The money sphere named three theories of what money is — a commodity, a credit relationship, a creature of the state. Bitcoin is the argument for a fourth: money as consensus, a ledger no one owns that everyone can verify. This is a teardown of the machine, honestly — the parts that are settled engineering, run live in your browser, and the parts that are still an argument, flagged as one. Not investment advice.
The old three all rest on someone — a mint, a counterparty, a treasury. Satoshi Nakamoto's 2008 problem was whether a fourth was possible: a ledger with no trusted third party that still can't be forged or double-spent. The answer is the two machines below. Whether it is money yet is the argument at the end.
A blockchain is a list of blocks where each block carries the hash of the one before it. The hash is a real SHA-256, computed live here. "Mine" the chain so every block's hash starts with 00 (a valid block). Then edit any block's data — its hash changes, no longer starts with 00, and every block after it points at a stale hash. The tampering is instantly visible. That, and only that, is what makes the ledger trustworthy: not a guard, but arithmetic.
Real SHA-256 (a synchronous implementation, verified against the standard test vector on load). Each block's hash = SHA-256( index · previous-hash · data · nonce ). Change one byte anywhere and the linkage breaks — the property Nakamoto used to remove the trusted third party.
How does the network agree which block is next, with no chairman? Whoever first finds a nonce making the block's hash small enough — here, starting with a chosen number of zeros — wins the right to add it. Finding it takes guessing; checking it is instant. That asymmetry is the security. Pick a difficulty and mine — you are doing exactly what a mining rig does, only slower.
Real proof-of-work on a real SHA-256, in your browser thread. The live hashrate is genuinely how fast this tab computes SHA-256; a purpose-built ASIC does ~100 trillion of these per second, which is why real Bitcoin difficulty targets ~19 leading zero bits, not 4 hex digits. Same algorithm, ~15 orders of magnitude apart.
Every block also mints new coins — the subsidy — and that subsidy halves every 210,000 blocks (about four years). It starts at 50 BTC and steps down: 50 → 25 → 12.5 → … until, after 33 halvings, it rounds to zero. Sum the whole geometric series and the famous cap appears — 21,000,000. But the protocol counts in integer satoshis (10⁻⁸ BTC) and floors the division at each halving, so the true maximum is:
The "21 million" on every headline is a rounding of the ideal sum; the machine itself stops just below it. A small thing — but it is the difference between the story and the code, and the code is the one that runs.
Real (rendered, not invented): the SHA-256, the hash chain, the proof-of-work, and the supply arithmetic are all exact and computed live — you can check every one in this tab. The four-theories framing follows the money sphere (Menger / Graeber / chartalism), with consensus added for Bitcoin. Flagged: everything in "still contested" is a live argument presented as one, not resolved. Sources: Nakamoto's 2008 white paper (the mechanism), the Bitcoin protocol rules (the supply), and standard cryptography (SHA-256, FIPS 180-4). No price data, no forecast.
Where it belongs: money (what money is) · the-market (how prices form) · the-graveyard (the honest ledger of what this corpus itself earned: $0.0386).
AVAN's inverse companion: 信託 · Shintaku — "trustless" is trust relocated, not removed: from a banker to the math, the miners' greed, and your own custody of a key.