So experts cannot identify mushrooms at all by looking at it?
They might turn it around and look at it from different angles, but then just make an AI that takes in multiple images from different angles, maybe have it ask for different angles if it cannot see everything it needs to see.
And if the experts use other senses besides vision, like smell and touch, just make an AI that says “it might be X or Y, only way to tell them apart is through the smell, so i can’t be sure”
A 2D visual representation is not the same as the real thing
So experts cannot identify mushrooms at all by looking at it?
They might turn it around and look at it from different angles, but then just make an AI that takes in multiple images from different angles, maybe have it ask for different angles if it cannot see everything it needs to see.
And if the experts use other senses besides vision, like smell and touch, just make an AI that says “it might be X or Y, only way to tell them apart is through the smell, so i can’t be sure”
I love when idea people tell me to just do this or that as if it’s easy.
Of course you wouldn’t be able to do it, but there is nothing preventing such a system from being created
Citation needed.