Yes, its true that HSM has been around for ages, but it's much better integrated now, with the arrival a few years ago of LTFS, which enables the tapes to be standalone (file metadata is stored on the tape with the files themselves) and portable between systems while still being able to form part of an HSM solution.
It's actually quite a cool innovation, if you can work out how to use it.
Of course, it does not prevent you using recent generation LTO tapes and drives as raw data storage under the control of, say, TSM.
Why do we have to wait for 15 years from demonstration to product? Instead of tinkering away and producing tiny improvements in mass manufacture, why not concentrate on introducing LTO12 in 2 or 3 years time? I want a 220TB tape cartridge and I want it now! Is everything held back by marketing departments? How about a drive using 8 x 16 Gb/s heads to fill the tape in about 2 hours? Many of us could go back to feeding the drive every day instead of using a tape library. Even better, let's have a longer tape in a bigger cartridge, which we change weekly or monthly.
To put it another way, what makes "Moore's law" perform so badly?
Backwards compatibility is vastly overrated
Every lto purchase I've made has been a 3 generation hop.
Most people who use large scale lto deployments are in the same boat - you buy a robot with drives and 5 year support. At the end of that period you buy a new one as maintenance past that point is more expensive than buying.
Data migration if needed streams from old tape to disk to new tape. After that theyre burned. We don't bother migrating backups. Just keep the old kit around until the tapes expire then chuck the lot.