To say HTML5 can go down the same dark path flash did in the realm of performance and bloat is sorta a silly misunderstanding. It would be like someone saying "Oh well C++ has gotten so bloated over the years". Adobe makes a product and authoring tools for that product. They expand it and they add new features to their authoring tool that makes the fast majority of Flash content.
Flash still has some advantages now. They come in discrete, independent files which can be used as standalone executables, for example. You know things will behave the same on any platform. But since html5 is a very open standard, as soon as someone comes up with a good way to package something made in html5 as a single, easily embedded, clean entity, it'll happen. People will want this to make it easier to upload games to places like kongregate easily and more securely. But if I were to guess, I'd say there are probably ways to do it right now. The question is more about which implementations will end up being the most popular and well supported.
So while Flash and HTML5 cover a lot of similar functionality they are very much NOT analogous platforms and there is little reason to expect them to share similar histories. HTML5 is a specification, flash is NOT. Flash has a single vendor, HTML5 either has no vendor or many vendors depending on how you look at it. It would be like saying HTML4 and PDF are similar "Things" with similar histories.
But anyways, I hate webgames! So whats the vague timeline for exe-export? If it's anywhere within the next year I'll be thrilled.