8080, hand built 8088, Ibm XT, 286 ps/2 - all very used in the mid to late 80's
Apple ][, PDP11/VAX (not personally owned!), home built wire-wrapped 6809, 68k Macs, PPC Macs G3 - G5, Linux Pentium Pro 200 then Athlon 700 then Athlon 3200+, Core 2 Duo Macs, Hackintosh i7-860 then Linux i7-4790K, i7-6700K. Various ARM boards (Pi, Odroid) and Arduinos. Now RISC-V HiFIve1 (Arduino) and HiFive Unleashed (1.5 GHz Quad core).
Now running X99 5820K with all the trimmings - 8th Gen I7 on the wish list but I'm happy with what I have now
Definitely still a great machine, with no reason to replace it unless you want 10 - 18 cores (i9) or similar performance in a smaller lower power package.
I'm very impressed by my new 100mm x 100mm x 35mm NUC with an i7-8659U CPU. I can put it in my pocket and travel the world, it uses 15W, and the performance matches a 2.5 years old 65W i7-6700 desktop CPU! (and beats it by 20% on single-core tasks).
Seem to have lost my passion for computers the last year or so, possible to do with no major jumps in tech
Wow! Even at 55 years old I'm the opposite!
Clock rates have peaked, yes, but that makes things MUCH MORE interesting than any time since 1990! It's no longer enough for Intel to simply spend their billions on a new FAB process and lazily crank out the same design as last year but at 50% higher clock speed.
Now they have to actually *think*.
So far Intel have mostly been thinking "Lets put more of the same old cores on on chip". That's reasonably exciting. It makes tasks such as compiling software or video transcoding or responding to a lot of web requests much faster without a lot of effort, but it's a very interesting challenge to figure out how to use this for other things.
But even more exciting is that YOU DON't NEED TO BE INTEL to innovate any more. Basic CPUs aren't getting faster any more which means that speeding up particular tasks often now requires special hardware to directly implement the main part of the task, with a conventional CPU there as well to supervise and control it.
There is an absolute *explosion* happening in people building this special hardware inside FPGAs, and the successful and high volume designs migrate to custom SoCs. Sometimes this special hardware can be done as a co-processor with relatively loose coupling to the main CPU (and you can license ARM cores for that), but often it works better if you can hook your special hardware up to custom instructions in the CPU and incorporate it in the normal program flow in a much more fine-grained way. You can't do that with ARM.
Some companies are creating their own proprietary and customisable CPU architectures and that's exciting to see that happening again (like in the 70s and 80s). But even though gcc and linux etc are pretty portable it's still a lot of work to actually port them, and that work doesn't actually add any unique value to your company. So more and more people are grabbing an existing open-source and license-free processor design and customising that with their own special sauce.
Mostly now, that means RISC-V, which is getting very good momentum.
And so I've made my way to a RISC-V startup where our business is making customised CPUs and SoCs. Exciting times!