A few points:
1. "EOL" has no standard definition. For a Temu "Gerberlink" bought today, that could be next week, while for a Unifi device, it could be a decade. This is why there needs to be some sort of standard created for this, because "what is profitable" clearly isn't delivering what's necessary.
2. The liability, as we are discovering, of having these old, vulnerable pieces of hardware on the internet, is massive. You aren't going to be able to force Joe Homeowner that his 3-year old Douche-Link from Dollartree needs to be replaced because the new Giga-douche 6000 is out, so I think the appropriate solution is to force the companies to provide timely updates to address any known CVE's within a defined period for what is determined as a standard for "useful life", which I think 10 years is probably appropriate.
This sounds worse/more complicated than it is. Most of these devices are just running customized versions of busybox with their own GUI/branding, there isn't a lot of effort put into the development. A long-term lifecycle requirement for firmware support would force them to standardize across products, which means better, more tested firmware on everything they produce, very similar to how Apple does IOS or Cisco does their IOS, or how Unifi does their firmware.
While perhaps painful for these companies at first, the end result is better for everyone.
But if the new devices have the same problem (and they will) then nothing is being addressed. We actually need some sort of well defined regulation as I've described above.
I agree, and it doesn't solve the problem. While it might prevent the proliferation of backdoored devices from foreign nation-state sponsored actors, it completely fails to address the issue of CVE's present in the millions of devices already in the wild, and the lack of robust and long-term support for firmware on devices of Western origin, which have the same problem as the Asian ones.