Proof of Nothing

People often imagine that the basis for consensus mechanisms is an objective proof that some work was done, or that some objective amount of value is at stake, but this is not the case.

Proof of Work is not proof of work

A traditional proof of work is, in fact, not proof that any particular amount of work was done.

To submit a proof of work in Bitcoin consensus, for example, a miner needs to find a string of data that, when hashed using a specific hashing algorithm, produces a hash that begins with a prespecified number of zeroes. This is a difficult problem because the only known way to learn what the hash of a string of data will be is to hash it. Furthermore, the hashed output appears random in the sense that two strings that differ only by a single bit may produce hashed output that are entirely dissimilar. In other words, for a prespecified hashed output, the only way to find the corresponding input is to “guess and check,” which, for sufficiently long output, is intractable for modern computers.

However, it is possible that the input needed to produce a specific hash could be found, by chance, very quickly. Therefore, finding the input string and submitting it as a “proof” of work, is not a proof that work was done in the mathematical sense, but it is an indication of the number of guesses (work) that would be needed to solve the problem on average.

Proof of Merit is not proof of merit

The mechanism used by the process described in the ƒractally whitepaper may be called “Proof of Merit.” An individual “proof” of merit is the final set of attestations of the members in a ƒractal regarding the value other members have contributed to the community. As the Bitcoin process describes how to quantify the work done by computers in the service of Bitcoin, so too does the ƒractally process describe how to quantify the work done by people in the service of a ƒractal. And just as any specific Bitcoin proof of work cannot precisely quantify the work done, any particular proof of merit cannot precisely quantify the value contributed by members. Both proofs of work and merit have significance proportional to the number of them being considered.

Conclusion

Coming to agree on a process to arrive at an objective view of anything in a subjective world is, in the blockchain industry, sometimes called the “oracle problem.” A consensus mechanism proposes a method by which we can gain an objective view of the amount of work done in service of a shared goal, even though the extent to which we value that work is subjective. Ultimately, therefore, a consensus mechanism is an oracle.

This is why the ƒractally team has been repeatedly explicit about the fact that the output of the weekly consensus meetings cannot be thought of as proof of anything. Proof is not a prerequisite for consensus. And consensus does not yield proof. Repeated consensus is proof of only one thing: that consensus was repeatedly reached. Regardless, it’s still useful to approximate the amount of value brought to the community by individuals on average.

Consensus is subjective and messy, but it is a prerequisite for coordination, and it is through coordination that we achieve things greater than ourselves.

James Mart
Twitter: _JamesMart