Alberto Brandolini, a software consultant, put it in 2013 as the bullshit asymmetry principle: the amount of energy needed to refute bullshit is an order of magnitude bigger than that needed to produce it.

It is an observation about cost, and everything useful about it follows from taking the cost seriously.

Where the asymmetry comes from

Producing an unsupported claim requires one sentence and no constraints. It need not be consistent with anything, it need not survive checking, and it can be generated as fast as a person can talk.

Refuting it requires establishing what is actually the case, which means finding a source, reading it, working out whether it says what it is claimed to say, and then writing the answer in a form somebody will read. That is minutes to hours against seconds.

Two further costs compound it. The refutation has to repeat the claim in order to address it, which spreads the claim and — through the way familiarity works on belief — can leave the claim better established than before. And the original is usually simpler than the answer, because the answer has to carry the qualifications the original did without.

The consequences worth knowing

A correction never catches the claim. By the time an answer is ready, the claim has been read by people who will not see the answer, and the mismatch is structural rather than a failure of effort.

Volume beats accuracy in any exchange scored by exchange. If the measure is who answered last, the side that can generate faster wins whatever the facts are — which is Goodhart's law applied to argument.

Engaging can be the losing move. Answering every claim is how a person with something to defend spends all of their time on somebody else's agenda. The exchange looks like a debate and is functionally a denial-of-service.

What to do about it

Pick the claims. Refute what matters and would change a decision. Most claims fail that test and answering them is a donation of time.

Answer to the audience, not the claimant. A person generating unsupported claims is not going to be moved by evidence; the people reading might be. This changes what a good answer looks like — shorter, and aimed at a reader who has not decided.

Lead with what is true rather than with the claim. Since refutation spreads what it repeats, the better form states the correct thing first and mentions the claim once, briefly, rather than quoting it at the top.

Use the burden of proof. Hitchens's razor exists precisely for this: an assertion offered with no reason does not have to be dismantled, only identified as unsupported, which is cheap and restores some of the asymmetry.

The limit

Taken too far it becomes a reason never to engage with anything and a way of dismissing inconvenient claims as not worth the effort. The asymmetry is real and it is not a license to decide in advance which claims are nonsense.

The distinction that keeps it honest: the law is about claims offered without support. A claim with an argument behind it costs real effort to answer because it deserves real effort. That is ordinary disagreement, and the asymmetry has nothing to do with it.