1. 1
    1. 2

      Seems like a lot of words to convey the simple concept that modus ponens only works in one direction. This is certainly a mistake that a lot of people make, especially in political discussions, and I wouldn’t be surprised if some ML researchers make it, but I suspect most people suggesting a connection between their models and the physical reality of how brains operate are not actually making this error, but rather suggesting that their model could provide insight into how brains work.

      As an analogy, correlation doesn’t imply causation, but it suggests that there could be causal connection of some sort.