It would change almost nothing at all for me, and probably a lot of other programmers.
First, true, because a lot of it would be abstracted away. Quantum computers will catch on in the marketplace at exactly the same time somebody releases an x86-compatible quantum computer. Yeah, it's going to be the biggest collective facepalm in computing history, but you know it's going to be true.
Second, because the kinds of things that most programmers work on, are not the kind of things that have anything to do with the problems that quantum computing can solve. Just look at the questions on SO right now, and imagine how many of them would be made any better by quantum computers. Regular expression not working? Need help with XML schemas? Trouble getting your CSS to work in IE6? (Well, if your users aren't upgrading from IE6, they sure ain't gonna upgrade to a quantum computer! You could serve up that CSS file from a quantum computer, but I can already buy a web server that's fast enough for any number of users I'll ever see, for pretty cheap.)
Me, I've got Excel spreadsheets to parse, I've got HTML and CSS to generate, I've got unit tests to improve. I'm no expert on quantum computing, but nothing I've seen has led me to believe that quantum computing will make people organize their spreadsheets better, or upgrade their web browsers more frequently. The web is today's big platform, but I think the effect of its simplicity is not so much that it's fast for computers to parse, but rather that it's simple and universal enough that people can grok it. I don't see a way to make it significantly better that would require what only QC can provide.
Finally, other technologies which promise big changes in performance for certain tasks remain niche, even after years. My database servers aren't doing GPGPU. My web server isn't running on an SSD. Lots of people have figured out independently that it's a lot cheaper and easier to string together last year's technology (in parallel, if you need the performance/reliability) than it is to adopt the latest and greatest new paradigms right when they're introduced. Most companies would rather keep old systems running on COBOL and FORTRAN than port them to something new, even if it's radically better.