[MUD-Dev] AI not worth doing in our games?

[Freeman]

From: Sasha Hart
> [Jeff Freeman]
 
>> I want to say: In most cases we don't need smarter AI.  We can
>> achieve the same or better results with more diverse AI and/or
>> some slight-of-hand.
 
> It depends on the problem you are trying to solve. A screwdriver
> makes a very poor hammer, even if it is a very good screwdriver.
> Sometimes you can fake it, sometimes you can't.

Agreed.  That's why I said "In most cases" rather than "In all
cases".
 
>> e.g. A "real" AI for pigeons in the park would be big, complex,
>> and result in flocks of pigeons wandering around the park.

> Not at all.

I should clarify what I meant by "real" AI, then.

Say you want loud noises to scare away the pigeons.  With "real" AI,
they are scared away *because* they have learned that loud noises
are associated with death and pain.  Loud noises make them scared,
because they dislike death and pain.  They run when they are scared,
because they've learned that this behavior helps them to avoid the
death and pain associated with loud noises.

It's the Learning that I associate with "real AI".  In most cases, I
think it's a bit over-the-top to make AI capable of learning from
past mistakes.

It's enough to just tell the pigeons to run away from loud noises.
Or even just to run away from things that scare them, whatever those
things may be.  I'm not opposed to abstraction in AI (believe me), I
just think that "Real AI" is over-the-top.

> This may be an argument that we don't need to attack ambitious
> goals like "pigeons which learn to avoid people on the basis of
> the appearance of their pants fabric and their past mistreatment
> of individual pigeons." We really don't NEED to (although if
> someone did a nice job with that, it might be pretty cool, and
> more importantly it might raise the bar or provide techniques that
> are generally usable; so I won't complain if they do). But
> certainly that sounds unreasonably expensive to me, and not all
> that desirable. These are valid considerations in my making the
> decision.

Yeah, that's the sort of thing I'm talking about there.
 
> There really isn't any good argument for why no one should try.

There are plenty of reasons why no one should try, but all depend on
what exactly is being attempted, who the particular "no one" happens
to be and what goal the no one is trying to achieve.

If the goal of better AI is to make for a "more fun" game, then
there are two reasons not to make *smarter* AI:

  a) Most of the time, it doesn't move you one step closer to the
  goal of "a more fun game".

  b) Time is not in infinite supply for any of us: The time you
  spend making "smart AI" is time you don't have to create a wide
  variety of diverse AI behaviors (which, as I've said, I think will
  take you close to the goal of creating a more fun game).

Of course, if your goal is to make smarter AI - and not even AI
which only seems smart, but for personal and unfathomable reasons,
AI which actually is sentient - then nothing else will do.  Your
in-game pigeons need neural networks.

_______________________________________________
MUD-Dev mailing list
MUD-Dev at kanga.nu
https://www.kanga.nu/lists/listinfo/mud-dev