[maaku]

You might be interested in learning about AIXI. It is in fact trivially easy to create a general intelligence effective in arbitrary domains ... if you assume infinite computational capability. It goes to show that the problem of artificial intelligence is really approximations and heuristics.

[kylebrown]

Its trivially easy to write some formulas down on paper ;) I was looking up AIXI the other day, but couldn't figure out what the acronym stands for.

[maaku]

Actual implementations of AIXI are also quite easy. I've written one before. Not that it's very interesting to run -- it is basically indistinguishable from an infinite loop.

As for what the acronym stands for, heck if I know.