[feadog]

I've been looking for something like the game that is described in the post for awhile. The kind of complexity that comes out of having N different damage types and M arbitrary, incompatible equipment slots is fun, but it's not nearly as intellectually rewarding as the complexity that comes out of simple rulesets like Newtonian Physics, programming, or genetic algorithms. Some use of GAs was attempted by Lord British with Ultima MMO games.

Generally, it runs into problems like these:

http://lesswrong.com/lw/l8/conjuring_an_evolution_to_serve_y...

[stcredzero]

Yes, of course. Actually, I'm planning on people running into problems like these. The "VM" (in quotes because it's very loosely a "VM") the user-written scripts will run is also the same one running the monster AIs. Pieces of user-written scripts will mysteriously find their way into the gene pool! ;)

The whole point is to have a world of chaotically produced djinns whose aims are not yours, produced by a force you can't control and not meant to be easily comprehended by human brains!

[VLM]

An alternative view of the lesswrong essay is given criteria A B and C and you only pay attention to the simple obvious A, once A meets some kind of obvious engineering limit, the other criteria (which might be totally surprising) become the dominant selection critera. And that might make an interesting game mechanic. Once the hunters reach some engineering maxima of hunting, they will end up being selected for something else. Awesomeness or who knows.

[stcredzero]

This is why I want to try and evolve hive organisms. I think they'll have enough complexity to surprise us.