Beyond the well reasoned response, I object to the premise that it's "just backups". Quality backups are a big ug lever for ensuring that others work is able to continue and exist to be appreciated.
Yeah, but generally the reason people's backups don't work aren't that they're using backup software written by programmers who weren't smart enough to win the Putnam. Generally when people's backups don't work it's because they don't do the backups, or because they set up the backups in a way that can't work (for example, forgetting to back up the files they really care about, or backing up a live database tablespace with a filesystem backup tool without snapshots), or because they're in some kind of a dysfunctional relationship with a vendor like Google that doesn't give them programmatic access to their own files so they can't do backups.
So, while I don't think it's bad for cperciva to have spent a lot of time working on Tarsnap, which is clearly a useful piece of software, I also think things like scrypt actually matter more in the end.
However, I also think cperciva is a better judge of what to spend his life on than I am.
In a sense, the difference between entrepreneurs and (idealized) academics is in making what people want right now versus making what they think people (in the future, might) want. So, trusting other people’s judgement over your own, including how to spend your life. Modulo Steve Jobs and lots of unearned capital.
