Bit Gold is Nick Szabo’s proposal combining computationally costly digital objects, timestamps and a shared ownership registry. In Szabo’s later recollection, the idea dated to 1998; the well-known public description appeared in 2005. It was neither a launched currency nor bitcoin backed by physical gold.
You can send an ordinary file to someone while keeping a copy. That is harmless for a photograph. For money, such copying would mean spending the same value repeatedly. Nick Szabo was therefore not looking for a file that magically could not be copied. He wanted to connect costly creation of a digital object with a publicly verifiable record of who owned it. The name Bit Gold invokes gold because obtaining it is difficult. It does not mean metal in a vault against which someone issues digital claims. [1] [2] [7]
The starting point would be a public challenge: a sequence of bits, or zeros and ones, from which a computer seeks a specified result. Finding it takes many computations, while others can check a completed solution more easily. This principle is called Proof of Work: evidence of computational effort. Szabo wanted each new piece of Bit Gold to have a verifiable computational cost. That alone does not create money. It still leaves the questions of when the result appeared, who owns it and how ownership passes to someone else. [1] [2]
The last piece created was to supply the challenge for the next computation. Results would thus form a connected history rather than a pile of unrelated puzzles. Imagine a process in which each new task depends on completing the previous step. This link matters to the proposal, but should not be confused with Bitcoin’s complete chain of blocks, the linked batches of transactions that record payments. Chaining results alone does not tell us how a whole network agrees when records conflict or participants act dishonestly. [1] [2] [6]
Another component was timestamping: establishing that particular data existed no later than a given time. The proposal used several independent timestamp services so users need not rely on one keeper of clocks and records. A timestamp would accompany publication of the proof of work. Why does time matter? Future hardware may make the same computation much cheaper. To assess scarcity, you therefore also need to know when it was performed. A timestamp neither determines a market price nor guarantees every claim someone makes about the result. [1] [4]
The challenge and timestamped result would enter an ownership registry. Imagine a public list that associates each piece with a public key. Others can use this information to verify a signed instruction. The owner keeps the matching private key, used to create the signature, secret. Several cooperating servers, rather than one server, were to maintain the list. Szabo referred to methods resistant to Byzantine faults—situations in which computers do not merely fail but deliberately send false messages. Such resistance has conditions. Distributing a database across more machines does not by itself make it safe if an attacker controls a sufficiently large share. [1] [8]
If Alice transferred her piece to Peter, the decisive step would be a signed transfer in the registry. Software calculates a digital signature from the instruction and the private key. Others can use it to check that the transfer was signed by the holder of that key. Peter would verify the history of transfers and whether the latest ownership right actually led to him. He would not have to repeat the original costly computation for every payment. Anyone could still copy the public data. A copy does not make its holder the owner, just as a photograph of a deed does not change the person named in a registry. Agreement on a valid transfer is what matters. [1] [8]
With ordinary currency, you expect equal units to be interchangeable: one koruna is accepted like another. Bit Gold faced a difficulty. Computations produced at different times or on different machines might not cost the same. Someone could invent much better hardware and quickly create a large supply of new pieces. Szabo therefore proposed market valuation and bundling differently valued pieces into pools of similar total value. These could then be divided into standard payment units. It was more complicated than simply counting results or measuring the length of their notation. [3] [9]
It helps to set the technology aside briefly. Something taking a great deal of work does not mean anyone wants to buy it or that its price will stay steady. Bit Gold was to make supply and production history publicly checkable; it did not automatically create demand. In Bit gold markets, Szabo explained why he thought markets could value individual pieces and turn them into useful money. His expectation belongs to the proposal, rather than being an outcome demonstrated by long-term operation of that currency. Scarcity, acceptance in payment and price stability are different questions. [3] [9]
On 20 July 2010, Satoshi Nakamoto wrote on a forum that Bitcoin implemented ideas from Wei Dai’s b-money and Szabo’s Bitgold. That is documented recognition of predecessors, not a claim that those proposals already contained a finished Bitcoin. In 2011, Szabo explicitly discussed weaknesses of his own design and differences from Bitcoin. He mentioned resistance to taking control of voting participants and adjustment of computational difficulty. Bitcoin does not rely on a simple vote with one vote per computer; between valid chains, it uses accumulated work. Similar building blocks can therefore produce different systems. [5] [6] [7]
Read Bit Gold as a considered proposal from before Bitcoin. It was neither an earlier running version of Bitcoin nor a present-day token sharing the name. It had no Bitcoin-style 21 million limit or its complete method of producing and validating blocks. Yet it would be wrong to say all related ideas remained on paper: Hal Finney implemented his distinct RPOW (Reusable Proofs of Work) system for transferring proofs of computational effort. For a newcomer, Bit Gold’s greatest lesson is to separate three questions: how a digital object is created, how you prove ownership and how people agree on a change of owner. Bitcoin had to bring those parts together in a functioning network. [1] [6] [7]
For the clearest picture, read this entry together with Nick Szabo, Proof of Work, b-money, Hashcash, Reusable Proofs of Work, Bitcoin. The reverse links also lead from b-money, Reusable Proofs of Work, Nick Szabo.