I think you’re right when it comes to SC2, but that doesn’t really matter for DeepMind’s ultimate goal with AlphaStar: to show an AI that can learn anything a human can learn.
In a sense AlphaStar just proves that SC2 is not balanced for superhuman ( https://news.ycombinator.com/item?id=19038607 ) micro. Big stalker army shouldn’t beat big Immortal army. In current SC2 it obviously can with good enough micro. There are probably all sorts of other situations where soft-scissor beats soft-rock with good enough micro.
Does this make AlphaStar’s SC2 performance illegitimate? Not really? Tho in the specific Stalker-Immortal fight, input through an actual robot looking at an actual screen and having to cycle through control groups to check HP and select units PROBABLY would not have been able to achieve that level of micro.
The deeper problem is that this isn’t DeepMind’s goal. It just means that SC2 is a cognitively simpler game than initially thought(note, not easy, simple as in a lot of the strategy employed by humans is unnecessary with sufficient athletic skill). The higher goal of AlphaStar is to prove that an AI can be trained from nothing to learn the rules of the game and then behave in a human-like, long term fashion. Scout the opponent, react to their strategy with your own strategy etc.
Simply bulldozing the opponent with superior micro and not even worrying about their counterplay(since there is no counterplay) is not particularly smart. It’s certainly still SC2, it just reveals the fact that SC2 is a much simpler game(when you have superhuman micro).
Those links were really interesting! My take them: any truly intelligent paperclip maximizer would not make any paperclips until it had under its control as much of the universe as it feasibly could. Or it’d turn into paperclips only those parts of its domain that could no longer help it expand its sphere of influence.
Basically a true paperclip maximizer would almost certainly not start turning the Earth into paperclips, since it would understand that using the Earth as a jumping off point for galactic colonization will produce many more paperclips in the long run.
This seems like a really effective counter to the naive presentation of the idiot paperclip maximizer. Has it been addressed and countered in turn anywhere?
I guess this self-improving and expanding maximizer would still view humans instrumentally, but it might still want to use humans as tools for it’s expansion. And indeed, depending on the trade-offs of neurologically modifying humans for obedience(or whatever), it might even leave the base stock more or less alone to the forces of evolution. It becomes more of Quixotic Crusader for Paperclips, with a suicide pact as part of the ideology at the very end of the crusade(once the universe is ours we all turn into paperclips).
Delayed gratification taken to cosmic extremes.