Artificial Intelligence- How far is too far?

[MonsterHunterMaster]

AI can't get too far imo.

there is always a way to beat a game.

 

Oh, have you ever tried to beat this guy without freaking out?

 

http://revistagames.files.wordpress.com/2009/07/psycho-mantis.jpg

 

It's psycho mantis from metal gear.

He reacts to all your controller inputs and is unbeatable.

He's also driving you crazy, cuz he is able to read your memory card saying what games you played.

You had to switch the controller into the other controller port so he wasn't able to read your actions.

That's the way you beat him.

There should be more bosses like this, you know, more difficult imo.

[Vagrant0]

The times a game is downloaded is no indication of popularity nor enjoyment of the game. You mention FO3 and Oblivion, but fail to notice the obvious facts that both games were regarded rather harshly for being too easy and having too many things set in favor of the player for anyone who even knows how to play... DA:O also features many encounters which provide little or no challenge unless you either suck, or have mods installed to increase the difficulty. Although god mods are common, they aren't popular compared to mods which increase difficulty in reasonable ways. Setting the bar lower and lower only results in more and more mediocre games. And although that seems to be the direction many game companies are going, even more gamers are getting sick of being cheated out of their money by a game that provides no challenge, no fun. You argue that people will have fun with an easy win, and I argue that people will get bored of winning without any effort.

 

Regardless, this is not the topic of this particular debate thread. As it pertains to this thread, I believe AI will continue to advance despite your suggestions to the contrary simply because people always like seeing virtual actors who are able to do something new. Take for instance the AI differences between Oblivion and FO3. In Oblivion, generally the most interaction you could get out of an NPC was interaction that you had to design and rig manually. In FO3, many of those things which had to be manually setup for each NPC were packaged into a single routine that decided what interactions to have on its own. While this is only a small change, many people prefer the FO3 sandbox AI simply because it adds new variation and is more sophisticated than what they saw in Oblivion... And likewise, people liked the AI in Oblivion because it allowed a wider range of reactions to the players actions than Morrowind... And likewise, people liked the AI in Morrowind because it was remarkably more dynamic than what they had seen in games prior. Just because the AI becomes more sophisticated does not mean it has to be more difficult.

[Surenas]

Marginal note

 

Actually verifiable user insights (different from individual impressions) in the special case of Bethesda are predominantly a function of dwelling time here on the Nexus and the Bethesda forum. So it was e.g. already clear in 2002/3 what's up with the AI in Morrowind. That's nothing new.

 

Sorry to say but the statement “the times a game is downloaded is no indication of popularity nor enjoyment of the game” is everything but serious cos the absolute download times of a mod or modlet are the sales and only that is the marker of popularity. No sales thus prove the absolute lack of it - unpopularity, the bankrupt in visible trade. That’s one of the basic principles in economy and not to be questioned, neither here nor elsewhere!

 

Right MonsterHunterMaster, but it's always a matter of be invested time and necessary effort...

 

So, the way is again open for an animated discussion of AI beyond the sphere of our main interest - viz gaming. Thanks for your patience, Vagrant0 et al. Have fun! :thumbsup:

 

 

http://www.abload.de/img/anne9hol.gifinside

[slygothmog]

I dont normally post in debates, but I remember reading once, if you took all the computers in the world and somehow managed to link them up, their processing power would still not equal that of a single human brain.

So from that, I doubt very much any kind of 'machine' will aquire the ability to 'think', before we wipe ourselves out.

[TheCalliton]

machines are limited, sure htey can compute 1349345 X 298417

but we can figure out how to use that number to solve a problem. A computor can only sove what it is told how to solve

[Vindekarr]

Well, the other thing to remember is that not everyone is smart, and not everyone is stupid.

 

Idealy we could calibrate an AI to our OWN personal prefferences. People like myself who thrive on dificulty and beleive that pushing yourself beyond your limits is the key to personal strength may choose to go with something hyper intelligent, or with preternatural reflex, or superior knowledge. Lazy couch potatos who fear such challenge or do not like to exeed thier limits may limit it to mere human levels of prowess.

 

And what about intelligence? The typical form of analysing this bizaar comodity is an IQ test. But these are based largely on how closely someone's answers hold to a predetermined normal. Higher or lower = higher or lower scores. I myself have a low opinion of these tests. They arent conducted scientificaly, or empircaly, and are usualy, alteast where I live, based on early 20th century ideas. typicaly fruedian. So to accurately choose an AI for someone's mental ability dynamicaly the programmer would either need to do centuries worth of study on the human mind and brain, or dodge an unending minefield of using the decades outdated single-intelligence "answer these questions correctly" test that rellies on knowledge rather than intelligence.

 

And then the problem of dynamic dificulty emerges, this is a problem because I've never seen a game that can do this properly. Valve went on for hours about left for dead's dynamic dificulty, yet when I played that game, neither I nor anyoen I spoke to ever saw it in action. The dificulty level was infact a hoax, no such changes according to skill have ever been seen either by myself or anyone I've spoken to*. And even if it did work, people could, the way systems work right now, manipulate it effortlessly.

 

 

 

*If you've seen this marketing ploy actualy produce a diference, by all means, tell me, I'd love to be wrong about hating that game.