This page might change but it's a good place to talk about something I've noticed and have come, over the past couple of weeks, to believe.
This is editorial.
For the past several years, I've talked about Mac lifecycle in terms of how long a particular Mac runs the current version of the OS, and ultimately ended up changing that to how long a particular Mac runs what I called a “supported” version of the OS, which as of this writing means Mac OS X 10.x-2. (Or: the current version and two major versions back, which is probably a better way to put it, what with the )
So, right now as I'm writing this, Apple is hypothetically “supporting” Mac OS X 10.13, 10.14, and 10.15.
When Mac OS X 11 launches this coming fall, the official support list will be Mac OS X 10.14, 10.15, and macOS 11.
This is only part of the story though. When macOS 11 launches later in the year, my 10.13 machine doesn't suddenly stop working. In fact, almost none of the software for it will stop working, certainly not immediately, realistically not for years.
At this moment, mostly everything needed to do remote work at my workplace still runs on Mac OS X 10.11, which runs on hardware from 2009 and newer. Firefox, (but not the VPN client we use) works as far back as 10.9. The Microsoft Office document formats haven't changed since 2007, so these older OSes have versions of Office that still work.
If you don't need to use your computer “for work” then you can keep going even longer. A surprising amount of stuff basically just “still works” even on machines as far back as 10.6 if you have the newest availability build of Chrome of Firefox. (This is not counting enthusiast builds and projects like TenFourFox or whatever might be out there to run newer Firefox versions on older OS X versions.
The other thing that exists, and lots of the machines now being dropped (which are minimum 7 years old as of this writing) are eligible to run newer versions of macOS using a patcher utility. I have used this on a MacBook Pro from 2011 to run Mac OS X 10.15 Catalina and it's “fine”. This isn't something I would recommend doing on something that was critically important, but it will give a secondary machine or a machine you (as an enthusiast) can support directly. I would not recommend doing for a machine you can't support directly.
So, I think the story here is more complicated.
Apple Silicon will end up being a confounding factor. If Apple Silicon is “a little bit faster” than Intel, then I think we can expect this pattern to continue. However, the last couple of platform transitions were very much not kind to the outgoing Macs.
The 68k → PowerPC transition took several years to finally finish. The last of the 68k Macs were discontinued in 1996 and they were almost immediately outdated. Office 98 is, I would argue, was the biggest problems for anybody who had a 68k Mac in 1996 or later. Introduced in 1998, it didn't run on any 68k Macs. The web was also very unkind. As soon as multimedia plugins became popular, 68k was dropped from their functionality.
The PowerPC → Intel transition was much better in terms of compatibility, but by the time 2012 rolled around and 10.5 stopped receiving security patches, the machines were already way too slow to be using daily. People did it, but it can't possibly have been good. The first Intel Macs were twice as fast as the outgoing PowerPC Macs at every single level. Often, 4x or more as fast as per-core speeds boost and almost every tier of Mac boosted core counts. Plus, long-standing problems like the continued use of 4200RPM disks were finally solved with that movement.
Apple has committed to “several” years of support for Intel-based Macs, but, we need to see ultimately what it is that means. I strongly suspect, however, that lots, if not most, software will still work for several years in the future, which is basically my point here.