Our Machines Now Have Knowledge We’ll Never Understand


Fascinating post. The author goes back to the very earliest writing we have on what it means to know something:

Back at the beginning of Western culture’s discovery of knowledge, Plato told us that it’s not enough for a belief to be true...knowledge in the West has consisted of justifiable true beliefs — opinions we hold for a good reason.

Here's my favorite line in the piece:

The machine-learned way of seeing might be more reflective of how the world actually is than purely human knowledge could ever be.

What if most knowledge isn't easily reducible to symbolic logic but requires a network with billions of weights to know?

The truth is that, as we begin to discover more about our own neural processes, we actually don't understand nearly as much about our own mechanisms of reasoning as we had previously believed. Knowledge, even our own knowledge, has always been incomprehensible to us.

Long read, but very worthwhile. Also worth a look: The Myth of a Superhuman AI.


Want to receive more content like this in your inbox?