Skyrim - Storytelling AI - future of Gaming?

Daddy32 said:

Haha, no. There's no limit to what you can do with zeros and ones (well, potential computational strength of the universe is the only limit).

Click to expand...

This isn't true, there are an entire class of problems that are undecidable with any algorithm, given any amount of computational strength. This is a fact of computer science, proven by Alan Turing using the halting problem. Undecidable problems are also far, far more numerous than decidable ones, an uncountable infinity to a countable one.

I am not so sure that consciousness is one of the decidable problems. It is hard to wrap the mind around what it means to not be an algorithm, but human consciousness is the most likely candidate in my opinion.

The A.I progression is disappointingly slow too, we still don't even have a decent A.I within our OS. Haven't happened much, if anything at all, since the Commodore C64 era really.

It's a lot harder than making prettier and more efficient graphics. Plus the benefits are not nearly as obvious...

Burress said:

This is a fact of computer science, proven by Alan Turing using the halting problem. Undecidable problems are also far, far more numerous than decidable ones, an uncountable infinity to a countable one.

Click to expand...

OK, I've got a bit over the top I should have said that I believe in power of 1s and 0s simulation-wise: I say, its possible to model a brain (or create computational model of many (any?) real life events, with enough computational power and knowledge.

Halting problem is different because it does not ask question about the next state of some model - it asks whether the event will happen at any point of time from now till infinity. (unfortunately, I'm out of the university for too long to consider the whole class of other problems this one represents).

As for the mind vs algorithm ... this is still rather philosophical (as in: unanswerable by other sciences). Someone would say that the whole universe is deterministic ie. algorithmical. Others would say there's a magic happening in our heads, which cannot be described or imitated.

But in the case of the consciousness vs algorithm we can take Turing's own stance: screw the consciousness and self-awareness. Machines that can emulate intelligent behavior - appear like intelligent to observer - are good enough.

---

The A.I. progression is slow but steady. You don't need to look further than google's products - "Did you mean", image search, language translation are all examples of advancements in various AI fields we use almost daily.

I agree that the march of AI is steady. I also think it's progress is inevitable, and it will be the biggest innovation that ever occurred in society as well. I think it is easy to speculate robots will eventually be the physical labor force of society. The big hurdle at the moment is something humans can do quite easily, recognize what things are in the environment. I saw an article recently where they have gotten a robot to accurately find all the coffee cups in a large office, so progress is slow and steady here. It may be incremental improvement, but there are only so many things to distinguish in normal environments. Programming them to pick fruit or do factory work would be cake once we could get them to distinguish objects in the environment. I take the stance that I don't care if they look or appear intelligent as long as they can do useful things.

I personally find the idea that I may be an algorithm distasteful. Since we understand what we know and communicate this understanding to others in the forms of algorithms, it is easy to make the leap that that is what everything must be. But we have creative aspects that are mysterious, and there may be some kind of adaptablility to consciousness in that the ability to spontaneously generate algorithms for what we know is something that falls in to the undecidable category that no machine can model. I hope that is the case, and it seems a very open question.

Daddy32 said:

The A.I. progression is slow but steady. You don't need to look further than google's products - "Did you mean", image search, language translation are all examples of advancements in various AI fields we use almost daily.

Click to expand...

I'm pretty sure you could do that 20 years ago too, it just wasnt done because it wasnt needed. We also got more CPU power to compute the A.I, that's the biggest difference and not that we've done anything completely ground breaking with the A.I itself.

Compare Ultima 7 to Skyrim and it's embarassing how little has happened with A.I in games.. It's been 20 (!) years since that game was made, its NPC schedules are miles ahead of any recent Bioware game, for example..

I asked this same question awhile ago, when I first put some thought into the Radiant Story of Skyrim. It really is an interesting system. I definitely think it could be the future of TES games. Not sure if it will be the future of all games, unless some companies decide to swipe the idea from Bethesda since it is such a cool idea. We'll see. But for TES games, I hope the future has more Radiant Story, with even more complexity and larger amounts of content. But that said, there should still be some hand-crafted stuff in the game world too. I think Skyrim has a great balance of the 2 systems. It's done well enough that for the most part, you never know, "is this a Radiant Story quest? Or was this handmade?" Which is a good thing.