Economic innovations are accounting innovations, and the ones that stick, the ones that become the next layer of economic infrastructure, are the ones that extend a write permission to the edge.
The economic record is not a description of the economy; it is the economy's computation. So every economic-media innovation is a change in who can write what into the shared record, and what the record can express. Read that way, the history has a selection pattern, and it repeats six times.
The tablet and the tally granted the permission to be remembered: strangers could contract, obligation survived memory and death. It paid by centralizing the record with the scribe and the temple. Double-entry granted to self-verify: provable economic state without an authority's blessing, the foundation of credit-at-a-distance and the merchant class. It paid by forcing everything into scalar balances, hardening surplus into profit. The market granted to value: the price is a public, continuously rewritten valuation entry that anyone's trade writes to. It paid by crushing all meaning into one signal. The platform granted to coordinate at network scale without a firm. It paid by enclosing its richer, multi-dimensional ledger: the edge got semiotic write and economic read-only. Blockchain granted to settle and to issue, self-custody and permissionless instruments, and paid by centralizing meaning totally, into one global record.
Each compression was a necessity of its era's computational constraints, not a conspiracy: when computation is scarce, meaning must centralize. And each enclosed register is exactly where the next innovation happened — a ratchet, edge-ward. The next permission in the sequence is the one every previous medium withheld, the permission to mean: write access to what value is. Its constraint (the informational cost of distributed semantics) is the one the LLM-driven agentic stack has just collapsed. Meaning is the last enclosure.
Betting on the layer that opens it is betting with the selection pressure, not on a rupture.
Full six-era treatment: The Evolution of Economic Computation (v2).
Related: · · ·