Which candidate is more fit, a computer programmer, a musician, or a chemist?

If you can't answer that question without further information, then the handicap/costly signal/PoW applied by each candidate is irrelevant to answering the question, too.

You can't even say which of the three is more likely to be a fit candidate based on how much they apply to the handicap signal.

This is because "fitness" is a qualitative comparison of something as it pertains to specific circumstances.

If your handicap has a direct relationship with those specific circumstances, then you can make an inductive chain of reasoning that connects from fitness in those circumstances to the enabling of greater handicaps being paid.

Here, the signal has circumstantial utility.

A deer cares not about the size of a peacock's tail.

This is because these two signals are not comparable, and the circumstances surrounding their utility greatly differ despite their similar use in gathering attention for the purpose of mating in their respective species.

How does one even compare the size of a deer's antlers to the size of a peacock's tail?

Even if one were to compare the calories 'wasted' for the purpose of growing these appendages (which is a directly comparable, quantitative, and objective measurement,) the comparison is still useless as it abstracts away from the circumstances of the signals themselves.

This is a nuanced problem that economists worked out long ago, and it is thanks to their efforts that we know value to be subjective, and subject to the ordinal nature of human valuations. As such, value is determined at the margin, with diminishing marginal utility.

To a peacock, calories are more efficiently spent growing feathers as a costly signal to peahens than they would be spent hashing arbitrary data to meet a difficulty target.

Likewise, deer are similarly improved by growing antlers as opposed to tail feathers.

This is not because they cannot choose to hash or grow tail feathers, rather, it is because they are presenting a specific signal to a specific receiver, and evolution has resulted in the most valuable signal being selected by their receivers. Genetics took care of the rest.

This qualitative difference between signals, combined with the circumstantial nature of those signals, results in an inescapable conclusion:

The best, most efficient signal is not that which is most 'pure' in its nature, but that which best connects cost paid to benefit derived.

Without this circumstantial chain of success begetting success and failure begetting failure, the signal becomes overwhelmed with noise from signalers seeking specific receivers who differ in their desired ends.

It's like everyone communicating over the same phone line at once. If you filter out everyone not speaking at the loudest volumes, nobody can be heard but those who speak loudest.

But that's not a test for fitness of those individuals to your values, rather, it's a test of their fitness in being the loudest voice on the channel.

Remember, signals are about presenting information.

If everyone is trying to use the same signal to declare the value of different information, it always fails because that value is subjective, meaning no comparisons can be made between different instances of the signal.

Those who continue their quest to use "costly signals" for arbitrary things are merely ignoring the hard lessons learned by economists long ago in the marginal revolution.

Those who follow them are being mislead by well intentioned, but wrong, people.

It's a shame that so many smart people are allowing themselves to be so mislead by those espousing long abandoned lines of reasoning.

And it's a shame that investments of time, intelligence, and resources are being diverted towards disproven theologies.