What's happening is that vendors of disk-to-disk arrays and other disk technologies are gunning to create market momentum, and a simple way to do that is to carve a rival technology's gravestone, tout yourself as heir apparent and hope no one notices the holes in your strategy.
Tape is an easy mark for this tactic: It's boring as brown dirt and one of the oldest technologies still in widespread use. We can't blame vendors for trying this ploy. By their nature, technologists--myself included--are always on the hunt for the latest thing. But we all need to resist the tendency to write off older technologies that have lots of life left in them. Tape backups still have an important role to play. Maybe it's the peskiness inherent in older tape backup systems that makes us so willing to bury the technology alive. Backing up to a single tape drive, usually one for every machine, meant lots of manual intervention, tape swapping and hardware to maintain. Even the advent of autoloaders and full-blown tape automation, which let you back up all your servers from a central location, haven't made life as good as it should be.
Restores are a major pain, especially for those good doobies who store backup tapes off-site like you're supposed to. Then along comes cheap, ATA-based storage. It's not only quicker, but it also facilitates restores and elongates the backup window while leaving production systems unaffected. Makes you feel like you've died and gone to heaven.
Hold the phone a second. How are you going to get all that precious backed-up data off-site? Unless the guy lugging your tapes to the storage facility is built like King Kong and you've got the coin to buy all that disk a second time, it's not going to happen. Proponents of the "disk everywhere" strategy have a pat answer, of course: "Set up your disk array off-site and add connectivity."