I know it's old hat, but I think it's worth pointing out that VR as depicted in Gibson is impending. There is every reason to believe that we will have 3D worlds rendered indistinguishable from reality in ten years, twenty tops. Video technology is almost there now, and work is being done on producing autonomously acting realistically moving characters in both the movie and the gaming business. Use a video camera and high quality microphone as input (instead of keyboard & mouse), and either goggles or Neal Stephenson style retina painting lasers, plus surround sound. All that's missing at that point is smell & touch, and terrain. I think that you won't notice the lack much, and if people do, and it's obvious what's missing.... someone will build it.

Of course, VR addiction (basically video game addiction on steroids) is likely to be a real, pervasive problem. If you could spend $3000 dollars once and have anything you wanted thereafter, just needing to provide rent for a small apartment, basic food, and utilities, would you go get a better job? Or would you do the minimum necessary to live your fantasy life, even if it is just a fantasy?

There is work available now that uses a combination of fractal, recursive generation and artificial "aging", and some common-sense rules to generate entire fantasy worlds from simple seeds, all automatically.

Dammit, I can't find links to the project that was doing the recursive generation and artificial aging of a world. I think the name was Other Gods. It was an extremely cool idea - first generate the world from some random chart, then generate the contentinents, then the terrain regions, then terrain features, then countries, then cities, then buildings, roads, etc, then characters & items, and evolve them over time to create an ultra realistic and ultra detailed world from nothing. You would be able to go back and 'fix' a parameter to be more to your liking and regenerate history with the new parameter. I think it was probably way before its time, though.

Of course, the one glaring thing that's missing is actors to interact with. Background characters that interact with each other are (relatively) easily generated, but we have a long way to go to build software that can carry on a meaningful conversation with a person. I think the gap there will be filled in with a combination of multiplayer games, an economic system for people playing 'bit parts', and people just putting up with crappy conversations in-game.

I don't think we'll see software that can consistently carry on meaningful conversations with a person until we can simulate a human brain in near real-time.