Proof-of-work was not invented for Bitcoin. Satoshi Nakamoto took up and assembled several older ideas, already explored in different contexts.

In the late 1990s, the problem of e-mail spam became significant. Indeed, if sending an e-mail costs almost nothing, a spammer can send millions. But if each message requires a small computational effort, sending a single legitimate message remains easy for a normal user, whereas sending millions of messages becomes very expensive.

This is the aim of Hashcash, proposed by Adam Back in 1997, which is considered to be the invention of the proof-of-work principle. The Hashcash principle is very similar to mining: produce a hash that respects a condition (having a certain number of zeros at the beginning of the hash). The proof then accompanies the message and can be verified very quickly by the recipient. If an e-mail is received that does not contain this proof, it can be immediately considered as spam, and therefore filtered. Spammers are then forced to expend a considerable amount of energy to send millions of messages, which drastically reduces (or even completely nullifies) the profitability of this type of operation, whether marketing or fraudulent.

Nowadays, Hashcash is not used for e-mail. Spam filtering is now largely based on centralized systems. Hashcash's advantage over current solutions lies in the fact that it doesn't require centralized filtering: anyone can adjust the parameters according to their own criteria. Nor does it require identification, since a hash search can be performed anonymously. Above all, it does not rely on a reputation system, which tends to introduce subjective forms of filtering.

Hashcash wasn't about creating money. It sought to impose a marginal cost on an easily automatable digital action.

In the late 1990s and early 2000s, Nick Szabo explored the idea of digital scarcity based on proof of work. His conceptual project, called Bit Gold, envisions the creation of units of value by solving a costly proof of work, then recording these proofs in a register to establish a form of ownership.

Bit Gold didn't result in a deployed system like Bitcoin, but it does contain several important insights: the idea that computation can produce scarcity, and the idea of timestamping elements over time to create a history that is difficult to rewrite.

In 2004, Hal Finney proposed RPOW (Reusable Proofs of Work). The idea was to produce proofs of work that could then be exchanged, rather than simply consumed. RPOW aimed to create digital tokens based on proof of work, with a system for verifying and transferring these tokens without duplicating them. RPOW, again, did not satisfactorily solve the problem of a totally decentralized registry as Bitcoin would later do, but it remains one of the great precursors of Bitcoin.

Hashcash, Bit Gold and RPOW use proof-of-work to impose a cost and create a form of scarcity. Bitcoin takes up this mechanism, but gives it a central and collective role: proof-of-work is not only used to create something, it is also used to decide who has the right to write the next page of the register (the next block), and to make this register costly to falsify.