The attempts before Bitcoin
Digital cash had been tried for twenty years. Understanding why each attempt fell short is the fastest way to see what was actually new in 2008.
Bitcoin is frequently described as appearing from nowhere. It did not. Every component existed beforehand, and the people who built those components are named in its own reference list.
DigiCash
The cryptographer David Chaum founded DigiCash in 1989, building on his earlier work on blind signatures — a method letting a bank certify a payment without learning what it was for. The privacy properties were genuinely excellent and remain impressive.
Its cryptography protected the user's privacy from the issuer. It did not remove the issuer. The company filed for bankruptcy in 1998, and the system could not have outlived it, because the system was the company.
Hashcash
Adam Back proposed hashcash in 1997 as an anti-spam measure: require an email sender to perform a small, verifiable computation before a message is accepted. Trivial for one message, expensive for a million.
This is precisely the mechanism Bitcoin later used, repurposed. The computation is worthless in itself; what matters is that it is provably costly to produce and cheap for anyone to check. See proof of work.
b-money and bit gold
- b-money, proposed by Wei Dai in 1998, described a system where participants maintain their own copies of a collective ledger and money is created by performing computational work.
- Bit gold, described by Nick Szabo around the same period, proposed chaining the output of one computation into the next so that a verifiable sequence of work accumulates over time.
Each described creating scarce digital units through proof of work, and each assumed some process for the participants to agree on the resulting ledger. Neither specified a mechanism that worked when participants disagreed, when some were dishonest, and when anyone could join or leave at will. That specific gap — agreement among an open, anonymous, untrusted set — is what remained unsolved.
What was actually new
The 2008 proposal combined existing parts and added one rule: when copies of the ledger conflict, accept the one with the most accumulated computational work behind it.
It sounds almost too simple to be the answer. It works because it converts a coordination problem into an economic one. Producing a competing history requires redoing the work, so agreement becomes the cheapest option for anyone acting in their own interest — including an attacker.
Why the history matters
Because it locates the invention accurately. Bitcoin did not invent digital signatures, hashing, peer-to-peer networking, or computational proof of work. It found a way to make an open group converge on one ledger, and then demonstrated that the resulting system could keep running with nobody in charge of it. Everything since is a variation on that.