[The_rationalist]

They do share some concepts to some extent, approximately. Notably the idea of activation function and of backpropagation. But yes clearly they're too different in most cases to make valid analogies between them.

It's still an open question why we can't port a biological neural brain on regular CPU (a neumomorphic system) One could say it's because emulating a brain (e.g Blue brain) require too much compute and/or that we lack observability (we can't easily records all the state data that constitute a brain) but that doesn't preclude emulating a simple brain such as the one of c ellegans (302 neurons whole nervous system!!) and I've never met a satisfying explanation as to why this is so hard and where can I track progress on this task.