> Sounds good in theory but how many people are actually qualified to interpret those algorithms? Enough researchers, at enough institutions with sufficiently

> Sounds good in theory but how many people are actually qualified to interpret those algorithms? Enough researchers, at enough institutions with sufficiently diverse motives, that if multiple of them identify a problem and agree about it, Joe can make a reasonably confident assessment that there's probably something there.