General means its not hard-coded to solve specific problems. Rather there are general purpose algorithms that can solve just about any problem you throw at it.

General means its not hard-coded to solve specific problems. Rather there are general purpose algorithms that can solve just about any problem you throw at it. It could be that some of these algorithms are somewhat specialized, but can still solve a wide range of problems never seen before. Then there's a question of how supervised these algorithms are during training. Unsupervised is ideal, but some degree of supervised training isn't bad either and is more practical.