I had read and filed away Why Minds Are Not Like Computers, which has revived some of my past mulling over what the problems with Artificial Intelligence are.
At root in the current debate as to whether AI is achievable is the soul-body dichotomy, i.e., “Mind is software, brain is hardware”.
I’m convinced the modern approach to AI is on the wrong track. Intelligence is our means of survival as humans. It is used in the service of our lives, which happens to be by means of solving problems related to meeting the requirements of life. As far as the popular scientific press goes, it seems like AI studies all focus on solving “interesting” problems, like pathfinding, or stomping on human opponents on Jeopardy.
I think it would be a fruitful line of research to develop autonomous artificial life, give it a survival condition, and let that new entity figure things out on its own. That would be the truest form of AI there is.