We're in such a wild time because it feels like Macs, even decade-old ones, are holding value way better than they did like 20 years ago.
The Plateau makes things so wild, too. Like, any Mac Pro 6,1 with a quad, 16GB of RAM, and working GPUs is arguably "on-Plateau" in a way that 2012/14 Mac minis with dual-core CPUs might not, at least for anyone with web-heavy and multi-tasking oriented work.
A 2012 quad-core Mac mini tiptopped with 16GB of RAM and a decent SSD seems like it "should" be basically fine for basic desktop usage, but I can 100% see how upgrading to the higher end system for multmedia authoring, technical computing, development, would be more appropriate.
Not to put too fine a point on it but: This is identical to what the state of play would have been with these two products in 2013 when the Mac Pro 6,1 was new, with the caveat that basic desktop usage would've been way easier on a duallie mini in 2012-13 than it is today.
But not only will the M1 outperform almost all Mac Pro 6,1 configs, the 2018 Mac mini will, core-for-core, outperform a 2013 Mac Pro (albeit less dramatically, it'll be like a couple percent probably). (So, you can make an argument about 8/12-core models, and/or any workload that needs the dedicated OpenCL grunt on the FirePro, ahead of just adding an eGPU to a 2018 Mac mini.)
It is
wild how good the newest Apple Silicon Macs are, and, genuine "finally" on 16GB of RAM being baseline.
Goes to show how uninteresting Apple has become. They release a new system and nobody seems to care.
This is a you thing. Almost everybody else noticed and there was much discussion about the meta of it. I'm pretty sure there was even a Lounge thread about the announcement, and likely a discussion about how the Mac Studio is the same computer but for half the money.
Apple Silicon RAM is not just soldered but it's on-package. In so doing, it's some if not the fastest RAM in the industry, with the highest memory bandwidth available anywhere, as far as I know off hand.
Flexibility comes at a price and sometimes that price is speed. The hot reality of it is that Mac Pros especially are professional computers used by professionals in their speed who need that speed above absolutely all else.
Entertainingly, the SATA ports are still present, so you can still add slow or slightly-less-slow bulk storage if needed.
Ultimately, for Apple, doing the right thing by their customers is providing the fastest possible computer. I would agree that their upgrades, especially on the Mac mini, which is hands down the best $600 in computing right now, cost too much, and that they don't make it easy enough to use external storage for certain situations where it would make sense.
But alluding to what joshc wrote, we've been doing this whole "computing" thing for a long time by now, and any reasonably technical customer who has some reason to buy a Mac, should be reasonably capable of measuring what their workload needs, and predicting how different options in a new computer will impact or improve their work.
If you want to get real silly, I would actually argue that, in terms of the environment, knowingly buying an underspecified new computer and then immediately installing a bunch of upgrades is bad, as it creates ewaste, as often the baseline components can't easily be reused elsewhere and might not be suitable for, say, upgrading an older system.
I have tens of thousands of words on this thought process but the short version is that historically speaking, running brand new commercial software on 10+ year-old hardware, with it's original CPU, is totally unprecedented. It's so nice that we have this option, but we'd do well to remember that this basically wasn't reasonable or possible a decade ago.
The rate of change on needs for basic desktop, productivity, and web usage has slowed significantly. I have no trouble whatsoever believing that a baseline M4 Mac mini purchased today for basic desktop computing will still be reasonably useful in a decade, for basic desktop computing.