The great paradox in Artificial Intelligence is this:

Neural networks

The neural networks can be non-computational however they lack the reasoning function akin to the human brain. They are intended for pattern recognition.

AI for science and engineering 

There is very poor achievement in reasoning AI. There are no AI systems that can fulfill the job of a computer coder, for example.

The background of AI’s current problem

is that the problem is approached in the wrong way.  To get ready for artificial brain development you must your change style of thinking. Change the approach to a non-computational one, and AI systems will be more similar to human brains rather than computers soon.    Here is an old video of a model of a mostly non-computational system to solve logical puzzles. This is not an AI system, but rather a non-computational SAT solver (Boolean Satisfiability Problem). It was based on extended Boolean logic. But it was the first step to a real non-computational AI, with a reasoning ability similar to the human brain. We are working on it now.  This video shows the solving of the task from the book “The Lady or the Tiger” by Raymond M. Smullyan (chapter 2, paragraph 12 “The logical labyrinth”) Should you wish to contact me, my contact is down below. Sergey Bulanov