Incubators Hot Again But Unlikely To Get Burned

Ravikant runs a tight ship, with hard deadlines. "The engineers have the freedom to experiment, but they have 90 days to ship a product," he says. "The product has to grow organically, without any marketing." Those that are still growing on their own 12 months later get to live.

They also receive more funding, and Ravikant plugs them into a distribution network he's building. Web products that don't hit their growth numbers get killed -- a far cry from the old incubator system, which kept bad ideas alive for years.

Another key difference is the cost of failure. Since a Web 2.0 idea can get to market for as little as $50,000, Ravikant can afford to try things, tweaking products after they ship, in a way that even the most successful incubators (such as Bill Gross's Idealab) never could in the dotcom era. That might prove to be a model for more than just incubators.