[swyx]

(OP here) thanks for the response!

so the tricky question is - why stop at 20% - why not 40% - or 80%? could you make a case for that?

[tigger0jk]

You could argue that Valve' s stated policy is 100% time. I'm not sure what the reality of it is but it seems like they really do want to let you work on whatever, although it might be more directed towards things that are valuable for the company and that you're good at, not just anything you desire to do on a whim. Employee Handbook: https://steamcdn-a.akamaihd.net/apps/valve/Valve_NewEmployee...

[wtracy]

Both Google and Valve use stack ranking for performance reviews, which is a pretty brutal system.

So the picture I get of Valve is that yes, you can pick whatever you want to work on, but choosing poorly will bite you in the ass.

[mondoshawan]

It depends on the need/success of the project. My first 20% project was a disconnected caching filesystem back in 2005 and it was rapidly shifted to my 80% project as it solved a real need in our infrastructure.