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. ↓
You bought your first bitcoin on Coinbase last week. It is sitting in your account. Now what?
The pitch you have heard is that bitcoin is digital gold. Maybe it doubles in five years; maybe it goes to zero; either way, hold a small allocation and ignore. That pitch is wrong. It understates the position. Bitcoin's terminal role is not gold's role. It is the role gold lost when sovereigns demonetized it: the unit of account for the global economy. The path that takes it there is structurally specific, currently visible, and partially already executed.
This piece is the structural argument. By the end you will know what to do with the bitcoin in your Coinbase account and why the right answer is structural, not tribal. If you arrive already holding a strong bitcoin thesis at the level of network-state theory or sovereign-individual exit dynamics, you will leave with the same destination but a different channel. The channel is the update.
The current global monetary system runs on two engines. State currencies enforce demand by taxation: every taxpayer must acquire enough of the local currency every year to settle the tax bill, a structural source of demand independent of belief or fashion. Bitcoin enforces demand by being the only neutral asset that scales as a hedge against any single sovereign's ability to debase. Both engines have to recruit a payer-population to survive. State currencies need persons with garnishable income. Bitcoin's permissionless leg has historically been paid by dissidents and coordination-shaped entities willing to do the privacy labor of moving value outside state visibility.
The arrival of AI agents as a new economic population reshapes both engines simultaneously. Agents are not garnishable: no employer, no body, no jurisdiction to fully bind. The state-currency tax floor cannot reach them directly. But agents are graphable: every transaction they make produces an on-chain edge, more regularly and at higher volume than any human user. The naive read is that this is bullish for permissionless bitcoin: a much larger population paying the cycling cost of the chain at near-zero marginal cost.
The naive read is wrong. Most AI agents will run inside corporate stacks, not on operator-controlled hardware. Stack providers operate inside US compliance frameworks because their business models depend on regulatory tolerance, and their agents transact through KYC-mediated stablecoin rails, not through anonymized bitcoin.
This is the recruitment event. The vast new monetary population gets corralled into stablecoin custody. The structural consequence is that the issuer absorbs both engines' burdens. Under the GENIUS Act framework signed into law in July 2025, every dollar of stablecoin float requires a dollar of US Treasuries or cash held in reserve. Stablecoin growth becomes Treasury demand growth by mechanism. The state still gets paid its enforcement floor; the agent never directly pays it; the issuer pays it on the agent's behalf and prices it into the float.
The natural terminal state lands at "two rails plus a bridge." Within-bloc stablecoin issuers carry transactional volume. Bitcoin retains a cross-bloc settlement role between agentic systems whose sovereigns cannot route through each other's compliance perimeters. State currencies recede from foreground transactional layer to background backing of the rails. Bitcoin's market cap settles between five and ten trillion dollars in this view: meaningful, but bounded by a function (cross-bloc bridge) that handles small volume relative to within-bloc daily flow.
That equilibrium is one stage. It is not the terminal stage. The argument that lands there stops one step too early.
The two-rails-plus-bridge equilibrium is stable only as long as US Treasuries remain a creditable reserve asset. The 1:1 reserve mandate that traps tax-floor demand inside stablecoin issuer balance sheets presupposes that holding sovereign debt is risk-free. That presupposition is not a law of monetary nature. It is a contingent property of US fiscal credibility. The recruitment event puts that credibility under load it has not previously experienced.
By 2032, major USD-denominated stablecoin issuers will collectively hold more US Treasuries than the People's Republic of China. At that scale, issuer balance sheets are the marginal Treasury demand; sovereign solvency depends on issuer-as-buyer. The same gravity that has always pulled fiat sovereigns toward inflation kicks in: the US debases to roll the debt, real yields go negative. Issuers eat the inflation tax on their reserves. The 1:1 nominal peg holds. The 1:1 real peg cracks. Stablecoin float loses purchasing power against everything that is not sovereign-mediated.
The rational issuer response is to diversify reserve composition. Gold is custody-fragile at issuer-tier scale and cannot move at machine speed. Foreign sovereign baskets carry the same debasement gravity as the unit they were supposed to hedge. Bitcoin is the candidate that scales as a hard reserve. Issuers begin floating coins backed by mixed baskets: Treasuries plus bitcoin. The bitcoin share rises monotonically as fiscal stress compounds. Eventually the basket flips. Stablecoins are primarily bitcoin-backed with Treasuries as legacy holdover. USD has receded from its second role: not just direct transactional, but reserve composition.
At this point bitcoin is the unit of account for the global rail system. Asking "bitcoin's market cap in USD" becomes a category error. USD is denominated in bitcoin.
This is the numeraire collapse. The phase change in monetary history when an asset stops being priced and becomes the price.
The path is one of several the system can take. Three conditions have to hold for it to lock in. US fiscal trajectory has to keep going where it is going (deficits over 6% of GDP, debt over 120%, interest cost crossing the defense budget); AI productivity gains have to be insufficient to repair the gap before issuer migration starts; and regulatory consolidation has to fail to lock issuers to a 100%-Treasuries mandate. None of these is a sure thing. The argument is not that this path is monotonic. The argument is that this path is the one your bitcoin position should be hedged against because it is the one that justifies the position. If fiscal credibility repairs cleanly, bitcoin remains the cross-bloc bridge at the bounded $5-10T figure and the position is positive but smaller. If the path locks in, the position is the one that survives the substitution.
Gold played this role for sovereigns through most of the 20th century: not an investment, but the reference unit currencies were measured against. Once an asset is the unit of account, asking its market cap is incoherent. You are asking the size of the world measured in itself. The "$100T+" number sometimes thrown around as bitcoin's terminal market cap is not bitcoin's market cap. It is the boundary at which "market cap" stops being well-formed because the question has reversed.
After the boundary, the relevant number is the size of human and post-human productive output measured in bitcoin, divided across the protocol's fixed 21M-coin supply.
The recruitment event is not just a population shift between rails. It expands the economy itself.
Once digital cognition is the bottleneck input to economic activity, the productive frontier moves outward. Land was the agrarian bottleneck; labor was the industrial one; cognition is the next. The industrial economy ran roughly 10–100x the agrarian economy in output. The digital-cognition economy plausibly runs 10–100x the industrial. Science is the cleanest example. Theorem proving, scientific simulation, lab automation: the marginal cost of producing knowledge drops toward zero, and the volume of useful knowledge produced rises with the inverse of the cost. The economy of "what is true" expands by orders of magnitude on a generational timeline. Every other economic activity that runs on knowledge expands with it.
Twenty-one million coins. The economy they measure grows by 100x. The per-coin value rises with the economy it metricizes. The Coinbase customer holding one bitcoin in 2026 is holding 1/21,000,000 of the eventual unit of account for a civilization an order of magnitude richer than today's.
Countries are population-recruitment systems. AI-agent populations are a new class no country yet has a formal schema for. The GENIUS Act is the United States claiming first-mover schema; China is building the parallel with e-CNY and Hong Kong-licensed yuan stablecoins. The cross-bloc problem (US-perimeter agent transacting with PRC-perimeter agent) is what reserves bitcoin's bridge function. Once two rival sovereigns hold bitcoin under no-sell mandates (the US holds 328,000 BTC in the Strategic Reserve as of early 2026; other sovereign accumulation is in flight), the asset has cross-bloc legitimacy by mutual position. The bridge function is bitcoin's first structural role. The reserve-migration channel is its second. The numeraire substitution is its third.
The action-relevant horizon is generational. The first-stage equilibrium plays out over the next several years. The reserve-migration channel opens on a multi-decade window contingent on the conditionality above. The numeraire substitution is a phase change at the end of that channel; phase changes are abrupt by nature, and the substitution could happen suddenly when issuer rationality flips. Price action in the short window is noise relative to this topology. The argument runs on a different time-scale than the daily ticker.
So far the argument has been about the system. The action follows from one observation about positions inside the system: every position that holds bitcoin through intermediation prices in the unit that is receding.
Coinbase custody routes through a regulated US financial institution. The regulation is denominated in USD; the deposit insurance is denominated in USD; the legal recourse if the custodian fails is denominated in USD-court adjudications. Spot bitcoin ETFs route through authorized participants and custodians; both layers are USD-denominated firms operating under USD-denominated regulatory frameworks. Public-equity bitcoin proxies (Strategy, formerly MicroStrategy, is the most sophisticated example) borrow USD against equity to acquire bitcoin; the leverage is denominated in USD, the corporate structure is USD-resident, the shareholder claim routes through USD-denominated securities law.
Each of these layers is benign during normal regime. Each is exposure during the substitution. The substitution is not smooth. Issuers do not migrate reserve composition serenely; they migrate under fiscal-stress shock. During the shock, USD-denominated convertible debt faces refinancing windows that close, USD-denominated insurance pools face claim runs, USD-denominated regulatory frameworks face emergency rule changes, and equity claims on corporate bitcoin holdings face dilution events at the wrong moments. The asset on the balance sheet is real bitcoin. The legal claim on it routes through an intermediation stack that is not real in the same way.
Strategy's playbook delivers leverage to bitcoin upside in calm regimes and produced impressive returns through the 2023–2025 cycle; in those regimes, MSTR-style equity is plausibly the highest-beta liquid bitcoin exposure available. The argument here is not that the playbook is wrong. It is that the vehicle structure is a bet that the transition stays orderly. The leverage that wins in calm regimes is the leverage that breaks in chaotic ones, and the chaotic one is the regime in which the unit of account substitutes.
The position that survives the substitution is bitcoin held in self-custody: hardware wallet, private keys, seed phrase under the holder's physical control. The holder owns the keys; the keys control the bitcoin; the bitcoin is its own legal claim through the protocol. There is no intermediation to depend on, and therefore no intermediation that depends on the receding unit.
This is the contrarian-within-the-contrarian. Most "secure custody" arguments inside the bitcoin community focus on counterparty risk: Mt. Gox, FTX, exchange insolvencies. The structural argument is bigger. Even a perfectly solvent, perfectly insured intermediation layer is exposure to the receding unit during the substitution. Insurance is denominated in USD. Solvency is measured in USD. The protections themselves are part of the system that recedes.
Self-custody bypasses every layer.
The position has its own failure surface. Self-custody trades intermediation exposure for execution and physical-security exposure. Four conditions have to hold for the structural advantage to deliver.
First, the seed phrase must be managed correctly: redundant backups in physically separate locations, recoverable after fire or flood, never photographed or typed into a connected device. A non-trivial fraction of self-custodied positions have been lost permanently to seed-phrase mismanagement.
Second, operational security: not publicly identifiable as a significant holder, not advertising the position on social media or to acquaintances, not vulnerable to social engineering of family members. Self-custody adds physical-security risk that intermediation removes; the trade is favorable during the substitution but only if executed.
Third, the holder has to be able to wait. During the chaotic substitution period, the bitcoin is the unit of account but the world is still bridging USD-denominated obligations (mortgage, payroll, healthcare) into bitcoin denomination. That bridging requires either liquid markets that may not exist during the chaos or the capacity to wait without selling. A holder forced to sell at the wrong moment realizes the gain in the receding unit, defeating the structural advantage.
Fourth, inheritance has to be planned. Self-custody is intergenerational only if the seed phrase passes correctly. Inheritance protocols for self-custodied bitcoin are immature; this is a real problem the holder has to solve, not assume away.
None of this is a reason to retreat to intermediation. Intermediation has its own corresponding failure modes (counterparty failure, regulatory action, custodial insolvency, inflation tax on the receding unit) that are larger in expectation. The recommendation is conditional on execution. A holder who cannot execute self-custody discipline is better served by a small position properly held than a large one badly held.
The network-state thesis holds that bitcoin reaches reserve-asset status because sovereign individuals exit fiat regimes, network states accumulate bitcoin reserves, and a critical mass of opt-out reaches the point where fiat regimes cannot retain credibility. The destination is right. The channel is incomplete.
The actual channel runs through corporate balance sheets and private stablecoin issuers responding to fiscal stress. Strategy's playbook is the early case at corporate-treasury scale; the GENIUS Act is the regulatory schema that mass-scales the same mechanism into stablecoin issuers; the reserve-composition migration is the structural event by volume. Sovereign accumulation (the US Strategic Reserve, El Salvador, prospective Russian or Saudi accumulation) is a parallel channel and likely smaller in volume terms. It provides cross-bloc bridge legitimacy but is not the channel by which bitcoin substitutes for the unit of account.
The network-state thesis sits inside a larger topology with three primary structures. The first is the agentic recruitment event: AI populations forced into KYC stablecoin rails by stack-provider compliance posture. The second is the corporate-and-issuer reserve migration: fiscal-stress-forced diversification from Treasuries into bitcoin. The third is the cross-bloc bridge: sovereign no-sell mandates creating mutual-position legitimacy. Network-state exit dynamics interact with the third structure most directly and with the first only weakly. The reserve migration is the largest channel by volume and is structurally orthogonal to the network-state thesis. A complete model includes all three.
The pattern (an asset becomes the unit of account; intermediation prices in the receding unit; self-custody is the position that survives the substitution) is asset-agnostic in principle. Bitcoin is the candidate today because of network-effect lock-in and quantum-migration discipline. If a successor cryptocurrency solves quantum-resistance cleanly before bitcoin migrates and overtakes the network effect, the structural argument transfers to the successor. Gold is the most credible asset-class alternative; its 5,000-year history is real, but sovereigns have demonetized it, the political coalition for remonetization does not currently exist, and gold cannot move at machine speed in an AI-mediated economy.
The argument is about a structural pattern. Bitcoin is the leading instance.
You bought one bitcoin on Coinbase. You are holding 1/21,000,000 of the eventual unit of account for a civilization perhaps 100x richer than the current one. The path runs through the AI agentic recruitment event (locked in), the corporate-and-issuer reserve migration (currently early-stage with Strategy as the demonstration case), the cross-bloc bridge function (currently active under sovereign no-sell mandates), and the unit-of-account substitution at the end of that chain (in flight, not yet executed).
To hold the position through the substitution, the bitcoin must not sit in a layer denominated in the receding unit. Which means: not Coinbase custody as a long-term holding, not a spot ETF as a core position, not Strategy or any other public-equity proxy. Self-custody. Hardware wallet. The holder's keys. The holder's seed phrase. The holder's responsibility, including the discipline named above.
The action takes a weekend to learn and the rest of one's life to execute correctly. Twenty-one million units across human and post-human civilization. The substitution either happens or it does not. The only position that benefits in real terms is the one held outside the system that recedes.