Exactly. You have a specific task and probably specific software for that task. If the software supports CUDA then you might want to spend money on Tesla cards over CPUs. Does it use Open CL? Then you might want to look at AMD GPU compute cards.
Do you need a large memory space?
Do you need a lot of threads or just a few really fast ones.
If you have 50k for the system then I suggest you spend a little of it on someone that really knows this subject.
It may make more sense to just use Amazon E2C.

Actually I knew people buying CP/M machines as late as 85 actually I knew some vertical markets that sold CP/M machines well into the early 90s.
Truth is that MS-DOS was not a lot better than CP/M for many years. It really was not until Lotus 123 and WordPerfect came out that MS-DOS was a lot better than CP/M. That combined with the price drop from the clone makers and you finally had the death of CP/M.
However by 1985 you had the Atari ST, Commodore Amiga, and the Mac. All of which were far better machines than the IBM intel based PC.
I blame the decline of those machine in part to the magazines of the day. They lived and died by ads so could never say the PC was really outdated. It is simple math. Do you want ads from IBM, Compaq, Kaypro, Corona, Sanyo, and all the rest of the clone makers or do you want ads from just Commodore, Apple, and Atari?
Pushing PCs meant more ads.

You left out that they failed when they did not listen to the husband/techie/mans advice and produced a PC compatible.
Had they done that they might have been Compaq.

Actually it was not Commodore that drove down CP/M prices. That was Osbourne then Kaypro.
Commodore is the the tragedy because Wintel PCs did not really match the Amiga until around 1995. The issue was marketing and expectations.

Not exactly.
It is pure evolution in action stranger=danger. It is a biological adaptation to trust the people that are most like us. You trust your close family more than your more distant family and so on. The easer to see the difference the greater the distrust. From an evolutionary point of view the amount people look, smell, and act alike reflects the amount of common genetic material they share. So you care more about their survival and they care about yours.
Frankly the fact that humans have come so far in ignoring this instinct show just how important community is to humans.

MS-DOS never ran on an 8080, because MS-DOS didn't exist for the 8080, because the first IBM PC was based on a 4.77MHz 8088 and MS-DOS as we know it was not yet a thing until IBM approached Microsoft about an operating system for said IBM PC.

Meanwhile, an 8086 was a 16-bit CPU with a 16-bit bus, and the (somewhat later) 8088 was a 16-bit CPU with an 8-bit bus. They used the same (16-bit) instruction set.

tl;dr There has never been an 8-bit IBM PC, therefore there is no historical reason for 8-bit MS-DOS.

Well, yes. Voters are no really working insurance against such catastrophes. While Hitler had a minority government, he was voted into office as Chancellor. Without that he would have it found far harder (or impossible) to take over the state. If the mood among the voters is right, something like that could happen in the US as well. And if at that time, checks and balances have been eroded enough, and law enforcement has gotten used to not being held accountable for what they do, the catastrophe is there.

Remember that it was not raging SS hordes that sent most Jews to the KZs, it was mostly ordinary police following orders. These things always happen over a while and today they are happening slower than ever, but the US has already tried out how to torture people, how to hold them forever without due process and police accountability is at an all-time low. I am also sure the NSA has nice little lists on who to arrest and who to shoot (while "attempting" to arrest them of course).

Only that they have those in power behind them. When law enforcement is not bound by law anymore, that is a police state, the precursor to a totalitarian state. The signs are well-known from past occurrences, as is the further story: Unless constrained very tightly by the law again, these people will eventually cause a total catastrophe. Checks and balances are not fluff, they are essential to keep the likes of these people in check.

Indeed. Until we have advanced robotics to that stage, real space-travel or colonizing even only the moon is completely illusory. May take a few decades though and maybe even a century or two.

And "actually secure"still fails, because every user carries two camera-like device and quite a bit of attached storage with them. And they will put sensitive data into that storage, because it is their brain.

The only thing that can make you really secure is if you can trust your people. No amount of threat and repression will accomplish anything that is actually secure. Your "actually secure" scenario just makes sure security does usually not get broken by accident. Mostly. And of course, many smart people, and in particular creative thinkers, will refuse to work in such an environment, so it degrades work quality significantly and sometimes catastrophically.

Yeah, do that. And while you are at it, why not fire anybody smart enough to get any work done? Because these are the people that break the rules because they have to in order to get results. If nobody does any work, then nobody has any incentives to break the rules.

What an utter fail.

Very much this. Also, people want to get work done, and are in fact obligated to do so. If IT security is standing in their way, then they work around it. This is a well-known (to people that actually have a clue, many in IT security do not) "insecurity caused by security measures" effect. Example: People cannot sent email with encrypted data. Hence they send sensitive stuff unencrypted and most still gets past the scanners. That, of course, makes things worse. Or "passwords must be changed every 4 weeks". Result: People use the worst passwords possible and write them down. Not good.

When done right, then restrictions must always be accompanied with a workable alternative to get the desired effect. If the alternative is too much hassle or does not work well, people will break the rules.

I work for several customers where it might be a thing to forbid smartphones, but they do not, essentially because they understand that such a prohibition would do more harm than good. IT security by Authoritarianism (forbid everything risky) basically always fails. Either it does not achieve its security goals or it kills productivity. The former is the typical thing in the private sector, the second is what typically happens in government.