Not far off, IMO. The move to SSDs and the earlier SSD performance jump made more of a difference to most workflows than the relatively small performance jumps in Intel CPUs for a period of time. Going from my 2011MBP to 2015 was nowhere near the performance boost we would have expected a decade ago (granted, I put in a 1TB EVO SSD into the 2011, but still). There have been worthwhile bits, like video codec native CPU handling, etc., but overall - definitely slowing down on raw computational benefit for the same core(s), more adding cores + die shrink for power use reductions + some specializations, although AMD is becoming interesting again (and happy for them!).It's funny, because due to Intel's slow updates of processors and apple's past 4 years, I've completely changed my approach to buying MacBook pros. I used to buy the latest and greatest every few years, there's no point now. When my 2013 died, I bought a 2015. ...
Future-proofing seems passé!
Future-proofing, even back in the days where I'd be building new systems routinely, overclocking and playing with voltages, was still always limited in reality - my own definition doesn't change too much today:
Works great for what you do today and 'very soon.'
Works pretty well for what you'll be doing in a year or two.
Can manage, or be upgraded to manage -beyond the few year mark.
Since for the most part 'upgrades' are out of the picture aside from external TB3 storage and GPU, as we don't have external easily pluggable compute nodes or added RAM via bus (aside from cloud, etc.), assuming not on an every year or two upgrade cycle (not needed for many IMO), it still makes sense to at least consider where you are on the current system, where it's coming up short, with a bit of leeway as I mentioned.
Even when I was building systems ground up there were limits as there are today, although I do think the last really significant jump was to capable SSDs for a while now, followed by the horizontal scaling (adding cores) for laptop/non-server CPUs...a better bus, whether for video or memory, leading to performance improvements of note - will still catch up to us.
Unless you’re buying your first computer, there should be no rule. You should be looking at the performance constraints on your current system and starting there.
Activity Monitor is invaluable in that regard.
Why would there be no rule? So, someone knowing relatively nothing, roll the dice, instead of looking at what everyone else doing <inset activities here> is using and using that as a baseline, let the Apple 'Geniuses' sell them on <whatever> or worse, see some of the guys buying 64GB for 'bragging rights' but in reality are browser/mail/snapchat/FB/netflix 95%+ of the time?
Agree on activity monitor for someone knowing a bit more, though - in looking at how you really use your system 'right now.'
I also agree with the general sentiment of buying what fits now. But not for the same reason here. People here seem to go for bi-yearly updates. But in my experience, the machine that fits you now will just last way longer than you think it will. I'm fully into VMs for testing out multimedia networked setups in a broadcast context. I can run 3 VMs fine on my late 2013 13" MBP with 8GB RAM. Of course, those are rather small Win7 or Linux VMs with 1GB each but they get the work done. And the mac copes fine with just 5GB left doing office work, including the cpu smoking Microsoft Teams.
If it wasn't for the dGPU power consumption bug, I'd seriously consider only getting the top default sku 16" with 16GB at a €500 discount because it should be plenty for the next 5 years. Also, BTO is never discounted so an upgrade to e.g. 32GB RAM effectively costs €1000 wich is kinda crazy for 'future proofing'. Ideally, Apple launches a 13-14" with 16/512 standard config and I'll get that. I don't mind if it's June. i have time: my 'old' MBP' manages fine while running the tech of a broadcast company from home
I'd go with the machine that fits you now, in some cases with a bit of room for growth (see comments on growing file sizes, program requirements inflating, etc.) this can be true. If I didn't need > 16GB of RAM, I would have kept plugging away on my 2015MBP for another year or so.
Good point on discounts vs BTO/CTO options - never thought about that, and it seems you can get 'some' level of discounts on BTO/CTO via B&H, Adorama and the like in the US (even though they're proxy ordering to Apple in at least most of these cases), but certainly nothing like on the standard models. If those fit your needs, it's a great consideration.
I'm happy with the 16" although yes, the Radeon power consumption is annoying, mainly as it keeps the fans a notch over where they'd be otherwise, which means it's overall easier to get it into jet plane mode, but it's a very nice system. If I were not so desperate for 32GB of RAM (or if they had replaced the buttefly keyboard, added the ESC key in, etc.) I might have picked up an older model, or waited to see if the 32GB 13" 4-6 core 14" MBP came into existence. Working in tech in both sw and hardware, VMs, compilers and RTEs, DBs, etc. - I can still get ~4 pretty good years out of a system if I plan reasonably when I do buy, so yes, definitely at least double what some believe. Having said that, for some specialized usage, I can see the 2 year upgrade cycle making sense, possibly in professional video editing where time may be worth the upgrade, although I'd wonder if an eGPU might not extend the system lifecycle there as an option.