In a recent trac ticket I brought a core modification up for suggestion.
In turn, it was kicked out and recommended to be brought up as a modification for WP, which would then filter down to MU.
Unfortunately, my opinion differs on this one, and here is why.
With WP, how many users (by percentage) actually have so many themes lying around that it would be a benefit to them? My guess is that there would be an extremely small percentage of users keeping an untidy house like that.
Now with MU, it's almost a given that you're going to have dozens of themes if it's being used as intended. That being the case, you now have hundreds, thousands or more little blog admins hitting that page. That's something that is more of a majority in MU, and multiplied by the number of users. Needless to say, a single page loading that many screenshots (which let's admit that most theme authors have really large file sizes on their screenshots) just kills your bandwidth and load times. Therefore discouraging the end user with an unpleasant wait time and experience.
I'd be happy for anyone to comment on this, and whether you feel it's a WP or MU issue.
Topic open for debate.