Brute forcing, guided by just enough expertise to generate a list of the most likely candidate answers (even a fairly long list—calculating thousands or millions of hashes is usually quite tractable), could be an issue unless the true answer really is extremely obscure amid a vast space of potential answers.
My instinct is that suitable questions (vast space of possible answers, but just a single unambiguous and precise correct answer) are going to be rare. But idk maybe you have a problem domain in mind where that kind of thing is common.
Thanks for the reply! I think the flaw you suggested is closely related to the “likelihood prioritization” augmentation of dictionary attacks from 3.2.1. Definitely something to keep in mind, though one measure against dictionary attacks in general is slow hashing, the practice of configuring a hashing function to be unusually expensive to compute.
For instance, with the current configuration, it would take you a couple years to process a million hashes if all you had was one core of a modern CPU. But this can be arbitrarily modified. The current slow hashing algorithm used also requires a memory burden, so as to make it harder to parallelize on accelerators.
Brute forcing, guided by just enough expertise to generate a list of the most likely candidate answers (even a fairly long list—calculating thousands or millions of hashes is usually quite tractable), could be an issue unless the true answer really is extremely obscure amid a vast space of potential answers.
My instinct is that suitable questions (vast space of possible answers, but just a single unambiguous and precise correct answer) are going to be rare. But idk maybe you have a problem domain in mind where that kind of thing is common.
Thanks for the reply! I think the flaw you suggested is closely related to the “likelihood prioritization” augmentation of dictionary attacks from 3.2.1. Definitely something to keep in mind, though one measure against dictionary attacks in general is slow hashing, the practice of configuring a hashing function to be unusually expensive to compute.
For instance, with the current configuration, it would take you a couple years to process a million hashes if all you had was one core of a modern CPU. But this can be arbitrarily modified. The current slow hashing algorithm used also requires a memory burden, so as to make it harder to parallelize on accelerators.
Curious if you have other ideas on this!