There are solutions besides flakes to pin versions of packages. niv and npins are notable. I like them better as they are quite stable and predictable.
Using flakes pull in a whole bunch of stuff that I'm frankly uncomfortable dealing with. Flake parts, flake utils,... It's like just getting comfortable with Haskell basics and then discovering type theory. Ain't nobody got time for that.
The thunderboltstream driver exposes /dev/tbstreamX devices on each host of a directly-connected USB4/Thunderbolt cable. From there data can be transferred using regular file-system operations, e.g. you can dd or cat from one system to another or similar commands.
This I like! Back when I started transferring stuff between computers, this was exactly what I expected to hallen when connecting two computers via USB. It was disappointing to find out only certain USB cables supported it. Maybe there were even kernel modules for it, but USB4stream will be so much easier.
Radicle doesn't have a 1-click solution to TOR and I2P, but there is a guide. Granted, radicle could do with a GUI configuration tool, but I think a fork and C rewrite are unnecessary to achieve that. A script around it could probably achieve the same. On nixos, it would be a simple matter of writing and sharing a configuration.
I watch stuff at 1080p but 1.5x-2x. A 1h video at that with 6kbps and 60fps can be quite consuming especially if a lot is happening. Have that on in the background for 8h plus doing other stuff (gotta pull docker images, or the blackhole that is npm) and you get over 10GB daily easily. Add a modern game or two a month and you're above 1TiB/month.
Blame the short term focus on profit and the system we exist in. There simply is no reason for it to be stable. It's not a life or death matter, neither for people nor for companies, hence, it will not be prioritised. The world would have to be very different for these "basic" things to be prioritised.
That is actually ducking hilarious. It had so many false positives that cops ignored it.