumpy, and a well-capitalized company in the space can take advantage of dislocations. When commercial aviation goes through a downturn, many companies hold on to their fleets; on the margin, it's better to keep operating and servicing debt. But since private planes often get acquired by individuals or non-aviation companies at the top of the cycle, the supply during a downturn is larger. Of course, overcapitalizing a company at the cyclical peak and hoping for a downturn is not an exciting model for investors—they're still buying an asset that will get marked down.

More Decarbonization

Yesterday was a very busy day on the low-emissions front:

Supply Chains as a Hostage Negotiation

A former advisor to Taiwan's national security council touts Taiwan's centrality to global electronics supply chains as a defensive asset. This is a more elaborate version of the "Golden Arches theory of conflict prevention," which holds that countries with close economic ties can't afford a war. While this is true, it has another corollary: those close economic ties mean that the marginal cost of total war compared to limited conflict is lower. If China's economy would be crippled by a serious diplomatic dispute with Taiwan, it removes one more disincentive for escalating beyond that point.

Coasian Parliament

A timely update to yesterday's piece on how nearly-split legislatures inject randomness into the political process: a Coasian defense of parliamentary systems. Small parties that back a narrow set of issues—but are absolutely committed to those issues—create a sort of marketplace in policy tradeoffs, which is a good way to aggregate public opinion.

diff.substack.com · by Byrne Hobart