I haven't come across a single program that tells you what is blocking the ejection even in Linux. Sure, if you use the right tools you can find the culprit, but the program that does the ejection task seems to love being vague.
It works with containers so I can create a setup where requests sent from the container goes through the VPN. I use it for my Redlib setup to bypass rate limiting by rotating its IP regularly. Unless you have your host to route all traffic through a certain node, it should work independently from Tailscale.
Yes, it makes them behave just like normal files. Internet connection is necessary unfortunately, and you can't do the "edit now, upload later" kind of thing.
You can mount your OneDrive on your filesystem with rclone. Once mounted, it works just like a normal directory, except with some latency. You can mount it to multiple devices too.
If you don't like sharing your files to Microsoft in plaintext, you can also create encrypted remote within Microsoft remote. With both set up, files will appear in plaintext on your desktop's filesystem, but will be encrypted on OneDrive.
They surprisingly were. Originally I went with Rust because writing them in bash script would be very inefficient in terms of the number of programs I need to invoke to fulfil the function. Then I ended up learning some systemd basics.
I haven't seen anyone using parasols/umbrellas when it's too sunny in UK, but it's pretty common in Korea. I don't think I've seen them in Europe in general either. No idea for anywhere else to be honest.
Oh that makes more sense. I've been to a couple of datacenters, though none of them are purely dedicated to AI workload itself. One thing I found was that there seems to be a mechanism that releases some sort of gas that blocks out oxygen in case of fire. At least that was what I gathered from the warnings around the whole facility. That could be mandated.
