One very small sample -- I had a student try training tensorflow to solve NP-hard problems, we couldn't get it to do any better than random guessing. I wasn't r

One very small sample -- I had a student try training tensorflow to solve NP-hard problems, we couldn't get it to do any better than random guessing. I wasn't really surprised, because you need to make many interlocked decisions, all in one go, and all correctly. Also, it's (at least for me as a human) almost useless to try to learn from other similar looking problems to spot patterns in answers.