First, if you haven't, go read Mitchell's Thunderbird post. Then come back and read mine. I'll wait :-)
I moved over exclusively to Mozilla Mail (the predecessor to Thunderbird) before even the core developers would trust it for their daily use (ask mscott and sspitzer, it's true.) That was 1999 and I've been using nightly builds ever since. No mail client, desktop or web-based, can rival Thunderbird for my specific needs (though PostBox, a Thunderbird-based product, was close until they changed the vertical view.)
But I'm part of a decreasing population. Every year, more and more of even my "oldschool" friends and colleagues have migrated to to web-based solutions. Some are living mostly in web alternatives to Thunderbird like Gmail or Y!mail. Others are using social networks like Facebook for just about all of their communications. Yet others are mixing and matching three or four different messaging services. Regardless of the specific choices, the trend is clear to me and has been for years. More and more people are doing messaging on the web and fewer and fewer are doing it from a legacy email client.
If Mozilla had infinite resources and focus, I'd want them to keep investing heavily in Thunderbird -- even if I was the only Thunderbird user on the planet (I did say if Mozilla has infinite resources.) But I have first hand experience with resourcing and focus trade-offs across Mozilla projects and products and so I understand the shift to a sustaining mode rather than an innovation mode for Thunderbird and I think it's the the right thing to do.
I will continue to use Thunderbird for as long as it is secure and stable or until a better (for my specific use cases) solution comes along. I'll continue to test nightly or any other pre-release builds and file bugs for any regressions that creep in and any issues I think should be fixed for stability and security. I may even find a way to contribute to some front-end polish issues that have been bothering me for a long time.
If you're like me, and you just aren't ready, for what ever reason, to move your messaging to a Web application, I encourage you to get on testing builds and call out regressions when they happen and if you've got the skills to do it, make some patches and help keep Thunderbird moving forward.
Also, I have some UI mock-ups that I think would really give Thunderbird a nice facelift without any serious functionality changes. If you're a JS hacker and willing to learn a bit about XUL (or already know it,) and you want to help improve Thunderbird's looks, I'd love to chat.