In order to learn the most important skill in any type of AI: Reactive machines, Limited memory, Theory of mind, Self-awareness; I believe, Mathematical skills such as Calculus, Algebra, Probability theory, Graph Theory etc... will be an important part to work and master on any latest concept/technology. Good understanding of deep learning and machine learning algos is an advantage.
Try to learn relevant Programming languages such as Python, Java…
Working skill ? OK. Thinking. Skill of the brain when it works.
Look what do our colleagues enumerate: Reactive machines, Limited memory, Theory of mind, Self-awareness; I believe, Mathematical skills such as Calculus, Algebra, Probability theory, Graph Theory etc...
Do this without a lot of thinking. In several domains.
Skill needed FOR (TO DO) General Artificial Intelligence: relevant knowledge in psychology and then the ability to use that knowledge to problem-solve, particularly the following:
Well, there were many mathematical works on pattern recognition in the last 50 years. So people worked out algorithm, machine did not do it itself. It is complex to formalize how humans do it, although machine can have faster speed and larger memory.
I am also mathematician by education, and I observe a bad trend among younger generation. They believe that machine can solve any problem; even without solutions. No. Computers are fast but more stupid than mathematicians. They even cannot understand the concept of continuous, because all algorithms are essentially discrete. So they are unable to understand human brain, because they cannot even understand all what mathematicians can do.
P.S. Going back to question. For computer to do fast the work requiring long memory and fast speed. For programmers to invent good algorithms. For other scientists - to specialize in the work, where computer cannot replace them (interdisciplinary areas that are not yet formalized by algorithms, philosophy, mathematics of continuous, etc)