I wrote at worst the speed can be one-sixth, which is about as correct as Apple's claim the M3 iMac is up to 2× faster than the M1.
But you attributed the one-sixth to the drive capacity, which is not correct. That's the problem. I could just as easily construct a test that shows the 512GB drive running significantly slower than the 256GB.
This isn't the same as the M3 being up to twice as fast as the M1-- that's comparing the two machines at their best. Your statement is comparing one at its best against one at its worst.
We've not yet talked about SSD lifetime and wear-leveling, for which a user should leave a large part of the drive empty to allow the SSD controller to equally distribute write commands over all cells.
I'm not sure what Apple is doing behind the scenes with their controller, but it's common to avoid this problem by moving static blocks around periodically if their erase count is sufficiently below average.
Very few users have a need for 8GB of RAM let alone 16. Swap memory usage isn't the big issue either, nor is it 27-inch displays, USB-C peripherals or any of the other topics widely discussed. The base iMacs are just fine as they are.
All of those depend a lot on what someone is using them for. I agree that more people worry about RAM and swap than should-- it's less of a problem than people seem to think it is, and people attribute every hesitation in their machine to RAM because they've become fixated on that one idea.
Still, there's no reason to add to that confusion with more bad information.
And those who pay up $200 for 16GB RAM are likely seeing no gains at all. Only a few seconds the one day when they cut a wedding video or something.
It absolutely depends on what the machine is being used for. If you're working with large datasets, it can make a huge difference. If you're just worried that a Chrome Tab you haven't viewed for an hour got pushed to disk then it's not making much difference at all.
He's making a huge mistake by only thinking about SSDs in terms of storage capacity and not system performance (and longevity). He's making a small mistake by thinking 40% free capacity is space than can be filled with data, without detrimental effects on performance (and longevity). And he's making a bigger mistake by already having decided for 16GB of RAM as the ideal way to future-proof his purchase. This money could buy him the fastest TB drive in the world or double the internal storage and maybe eliminate the need to manage external storage completely. Not to worry about running a lean system and which data to store internally or external is a quality of life improvement.
They explained their use case pretty clearly-- they aren't close to filling the drive and plan to use cloud storage as needed. They already have a 1TB external drive available. There's no indication of what they use the machine for so I can't judge whether the 16GB of RAM is worth while.
Taking the multidimensional view you're proposing is the right approach, but you're over dramatizing it and supplying bad data to support it based on a video made by people who, frankly, seem to know less than you do about system performance.
