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. ↓
In a 2020 interview re-released this year, Michael Saylor compressed twenty-five years of investing into one sentence: find a digital dominant network that has dematerialized some fundamental thing. The line reads as a slogan and is the load-bearing claim. It states a structural property of digital networks that physical networks do not have, and the rest of his thesis (a $250M bitcoin treasury bet in 2020 and the construction of a bitcoin-banking layer in 2026) follows from it.
The property: a digital network past roughly 10x dominance has no equilibrium for a runner-up, because the substrate has no edges.
Every physical network has an edge. Walmart's marginal cost of reaching the next rural town rises with reach until it crosses the marginal value of that customer. Tokyu, the Japanese railway-and-city builder, owns one corridor and cannot extend into Hokkaido without building Hokkaido's tracks. The dominant physical network's economics break before it reaches the population frontier, and competitors survive in the gap. The edge is what allows a long tail.
A digital network has no such gap. Marginal distribution cost is approximately zero; marginal production cost approaches zero; marginal value of the next user under network effects is positive. The two curves never cross within the addressable population. There is no geographic, demographic, or economic frontier where the dominant network's economics break and a smaller competitor finds shelter.
This is the lock. Saylor calls it dominance. It is more precisely structural exhaustion of the niche.
The 10x ratio is not numerology. Network value scales superlinearly with size: n² in the Metcalfe form, n log n in others, depending on connection density. At 10x size, the smaller network cannot offer any user something the larger does not offer better, except in idiosyncratic cases that do not aggregate. The smaller user base becomes defection-prone in every direction, and the defections compound until the network unwinds.
Below 10x, runners-up persist. Pepsi at half of Coke is stable. Bing at a few percent of Google survives on Microsoft's distribution rents. The 10x line is where the structural niche disappears, distinct from the commercial niche which can persist on adjacent rents indefinitely.
Above 10x in digital networks specifically, Saylor's empirical claim is that no monster-scale network has been vanquished. Information: Google. Social: Facebook. Retail: Amazon. Mobile devices: Apple. Crypto: bitcoin. The claim is falsifiable by counterexample. None has surfaced.
The graph already names a kill condition for monopolies generally: irrelevance, not competition. Newspapers did not lose to better newspapers; classifieds became free. The frame sharpens here. The only kill condition that survives the digital lock is substrate-level redefinition above the network's own layer.
Apple is locked within mobile. Apple is fully exposed to mobile being demoted to a legacy substrate by ambient computing or neural interfaces. Google is locked within web search. Google is exposed to retrieval being subsumed by a generative substrate that reframes what a query is. The position is unassailable inside the substrate; the substrate is mortal.
The historical record honors this distinction. Yahoo dominated portals at $100B-class scale and was not vanquished by a competing portal; it was vanquished by the substrate becoming "search" (Google) and "social" (Facebook). IBM dominated mainframes and was not vanquished by a competing mainframe vendor; it was vanquished by the substrate becoming personal computing, then cloud. Nokia dominated feature phones, displaced not by a competing feature phone but by the smartphone substrate. Blackberry the same. Every example offered as a vanquished dominant network is, on inspection, a substrate redefinition. The lock holds. The substrate doesn't.
This implies a time bound. Empirically, digital substrates have shown lifecycles roughly in the ten- to fifteen-year range before redefinition pressure accumulates. The lock is real within that window. It is not permanent. A bet on a dominant digital network is implicitly a bet on the substrate's remaining lifecycle, and that variable is rarely priced explicitly. Saylor's framework is correct about within-substrate dynamics and silent on substrate cadence. The cadence is where most of the actual error in dominant-network bets lives.
Tokyu is the structural opposite of bitcoin. Both are dominant in their substrate. The difference is that Tokyu's dominance is reformable from outside.
Switzerland runs world-class transit on subsidy. The substrate has edges; subsidy can fill the gap between fare revenue and full investment value. Singapore captures land appreciation through state leases. The variable is alignment, not ownership. The lever exists because the network has a frontier where intervention is meaningful.
A dominant digital network has no such frontier. Subsidy cannot enable a competitor to capture substrate-edge value the dominant network leaves uncaptured, because there is no substrate edge. The reform options collapse to two: antitrust-style fragmentation (which fights the network effect itself) or substrate redefinition (which is replacement, not reform). The lever that worked on JNR cannot work on Google. This is the structural reason digital-network dominance feels permanent without observers being able to name what is different. Within its substrate, it is.
There is one place the no-edge claim does need qualification, and it is more important than threshold-fitting.
A digital network's nominal substrate is borderless; its effective substrate may not be. Capital controls, KYC requirements, jurisdictional compliance, and content moderation pressure can carve a nominally-borderless substrate into legal partitions. The lock holds within each partition. It does not hold against legal action that fragments the effective substrate into pieces small enough that none has a 10x dominant network.
This is the real attack surface for any large digital network and is usually misframed as competition. China did not produce a competing search engine that beat Google on technology; it produced a regulatory partition inside which Baidu's dominance is locked and Google's is structurally absent. The competition framing obscures what happened. The substrate-fragmentation framing names it.
For bitcoin, the corresponding question is not whether another chain will compete (that pathway is closed) but how robust the global monetary substrate is against legal partition into sub-substrates inside which different networks dominate. That question is open and is where the analysis is least settled.
The framework has one application-level failure mode that is worth naming explicitly. The substrate boundary is not always cleanly drawn. "Digital monetary network" is one substrate to a bitcoin-first observer, "smart-contract platform" to an Ethereum-first observer, "stablecoin payment rail" to a third reading. Each definition produces a different dominance ratio, a different lock claim, and different bets. The frame's structural rigor is real. The application's rigor is bounded by substrate-definition discipline, which is partly ideological in contested cases.
This means the framework licenses high-confidence calls only where the substrate definition is broadly settled. Web search is settled; mobile is settled; retail-marketplace is settled. The crypto substrate is not. Saylor's bitcoin position is consistent with the framework if his substrate definition is correct. It is not separately a proof that his substrate definition is correct.
The framework, if real, should be portable across substrates. Saylor's career is the test. Four catches over thirty years, in four uncorrelated substrates: enterprise data (MicroStrategy, 1989), mobile (The Mobile Wave, 2012, predicting dematerialization several years ahead of consensus), crypto domains (Voice.com sold for $30M in 2019, the largest such sale ever), monetary networks (bitcoin treasury 2020, then Strategy as bitcoin-banking infrastructure by 2026). Each catch followed the same procedure: identify a dematerializing substrate, locate the network winning it, hold past the 10x threshold. Four uncorrelated substrates is too many for luck.
This is also why Saylor reads more like Elon than like a standard finance figure. Both run substrate-compression operations. Elon at the engineering-physics layer where rockets, cars, batteries, and neural interfaces share manufacturing-and-physics ground truth. Saylor at the dematerialization-and-network-effects layer where each new network is a fresh application of one frame. The structures are isomorphic. Only the substrate differs.
The frame, if held, makes some bets and forecloses others.
It licenses bets on dominant networks within stable substrates against challengers operating in the same substrate. It forecloses bets on dominant networks above their substrate frontier, where the bet implicitly assumes the substrate itself is permanent. Apple within mobile is locked; Apple's broader position depends on mobile remaining the central computing substrate. Different bet. Different sizing.
It licenses suspicion of any "we'll out-compete on technology" pitch against a >10x dominant digital network within its own substrate. That pathway is closed. If a challenger is real, it is operating on a different substrate or a different partition.
It licenses pattern-matching on operators with portable substrate frameworks, who are rarer than the institutional landscape suggests because most institutions select against the cross-substrate generalist who can hold the frame across decades.
The interesting move is not to debate whether bitcoin specifically wins. It is to ask which substrates are dematerializing now, which networks within them are crossing 10x, which substrates are nearing redefinition by the next layer up, and which remain uncolonized. The framework is upstream; the asset is downstream. Saylor has been working upstream since the 1990s. The bitcoin position is one application of a frame, not a one-shot conviction.
Source: Anthony Pompliano, The Pomp Podcast #385, "Michael Saylor On Buying Bitcoin With His Balance Sheet," recorded September 2020, re-released 2025–2026. Saylor's framing (the dematerialized-dominant-network recipe, the 10x dominance criterion, the empirical claim that no $100B-class digital network has been vanquished) is verbatim from that conversation. The no-edge mechanism, the substrate-redefinition kill condition, the substrate-lifecycle time bound, the reformability contrast with physical networks, the regulatory-partition qualification, the substrate-definition failure mode, and the substrate-compression framing of his career are this node's.