default header

Theory

Artificial Intelligence

Moderator: JC Denton

Artificial Intelligence

Unread postby icycalm » 24 Aug 2018 20:05

https://forums.uberent.com/threads/plan ... st-1148133

tatsujb wrote:starcraft meta ...and even gameplay as a whole is an utter joke.

there's no fun in the actual play the only fun is derived from crushing your opponent to which end you're forced to play the game like an utter machine.

and whoever plays the game most like a machine wins.

I expect DeepMind to beat StarCraft players anyday soon and while that doesn't mean much as inevitably Ai wil beat man at all games I make a distinction between those games Ai will have to hit the singularity to be able to beat humans at.

And that's what I'm driving at : gaps in time

chess was beat in 1996, checkers was beat (as I recall) in 2007, while go was in 2016 and chess and poker in 2017
(we notice a pattern of acceleration going on).

Starcraft should happen in 2019.

but I wager for Supreme Commander Forged Alliance, neural networks and machine learning won't cut it as there's too great a complexity curve and no real counter-less strategy.

that being said I can see it happening right before the singularity, I just see the singularity as something that happens in 2050 if humanity survives 'till then.

And all that leaves me with is that clicking to commit to the imposed meta-ed tactic rather than dictating your all-new well pondered strategy is inferior.

and much less fun.

and less something we should shy away from rather than something to strive towards.


Evidence that game complexity can be measured, and is therefore an objective quality.
Image
User avatar
icycalm
Hyperborean
 
Joined: 28 Mar 2006 00:08
Location: Tenerife, Canary Islands

Return to Theory

cron