Chicken sexing is the act of telling whether a young chick is a male or female

Difficult; “chickens don’t have penises”

Trained sexers “can sort a thousand birds an hour with over 98% accuracy”

Trained sexers “can’t explain how they know [the sex of a chick]. They just ... do."

New sexers are trained purely via feedback: have them sex a chick, then say if they were right or wrong; rinse and repeat “over a period of a few years”

Chick sexing is thus an example of [[tacit knowledge]]: the sexers are able to know the sex of a chick, but can’t explain how they know

Article defined tacit knowledge as “knowledge that can’t be explicitly communicated”

Machine vision algorithm developed to sex chicks

“Accuracies ranged from 50% to 89%", strictly not worse than luck

Proves that the tacit knowledge of chicken sexing __can__ be codified, if in the fuzzy language of machine vision

Seems to suggest, counter to the previous article, that it’s not that the knowledge __can’t__ be communicated (a machine vision algorithm is a perfectly precise method of communication), but rather that it’s not easy to do so in a human manner; perhaps this is innate to the problem (we will never be able to explain why we can sex a chick) or maybe it’s just because we haven’t yet made the requisite breakthroughs