Say you change your workload model there might be ~20% improvements between bare metal, virtual machines and containers. If you simulate a part of infrastructure using containers you may not need more RAM, but more CPU would be nice. But when you then want to do a lot of recording/capturing and process that, RAM gets more important. Just upgrading the RAM wouldn't help much because without a CPU to generate the data you might as well offload the whole thing.
This is definitely an echo-chamber/bubble point of view. The vast majority of users out there don't even know or care what AVX is, don't use external GPUs, and don't know or care about the difference between Thunderbolt 2 and 3.
If you personally need these things and want to buy a new machine every few years, then that's great, you should do that. But there are a ton of people who would benefit from an easily-repairable, easily-upgradable (RAM, storage) machine that end up dropping $1500 every few years instead of the couple hundred they could instead spend for a reasonable upgrade.
While I bet that there are a lot of people that do want to modify their systems, they are such a minority that it's not very logical for a large multinational to invest in that to the detriment of other goals. It might simply mean that you are not the target audience for their product(s).
Some other manufacturers/brands have the same, while others do a mixed portfolio to cater to smaller groups as well. We also have large manufacturers that cater to the classical enterprises which still run on the old idea that you need a fleet of identical machines and then swap out parts all day long, so machines that have facilities for that exist. Most notably Lenovo, HP and Dell do that.
What does this mean? Do you mean your current device starts working poorly (but that would mean "existing device degrades per se", which you ruled out) or that the existence of a new device automatically makes you perceive your device as being of "degraded productivity"?