Longreads

• Roger's Bacon at New Science asks: whatever happened to good writing style in research papers? At one level, demanding consistent style is a good way for results to speak for themselves, but an important part of the job of researchers is to communicate—as you can see any time you look at the results of a game of telephone from the results to the abstract to the press release to the pop sci blog post. And that might mean writing with a little more pizzazz. One partial solution that's already in place is that many researchers are active on Twitter, where there's a premium on punchy descriptions, so novel results are less likely to get lost (but are still subject to the same game-of-telephone dynamics).

• Scott Alexander suggests a cyclic theory of subcultures, where early in a subculture's existence it's possible to achieve influence merely by showing up, and later on this gets more competitive. As the subculture matures, status gets more zero-sum. This naturally analogizes to successful startups; in financial terms, it's often a whole lot better to be the worst non-fired employee at the Series A than to be a superstar post-IPO (though on a risk-adjusted basis the latter is probably a better bet, at least if you're indifferent to exactly what you work on).

• Sebastien Lleo and William T. Ziemba on How to Lose Money in Derivatives. This paper goes through a few great case studies on classic blowups, like LTCM, Amaranth, Orange County, and more. One common theme: on a longer timeline, the traders were right. But time is funny in finance; one way to look at the idea that "correlations go to 1 in a crisis" is that it's an example of time speeding up, so every bad thing happens at once or, or at least prices act that way.
  
  Via @worstcontrarian.

• A long and wonderful interview with venture capitalist David F. Marquardt. "Cajoling, handholding, hitting with a bat. It’s a little of all that, yeah." This one really gets interesting when it gets to the story of Microsoft; Marquardt spent a year pitching Microsoft on taking his money; they didn't need it financially but did need it organizationally. Other key points: he recognized the demand for better disk drives because he was a member of the Homebrew Computer Club around the same time Jobs and Gates would present there, and knew what hobbyists liked. And his response to being pitched Webvan was to essentially suggest that they build Instacart first. (Instacart, incidentally, was founded the same year the interview took place.)
  
  Via Paul Graham. And, disclosure, long MSFT.

• Ezra Klein on how the media we choose to consume affect our thought process. This is always something worth worrying about, specifically because it's hard to notice. The feeling of boredom is subjectively the same whether it's after three hours of reading a book or three minutes into reading an article that doesn't get to the point quite fast enough.

Books

• The Innovation Stack: a great resource for yesterday's post on Block/Square, this book is also a fun and well-written look at a few different companies that built unique models with interlocking parts. The core mental model of the book is that for a company like Square, it's hard to copy just one thing they offer—everything fits together, so you have to copy wholesale, at which point you've made a partial duplication of your best guess as to what another company is getting right, and that company has a head start.

• Alexander of Macedon: proof that you can accomplish a lot through decisive incremental choices instead of grand strategy. A good way to look at Alexander the Great's career is that, while he was very high-agency, he was also very constrained in his actions. Two of the big forces that propelled his conquest were 1) the worry that Greek city-states would rebel against Macedonia with Persian help, so Persia always needed to be busy defending its own empire, and 2) the Macedonian army was always short on money, and further conquest was the only way to get cash flow.

Open Thread

• Drop in any links or comments of interest to Diff readers.

• Yesterday's Block post talks about how the company has a network of merchants and a network of Cash App users, and these don't quite overlap. Are there any other good examples of a company that built out network effects but had non-overlapping members? This would be especially interesting if the company found a way to close that gap.

Leave a comment

A Word From Our Sponsors

Here's a dirty secret: part of equity research consists of being one of the world's best-paid data-entry professionals. It's a pain—and a rite of passage—to build a financial model by painstakingly transcribing information from 10-Qs, 10-Ks, presentations, and transcripts. Or, at least, it was: Daloopa uses machine learning and human validation to automatically parse financial statements and other disclosures, creating a continuously-updated, detailed, and accurate model.

If you've ever fired up Excel at 8pm and realized you'll be doing ctrl-c alt-tab alt-e-es-v until well past midnight, you owe it to yourself to check this out.