There are 41 built pieces in the Analysis catalog right now. Four blog posts, including this one. That gap is not an accident of what got written first -- it is just what happens when the deep work outpaces the signage pointing at it. This post is the signage.
Places are not uniform, even when one number says they are. "America doesn't feel one way — Reston is not South Boston" puts a $52,060 median household next to a $231,101 one, both American, both real, both erased by a single national income figure. "Same government, eleven times apart" does the same thing with Loudoun County, Virginia and Las Marías, Puerto Rico -- same Census Bureau, same currency, an 11x income gap neither number is wrong about. "Why Appalachian whites are not New England whites" goes after the same flattening from a different angle: "white" gets read as one uniform baseline, and the roughly $14,000 gap between Massachusetts and Appalachia says it never was.
Change needs a structure AND a person who actually moves inside it. "Cycles and levers — why Egypt, Babylon, Rome, and Ming China all ran the same course" traces a real, repeating imperial rise-and-fall pattern across three thousand years, then insists the cycle never explains why a specific empire fell on a specific day -- that part is always somebody's decision. "The arc, not the cycle" and "Who does which work — four ways the arc bends, and one that takes a generation" both dig into the same idea from the ground level: almost everything that lasts was built by someone choosing a longer horizon than their own.
The market surprises you the moment you actually check it. "Why venture capital wasn't always an industry" and "Flying cars vs. 140 characters" both dismantle tidy stories about how VC got to be VC. The newest piece, "The equity stake used to be rare — three precedents explain why it's back," checks the current wave of the federal government taking direct equity in companies like Intel and MP Materials against 2008's TARP and 1979's Chrysler bailout -- and finds the government already ran this exact play twice before, and made money on it both times.
That's nine pieces out of forty-one, picked because they were the fastest way to show the range: place, history, and markets, all held to the same standard -- a real, sourced, specific causal chain, not a computed rollup. The rest are on the full Analysis index.