SL _should_ be a Flash killer but it's almost dead in the water
SilverLight is beautiful, delightful to use, works from ANY webserver and offers a range of services, functionality and client/server integration that Flash can only dream of. So why is it going to fail? Because MS are leaking what should be Alpha versions onto the net and promoting it (like the original AJAX extensions) in a way that suggests it's a developer's weekend hobby project that the marketing department have suddenly gained an interest in.
So instead of proper databinding, version 1 has Flashy (note capitalisaton) video processing and filters. Instead of datagrids we get arty-farty 3D image galleries. Instead of a usable presentation layer for server-side code we get silly pointy-clicky-drag-round-the-screen-ooh-it-looks-like-we're-in-Torchwood navigation that no one really needs.
Make it run on Windows Mobile, release some FREE tools for C# SilverLight development, give us a WYSIWYG XAML editor, let me create Silverlight assemblies in VS2005, stop releasing half-arsed betas and promote SilverLight as a serious competitor to Flash and it could (and should) find a very good share of the market. As things stand, it'll be dead within 18 months. And deservedly so. Because it sucks. Big time.