Developers will be limited by the lowest common denominator.This is always true, but it also is always qualified by "of the audience they intend to reach". If you're trying to reach "everyone", then your lowest common denominator is still going to be some version if Internet Explorer for some time to come--even IE9 won't support some major parts of HTML5. Beyond that, you're implicitly excluding some set of potential users -- it's only a question of which. Targeting a plugin (or bleeding-edge standards) never gets you a bigger audience than standard web technologies, just a different audience.
(Flash is really the only historical example of a plugin becoming so ubiquitous that you could safely assume that you weren't excluding many by depending on it, but that is obviously changing back to the more typical state.)
What this should mean, I think, is that the bar for developing on plugin platforms is raised. The smaller the installed base, the more valuable the experience you have to deliver for it to be worth your while, and the user's. I think this is as it should be. The only developers who will choose to limit themselves so will be the ones that weren't really doing anything that necessitated the plugin in the first place.
Don't like that Apple has chosen H.264 over VP8 or Theora?
If you don't like that, you probably aren't using Mobile Safari in the first place.