I understand discontinuation in the case of something like itanium for being relatively low volume entirely enterprise hardware of which most has been scrapped and it's hobbyist community long gone, but i468 had massive volume in the consumer space and there genuine purpose in keeping it and other widespread preenshittification tech alive.
This decision shouldn't be made on grounds of hardware speed. Most should agree that pushing up the bar for hardware just because your software runs poorly will lead to suboptimal programming. Speed is not the burden that should be weighed against. It's the testing and patching overhead for architecture quirks
I argue that the benefit is not on the basis of how many people currently using modern Linux on these machines get updates. Free software should be developed and maintained on to basis of the potential to be used in place of nonfree software; both in the present and future. Quite a few of these machines are still in use today but do not use linux because it is not good enough to replace nonfree software. Many more of these machines are no longer used but exist in storage because they are of good quality, but there are not software solutions to make them practical. Moving an entire class of processors to LTS is to say that their exclusion is inevitable, which means that at some point these machines will be unusable with untrusted data and software. This is actively throwing in the towel and saying Linux and its surrounding software ecosystem will never get efficient enough to revive these beautifully basic machines. These machines will never be usable in the modern era and I don't like that.
