geez....... i didnt know that prototyping 3d prints wastes so much filament...... every day you learn something new......
It doesn't.
One of the biggest wastes of plastic is inexperience...
CAD is done in a virtual world, you can make screws without access, or worse, forget them entirely. When you work with say wood, you can test fit, hold it in your hands and see it's missing a screw hole just by looking at how the physical parts fit together. You can then add a hole, shave off a bit, etc. In cad, you often work on a single part at a time and things can be missed and then you fire it off to a printer and you can't test anything (as a whole) until it's done, and you have limited ability to make corrections without redoing the cad and re-printing it all over again. On the good side, once you get it dialed (final prototype) you can then fire off numerous ones without issue. This was what 3d printing was initially meant for, get the details worked out and then send it to more traditional manufacturing methods.
The other reason 3d printing uses plastic is that you can easily get caught in revision hell.
If you make a keyboard in (again) wood, you make it and test as you go, want a second, you start over. Want a third, start over. You might make minor changes, but say you decide to change the whole shape, or want to add a screen, it probably requires starting completely over, whereas a 3d printer, just add it to your cad drawing and hit "print" and then wait (same with CNC milling). So what ends up happening is you make a design, pull it off the printer (after a couple prototypes) and decide that you don't actually need that extra key, so you remove it and print it again, it's easy, too easy. That one comes off and now you decide you actually do want the key, just slightly moved, so you make that change and fire it off again. This time you didn't get it quite right so nothing fits so you fix it and as you do you notice a way to simplify the design and make it print faster, great, change it and hit print. That one comes off and you suddenly see that you dislike the shape or it just sits slightly off so you make a change and again hit print. By the time you get done you have 13 keyboards, some of which are usable but not great and one pretty much the way you want, you just got bored and decided it was close enough. There's probably a few print failures along the way (since it's a tough print), another ran out of filament because you forgot to swap the roll, forgot to clean the bed, etc... In the end your keyboard only used 1/2 Kilo of plastic, but you used 6 kilos getting there and most of it was your search for perfectionism that caused it.
Frankly, that's how all manufacturing and prototyping is, you find problems and fix them and do it again until you get it close enough, the difference here is you can make revisions easy, fast and cheap, so you do. If it was costing you $500 and 6 months per revision you can bet good enough will be satisfactory rather quick but when it only costs $5 or $10 and it's waiting for you in the morning you can afford to be a perfectionist and so you do.
Other wastes of plastic is cheap printers...
A cheap printer may go 10-50 hours between needing adjustment/repair/filament jams/etc. It can increase some as you get to know the machine and that's fine for small parts but when you have prints that take 18-24 hours (like a Dactyl main body) that means every 2-3 parts is going to fail or require adjustment. Now imagine doing something like a 65% tub which can take 45 hours on a slow/cheap printer, what's the odds of it completing?* God forbid you try doing a print like my big machine handles, that would be closer to 300 hours.
*Honestly, it would fail for other reasons, mainly warpage but that's
another discussion...