This stuff is both great and depressing. The problem is that some companies, like Google, 37signals, etc. let their workers do these kinds of projects. Yes, some of them will bring those companies more customers, but the reality is that it's not clear if the hours put into this type of thing will ever pay off. Despite that, these companies can afford to have 20% time, or whatever it is, and let their employees mess around.<p>What makes me sad about this is how this kind of stuff is perceived by the rest of the community. People seem to follow the cargo cult of "Google has 20% time, so should we", thinking that it's the 20% time that makes Google special. I think copying this behavior until you are flush with cash is dangerous. I tried to implement it with the previous team I ran and it was more or less a disaster. It became an entitlement which you couldn't take away, but did not produce any direct or even indirect positive results no matter how we spun it. Productivity actually suffered as a result, while the 20% projects were completely unusable.<p>Has any startup been able to get away with this kind of thing, and if so how did you do it?<p>Edit: I should add that my previous team was a mess on many levels. While we had brilliant individuals, getting the company to move in lockstep in one direction was close to impossible, so there are many confounding factors to my experience. Perhaps a happy team that hadn't been terrorized by the upper management for a year and a half would have done better.