A platform is either growing or it is dying.
It's true that there will continue to be work in Flash and Silverlight and AIR and XAML and Java Applets and Fortran and COBOL and Deplhi and Windows 3 and OS/2 and MSDOS and VMS for years yet.
The real question is are new developers coming to the platform? If not then new programs will slowly stop being written it, and the only work will be legacy. The good thing is that you can make good money doing that if that is what interests you.
And Adobe has clued in to the fact that Flash development is well past the inflection point on the way to decline. They could ride Flash into the ground or they could change horses when they're still doing well. If Adobe lets another player dominate marketshare in the new ecosystem while they are distracted with squeezing the last drops of blood out of the Flash turnip then they would miss out on a huge opportunity and suddenly be in the worst strategic position imaginable.
Right now, before Flash is well and truly dead, is the smart time for Adobe to transition and try to plant their feet firmly in the territory where it appears web development is heading (html5).
As for me, I will be coding in Flash for a long time in the foreseeable future. Flash is huge in the UX research space and I don't see HTML 5 replacing it any time soon unless the tools reach the sophistication and speed of the Flash IDE. There isn't even a standard graphics library yet for Canvas. How long did it take for the industry to standardise around jquery?