Virtual worlds, like PCs, are disruptive technology, with unforeseen consequences, Kapor said. They will become mainstream quickly, but -- like PCs in the very early years -- they're now a very marginal phenomenon, Kapor said.
Like PCs, virtual worlds will enable people to do new things, and will create new economies of winners and losers.
But virtual worlds are still in the early adopter stage. The next, larger stage of users -- pragmatists looking for a payoff in uselulness -- has yet to begin.
"Virtual worlds are now at a tipping point," he said. "There is a critical mass of early adoption."
Virtual worlds are succeeding now, where they've previously failed, because of faster PC hardware, global broadband, and an Internet culture which now accepts an "ethic of participation" in areas such as open source, free culture, GNU/Linux and Wikipedia, Kapor said.
But virtual worlds have a long way to go until they become mainstream, Kapor said. They need the equivalent of the Web application server -- building content in virtual worlds is still equivalent to hand-coding Web pages and code. They need an improved user interface; Second Life is dificult to use. They need to be decentralized, to permit creation of private spaces -- the equivalents of intranets and extranets.
Linden Lab is taking steps to decentralize. It open-sourced the client in January, and plans to allow people to put up their own servers and attach them to the main Second Life grid. They're moving to eliminate proprietary protocols. The company is driven to do this by the conviction that its biggest threat is not an existing company, but rather a future virtual world that runs on those principles.