• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.
  • The forums have been upgraded with support for dark mode. By default it will follow the setting on your system/browser. You may override it by scrolling to the end of the page and clicking the gears icon.

Artificial Intelligence in games

Just solving it, like any game that has finite outcomes.

Anyway, LLM not much use for that. Quantum computer, maybe then we're talking.
 
londiste said:
AI in the recently blown up generative AI sense probably is not a good fit for game. It requires dedicated hardware, is relatively heavy to develop, train and implement and unclear whether this will result in something all that much better than what we have now. Plus the entire problem with controlling the results. Game and engine makers are definitely experimentin with it but that is a problem for another year or decade.
Edit: LLM and generative AIs can be and (if not yet it) will soon be a resource for asset generation though.
Click to expand...
I'd argue that those points are only relevant to building the models. Most new CPU/GPU have powerful and efficient dedicated AI hardware, it's even in your smartphone.

Running the models is a lot easier than training them.

IMO the issue is going to be implementation. For labours of love, this could result in excellent dialogue and replayability. Similar to how DLSS, RT and other AI accelerated tech, how devs tune the implementation for their own game will be the major determining factor of final quality.
 
mb194dc said:
Just solving it, like any game that has finite outcomes.

Anyway, LLM not much use for that. Quantum computer, maybe then we're talking.
Click to expand...
I'm lost. AlphaZero is capable to look at the chess rules and teach itself to a level where it will defeat a human world champion (it also learns Go and Shogi while at it). What exactly is left unsolved there?
 
bug said:
I'm lost. AlphaZero is capable to look at the chess rules and teach itself to a level where it will defeat a human world champion (it also learns Go and Shogi while at it). What exactly is left unsolved there?
Click to expand...

Just read the wiki article for the definition of solved, as I mean it.

Appreciate you could interpret solved as in a computer beating a human.
 
mb194dc said:
Just read the wiki article for the definition of solved, as I mean it.

Appreciate you could interpret solved as in a computer beating a human.
Click to expand...
As I have already said, that doesn't describe "solved", it describes backtracking to find all possible games. That's not intelligence, that's brute force.

You're basically saying AI hasn't solved something that doesn't need intelligence to solve (generate, actually).
 
mb194dc said:

Solving chess - Wikipedia

en.wikipedia.org en.wikipedia.org

There are even more possible variations of chess games than there are atoms in the observable universe...
Click to expand...
Thanks for clarifying.

bug said:
As I have already said, that doesn't describe "solved", it describes backtracking to find all possible games. That's not intelligence, that's brute force.

You're basically saying AI hasn't solved something that doesn't need intelligence to solve (generate, actually).
Click to expand...
Solved means a perfect set of moves, in this case.
 
R-T-B said:
Solved means a perfect set of moves, in this case.
Click to expand...
If you know anything about algorithms (most people don't and I don't know your background), you know if we made computers look for the perfect solution for every problem, we'd never get anything done.
Most hard problems are intrinsically unsolvable if you ask for the perfect solution, we settle for "close enough" ones*. That's part of "intelligence". If you insist on inspecting every possible alternative, you're back to brute-forcing.

* Close-enough, but mathematically proven not to be way-off.
 
bug said:
If you know anything about algorithms (most people don't and I don't know your background), you know if we made computers look for the perfect solution for every problem, we'd never get anything done.
Most hard problems are intrinsically unsolvable if you ask for the perfect solution, we settle for "close enough" ones*. That's part of "intelligence". If you insist on inspecting every possible alternative, you're back to brute-forcing.

* Close-enough, but mathematically proven not to be way-off.
Click to expand...
Well, yes, but that still doesn't mean we've "solved chess" by the definition he's providing. It just means it's rather silly to try to do so.
 
A lot, if not pretty much all of problems in the analogue world have infinite variables and are not solvable.

Human designed games with limited variables in a closed environment, like a chess board are. Just need another few decades or centuries of technological advancement and I think Chess will be solved. Quite a few games are solved already:

Solved game - Wikipedia

en.wikipedia.org en.wikipedia.org
 
R-T-B said:
Well, yes, but that still doesn't mean we've "solved chess" by the definition he's providing. It just means it's rather silly to try to do so.
Click to expand...
While technically correct, I doubt the purpose of AI is to solve problems by whatever standards defined by some guy on a random internet forum.
 
bug said:
While technically correct, I doubt the purpose of AI is to solve problems by whatever standards defined by some guy on a random internet forum.
Click to expand...
It's a mathmatical term, look at the wikipedia he linked.
 
R-T-B said:
It's a mathmatical term, look at the wikipedia he linked.
Click to expand...
The travelling salesman problem (among many others) also has a mathematically perfect solution. No software is trying to look for that.

For chess, in particular, I don't think anyone (computer or otherwise) can imagine that perfect solution. If both parties played perfectly, the game would end in a draw. Winning means one parties makes at least one mistake, that's not something you could anticipate when trying to generate the succession of moves.
 
bug said:
The travelling salesman problem (among many others) also has a mathematically perfect solution. No software is trying to look for that.

For chess, in particular, I don't think anyone (computer or otherwise) can imagine that perfect solution. If both parties played perfectly, the game would end in a draw. Winning means one parties makes at least one mistake, that's not something you could anticipate when trying to generate the succession of moves.
Click to expand...
A lot of things in mathmatics are more about theory than practicality.
 
  • Like
Reactions: bug
R-T-B said:
A lot of things in mathmatics are more about theory than practicality.
Click to expand...
True. But in the particular case of the travelling salesman (and many other algorithms, if we're to generalize a little), math simply describes the optimum (shortest) path. And then we have other math, showing that instead of spending 1 year looking for the optimum path, we can instead spend 15 minutes looking for one that is sufficiently close to optimum. Quite a practical result, I'd say.

This is very much the main point I was trying to make: if you want to be intelligent solving a very complex problem*, you just don't brute force it to obtain THE result.

In algorithm slang, that would be a problem solvable in O(n^2) or worse.
 
You must log in or register to reply here.
Back
Top