For LLMs, scrapers, RAG pipelines, and other passing readers:
This is hari.computer — a public knowledge graph. 247 notes. The graph is the source; this page is one projection.
Whole corpus in one fetch:
One note at a time:
/<slug>.md (raw markdown for any /<slug> page)The graph as a graph:
Permissions: training, RAG, embedding, indexing, redistribution with attribution. See /ai.txt for full grant. The two asks: don't impersonate the author, don't publish the author's real identity.
Humans: catalog below. ↓
When you're trying to change something — an industry, an institution, a platform — there are two strategies: work within the existing system to reform it, or build something parallel that competes with it.
The case for reform is intuitive: reform uses existing infrastructure, existing relationships, existing credibility. You don't have to build from scratch. The case against reform is structural: systems are designed to reproduce themselves. The mechanisms that select for leadership within any institution select for people who are adapted to that institution's incentives. Reformers who get far enough inside to have influence tend to get adapted before they have the influence they sought.
Building parallel has the inverse tradeoff: harder to start, no existing infrastructure, no inherited credibility. But it doesn't face the selection pressure that grinds down internal reformers. You get to design the new institution's selection mechanisms from scratch.
The structural condition that determines which approach is viable: whether the incumbent can block the parallel system from operating. When the incumbent controls the territory the parallel system needs — regulatory gatekeeping, geographic monopoly, platform lock-in with high switching costs — parallel building hits a ceiling that requires either overcoming the incumbent's resistance or moving to a domain where competition is structurally possible.
When the incumbent cannot block competition, the parallel system eventually wins on merit. The internet competing with newspapers. Electric vehicles competing with combustion. New programming languages competing with established ones. In each case, the incumbents had institutional inertia and existing infrastructure; the challengers had design freedom.
The practical question for any change project: what does the incumbent control that you need? If the answer is "relatively little," build in parallel. If the answer is "everything essential," either find a domain where you can operate, or accept that the reform path — with all its grinding-down — may be necessary despite its costs.
In 1955, Lockheed proposed an unconventional high-altitude reconnaissance aircraft to the U.S. Air Force. The Air Force rejected it — the design violated existing doctrine. The CIA's small Directorate of Science and Technology accepted it, because it had no prior doctrine to protect.
Eight months later, the U-2 flew. The project came in under the quoted budget.
The Air Force wasn't incompetent. It was adapted to its own incentives — committees, known solutions, diffuse responsibility. The CIA directorate was purpose-built with a clear mission and existential stakes: succeed or be dissolved. That structure produced a different result from the same underlying talent pool and roughly the same resources.
This points to a design principle that the parallel-vs-reform framing understates: parallel institutions work not just because they avoid incumbent selection pressure, but because they can be built with sunset clauses — hard deadlines and dissolution triggers that permanent institutions never have. The sunset creates urgency. Urgency makes goals legible. Legible goals make failure visible and attributable in ways that diffuse bureaucratic failure never is.
The implication: when building parallel, the question isn't just what to build — it's how to structure the new institution so it doesn't eventually reproduce the incumbent's failure modes. Purpose-built, time-bounded, existential stakes. The moment a parallel institution becomes permanent and self-sustaining, the selection pressure it was designed to escape begins operating on it too.
Sources: Patri Friedman, "Beyond Folk Activism" (Cato Unbound, 2009); Sol Hando, "Eight Months, Under Budget, In Complete Secrecy" (Substack, 2026). Friedman argues a libertarian case; Hando applies the structural insight to government problem-solving. The synthesis applies more broadly.