b-money is Wei Dai’s 1998 proposal for cryptography-based money and contracts in the cypherpunk milieu. In the 21ideas framing (Genesis Files, Part III), it is one of the closest “first drafts” of Bitcoin’s basic pattern: pseudonymous keys, signed messages, and a shared ledger — but with critical gaps in practicality and consensus.

What b-money tried to achieve

Genesis Files, Part III emphasizes Dai’s motivation: enabling online cooperation (money + contracts) among pseudonymous participants, resistant to regulation-by-violence.

This goal overlaps with Bitcoin’s later “don’t trust, verify” posture and the attempt to remove privileged intermediaries.

Two designs (version 1 and version 2)

According to Genesis Files, Part III, Dai described two variants.

Version 1: everyone keeps the ledger

In the first proposal:

  • every participant keeps their own copy of the ledger,
  • transfers are broadcast as signed messages,
  • identities are public keys (not legal names).

The source notes why this sounds familiar: it resembles a key intuition behind Bitcoin’s public verification model.

Version 2: “servers” keep the ledger

Dai’s second version introduces specialized ledger-keeping “servers” connected via a broadcast network. Regular users verify transactions against a subset of servers, and servers post bonds for incentives/penalties.

In 21ideas’ telling, this starts to resemble later ideas where ledger power concentrates, creating new trust and governance problems.

What b-money did not solve (per the sources)

The Genesis Files article explains why b-money was not a complete practical system:

  • the design assumed network/broadcast properties that did not scale well,
  • double-spend prevention and robust consensus were not concretely solved in the first design,
  • the “server” model introduces a different trust surface,
  • money creation required participants to agree on the “cost” of computations (including Proof of Work), a fragile requirement in an adversarial world.

This is why 21ideas treats b-money as conceptually close but not sufficient.

Why it matters for Bitcoin history

Two connections are highlighted:

See double spend and Byzantine Generals Problem.

Sources