A quick fix might be to disable any ipv6 addresses if you don’t specifically need them. The vpn /could/ be ipv4 only, which /could/ leave your ipv6 free to leak or make ipv6 dns requests.

Assuming that you trust what Proton says, when they receive a (possibly unencrypted) message they re-encrypt it with your key as soon as possible and they don’t log the content. So, after that point, they (or anyone else) can’t read the email contents. If it was also encrypted in transit, then there’s only a small window inside their email processing system where the plaintext was passed from one encryption to the other. It’s only decrypted again in your browser or proton mail app with the key that only you have. It’s not bulletproof, but it’s better than most providers.

Agreed. Theoretically possible, but practically not possible unless you are an embedded hardware engineer with access to Sony’s datasheets and potentially crypto keys. Some sort of external box is much more practical.

assfish are soft and flabby

Your domain name could be ordered to be removed from US-based dns providers, no matter which TLD it is. That would essentially block your website from most US-based viewers without actually shutting down your hosting. Advanced users could still get to it, though. Consider hosting through Tor and a .onion address for more resiliency.

Look into https://simplelogin.io/

They make creating random aliases for custom domains like this easy.

As for the domain name itself, anything that already looks like a mail service is good. “examplemail.com” or “mailexample.com”

Tailscale (https://tailscale.com/) works great for remote access to your private services. Once the wireguard tunnel is established, then the traffic is peer-to-peer (assuming it’s configured correctly) and not through their centralized servers. Even from a mobile device.

You might enjoy reading Extreme Privacy by Michael Bazzell

https://www.goodreads.com/book/show/217289412

How much money did you spend, and how much time overall? Would you do it again?

Agreed. Windows updates will very likely break your single-drive dual-boot at some point. So, use two different drives and use your bios/efi to choose which one to boot.

Edit: check out https://hackaday.com/2021/11/30/linux-fu-the-ultimate-dual-boot-laptop/

Edit 2: Framework 16 looks like it would meet your needs. It has two M.2 sockets for drives. https://frame.work/products/laptop16-diy-amd-7040

The vendor will absolutely take that custom code and use it to extract maximum profit from a different customer. I’ve experienced it from both sides of the transaction. Open source at least allows the functionality to be “developed” only once.

It doesn’t even have to be offshore accounts. Just a fat long-term maintenance contract would be enough to hide a lot of corrupt costs.

Maybe instead of choosing an app, you could consider the overall process of tracking what you need to do. Read David Allen’s book “Getting Things Done”. It’s short. Then you can apply that process using your favorite app.

https://en.m.wikipedia.org/wiki/Getting_Things_Done

My favorite app for GTD is emacs and org-mode.

U.S. Immigration and Customs Enforcement www.ice.gov

800TB of bandwidth per month?

Check out https://www.firefly-iii.org/

If you’re not already, use it as your main system. Don’t dual boot. Stop using windows and mac. When you run into something you need to do, figure out how to do it on linux. It will be slow going at first, but after a few months you will pick up more productivity than you had before.

Another commenter recommended the fish shell, but I disagree because fish is not posix compliant. Almost all of the shell script examples that you will find assume posix compatibility and will usually have to be modified to run with fish. Once you get comfortable with a posix-compliant shell, then maybe consider fish or another “modern” shell.

On the topic of shells, read the bash manual. It’s long and informative. You don’t have to memorize it, but be aware of the different concepts there, and refer to it when you need to. It’s pretty horrible as a programming language, but it’s what glues most of Linux together.

First, make sure it’s enabled in your kernel. Check the value in /proc/sys/kernel/sysrq to see if it’s enabled. Then see if you can trigger it by writing to /proc/sysrq-trigger. Then try an external keyboard that has real SysRq key. If all of those work, you may have to ask Framework support if their keyboard supports generating that keystroke.

The “Magic SysRq key” may be helpful as a last resort.

https://en.m.wikipedia.org/wiki/Magic_SysRq_key