In theory the idea was that everybody should want to do a 20% time for the good of the company, and thus failure to seize the 20% opportunity would count agains

In theory the idea was that everybody should want to do a 20% time for the good of the company, and thus failure to seize the 20% opportunity would count against you in the promotion process. In practice though your promotion was really just graded against your main job and nobody cared about 20% time, thus actually doing something in 20% time would be bad because it would hurt your performance for your main job. Even in 2007 everybody would talk about how things were so much better in 2004 and if only you'd seen it then, and so I imagine that maybe back in 2004 you could do 20% time and it would be beneficial for your career, but you'd have to ask somebody who worked there at that time. As far as I could tell there were a bunch of very talented, very senior engineer who could basically just do whatever they wanted as 100% time, and they did a lot of really great stuff that was helpful to the company as a whole but didn't fit neatly into a specific team. But it didn't seem like that was accessible for mere mortals, nor would you be able to make as large an impact if you were only working on it one day a week.