stuner

No, they're almost entirely unrelated. Almost all CPUs will idle close to 0 W (with correctly working drivers). The main idle power contribution comes from the mainboard and other devices (e.g. disks). The Mini PCs you mentioned should have a very low total idle power, probably below 10W.

Check out Wolfgang on YouTube, he has some great videos on the topic: https://youtu.be/Ppo6C_JhDHM

I've also recently built my own NAS and I've gone through similar considerations. One of my mayor decisions was not to use btrfs because it's not recommended for Raid Z1/Raid 5. With that, I landed on ZFS and TrueNAS Scale. Note that RAID expansion should be landing in both very soon.

Things with TrueNAS were pretty easy, very quick, and everything worked nicely. However, I noticed that it was constantly accessing the disks and preventing them from spinning down. I really wanted to keep the power consumption low (<20 W idle), so I eventually decided to just go with Vanilla Debian + ZFS. I can recommend that if you want to tinker with things yourself. Otherwise, I'd recommend TrueNAS Scale.

As for migration, you might be able to create a degraded pool initially, copy over the data, and add the parity disk last. Raid expansion would ofc also help there...

What kind of junk energy is there to harvest from a car (in meaningful amounts)? I guess breaking is the obvious answer, but that's already covered by regenerative breaking. Most car-based energy harvesting systems seem to employ speedbumps that clearly take useful (kinetic) energy away from the car (probably at a very poor efficiency).

How would a turbine that takes energy from the air current generated by a passing car decrease the energy of the car?

Not sure where you got that idea from, but how would that generate a meaningful amount of energy? It seems very unlikely that such a system would ever recover the energy spent on its construction.

Sure, but those are completely different approaches. Dams have the advantage that they have a much larger capture area for water and that they can accelerate the water beyond the 10 m/s terminal velocity of raindrops.

Raindrop energy harvesting is a rubbish idea. The raindrops simply don't have a meaningful amount of energy to begin with: https://news.ycombinator.com/item?id=36907674

One way to do it is have a small Python (or any other scripting language really) script that performs text replacements in the Latex source file. This is much easier in Latex because it's plain text. I don't know of a solution that doesn't involve writing your own code (apart from LO/Word serial letters).

Is using Latex an option? I've done that and it works quite nicely. You can easily populate a template e.g. using Python.

The article says they are aiming for 1W in the next couple of years, which can probably do it.

They won't magically improve the power density by three orders of magnitude. They're just trying to defraud their investors.