It is possible for there to be no problem with the design and still get large numbers of failures. It could be that the assembly procedure is no sufficiently well specified and/or controlled. It could be that the glue specification is not suitable for some use situations, or that the wrong glue was being used. It could be that a certain finish was specified for the contact surfaces but the parts supplied didn't achieve it, and so on. We just don't know enough and Shimano will most likely obfuscate in order to hide what really happened.
Fair points, although probably more to do with semantics than principle. The bottom line is that these products have been in service for years, over which time a significant amount of potentially very dangerous failures have been well documented in the public domain (and presumably this coverage has also reached Shimano - directly or otherwise) and it's taken them this long to issue some form of remedial action following years of "la la la fingers in ears not listinging all's fine don't know what you're talking about mate".
I don't think they've done themselves any favours re. the difference between advice in the US ("cease using") and everywhere else ("crack on, we'll take a look and if it doesn't appear shafted crack on; having signed away your rights should the unit fail in future").
Tbh as a generally paid-up Shimano fanboy I'm disappointed by this outcome / the way this has been handled and tbh expected far better of the company.