Not necessarily. There is talk of energy being 10,000x more abundant for humanity if we were to put development into the LFTR reactor. If we have cheap electricity via safe nuclear power, then using some of it to generate fuel from sea-water is surely a lot better than putting the effort into getting it out of the ground and then shipping it half-way around the world.

Then again, with cheap nuclear power, we can also effectively supply hydrogen (which is obviously much cleaner) for other internal combustion engines.

Um. When safety standards change in the industry I am in (Mining) they do actually require everyone to immediately replace what is "working" with something deemed to be safe against the particular mode of catastrophic failure that has been observed.

XP can be made safe if it is kept patched or isolated from the network. Choose one. If it is isolated, who cares. If it is not, then you're simply rolling the dice.

Add the cost of re-training, software compatibility testing, a pilot program, etc. and those costs will blow out MASSIVELY.

Anyone in IT worth their salt knows that the software license cost is a tiny part of the TCO or cost to change. There are huge amounts of other costs involved and they are really hard to calculate. Switching platforms is a risk. Switching from XP to say, 7 is a big enough risk with big enough costs and there's a high level of application compatibility there. Switching to ChromeOS? Lol. Even if the software and hardware was FREE, it would still cost money. A lot. A very difficult to calculate number. Business decision makers do not like large, difficult to calculate $ values for risk. With good reason: being able to budget effectively goes out the window.