I use both.
Depends on what I'm making. I tend to brew a lot of pale ales, IPAs, APAs, and so on with US-05, and I tend to brew a lot of stouts, brown ales, and porters with S-04 (or other dry yeasts such as Nottingham or Windsor).
On the other hand, if I'm making a Belgian ale or something more "interesting" (such as a blend of two different yeast yeast strains or something more unique/unusual), I'll use liquid yeast.
They both get great results. Just depends on what you're looking for. The main advantage of liquid yeast is variety. Dry has a much more limited range of yeast strains available, but there are some pretty good ones there.
Not true at all. About a year or two ago, I bought a packet of S-04 from an LHBS and pitched it on my brew day two days later into well-aerated room temperature wort (around 68F) with Servomyces, and 4 days later, the fermentation still hadn't started. Up until then, I had just thought it was the slowest fermentation ever (normally fermentations for me start within 3-6 hours after pitching) so I went to the LHBS and grabbed a new pack of S-04, pitched it, and the fermentation started less than 4 hours later.
So, although dry yeast fails less often than liquid yeast, it does fail sometimes. (I'll note that I've never had a vial of liquid yeast fail, so if I was basing my conclusions on anecdotal evidence alone, I'd think liquid yeast is less likely to fail than dry -- not true, of course, but I think you get my point. Dry yeast can fail, and it does sometimes).