In fact, a growing number of submissions can be a problem for a social news aggregator. Is it fair to assume that curation capacity will scale linearly with submission quantity? Perhaps not. Even if a linear growth in submissions represents a scaled growth in visitors and user accounts, curation capacity will only scale if new user accounts have the same curation appetite as older user accounts.
If you get lots of new submissions and lots of new visitors but a smaller number of new people voting on submissions, average quality of front-page material will suffer. It's possible there will be a positive feedback effect if older users with a higher appetite for curating content are more sensitive to a decline in quality and thus more likely to leave the site, exacerbating the problem.
When a social news site is young and small, (non-spam) submissions are a Good Thing. But as a user-driven site matures, it's much more valuable to have better submissions rather than more submissions.
I hypothesize that one solution to the seemingly inevitable decline in quality of large social sharing sites is to introduce a cost to posting new content. A trivial example would be scaling the minimum time between submissions to the user's karma for previous submissions. (For this to work without incentivizing spam via sock puppets, there would have to be heavy restrictions on posting content for new users.) The time-between-submissions could be large for less proven accounts - think days rather than minutes.
There are problems with this approach, but they may be better problems to have than the alternative.