I agree. I think the current state of AI can roughly be described as "we want to fly like birds" rather than "we want to fly". Aviation is a great example of ho

I agree. I think the current state of AI can roughly be described as "we want to fly like birds" rather than "we want to fly". Aviation is a great example of how the formulation of a task makes a huge difference. We still can't build machines that would fly exactly like birds, but instead we have built machines that fly faster and carry a lot more weight.