>merely that it's the most successful we've ever come up with to date.
That is a very good reason to be cautious about throwing something out wholesale. But it is also a good reason to not fear experimentation, because the official story is that it will show the incumbent to he superior.
> Maybe. Ask me again when all states disappear entirely and fail to re-arise.
Rather than ask a question, I'll make a statement:impersonal emergent forces of the kind we are talking about use a very greedy algorithm, and are very prone to local optima. There is a place for human assistance in finding where they "ought" to be.
That is a very good reason to be cautious about throwing something out wholesale. But it is also a good reason to not fear experimentation, because the official story is that it will show the incumbent to he superior.
> Maybe. Ask me again when all states disappear entirely and fail to re-arise.
Rather than ask a question, I'll make a statement:impersonal emergent forces of the kind we are talking about use a very greedy algorithm, and are very prone to local optima. There is a place for human assistance in finding where they "ought" to be.