On @DeepMindAI AlphaStar: It's _very_ impressive! No question about that. But I'm not so convinced by that APM argument re "unfair advantage" wrt microing. And the whole focus / framing / accessing unit properties story is also pretty fishy.
-
-
Isn't 500ms a bit harsh? 200ms seems more reasonable for a focused pro player.
-
Are we talking "fair" or are we talking "convincing threshold" for definitely superhuman intelligence. Because I'm talking about the latter and tbh I think 500 ms is super short for making a meaningful strategic decision.
- Show replies
New conversation -
-
-
This Tweet is unavailable.
-
I don't think deepmind sees itself as "game ai" shop but I think I understand what you mean. I just think the current framework can and should not be used to make any statements regarding the "superhumanness" of AlphaStar.
End of conversation
-
-
-
IMO forcing a chess AI to physically move the pieces may make it more difficult for machines to compete with humans at chess but its not a more interesting competition.
-
But AlphaStar gets information that a human player wouldn't get unless they move the section of the map they see. That's totally different from the chess example.
- Show replies
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.