One of the most common answers when asked about "How can I achieve more privacy and security with my OS?" is to switch to FOSS/Linux.
The biggest issues I have is that, when you search for privacy oriented Linux distros - they almost always boil down to the following: Qubes, Kodachi, Kali, Whonix, Tails, BlackArch, and variants.
My issue is that these are not very user-friendly or user-oriented distros. They are more or less distros for either security experts, the very paranoid, or penetration testers.
My thoughts on a private OS isn't necessarily one that's safe from a direct attack or something you need to secure spy-level intelligence, but one that does not leak private information, does good-enough encryption by default and offers great usability while allowing access to a good set of applications that are also privacy-oriented. Applications and the OS should not leak telemetry data by default (opt-in only) or have no private data leaked at all.
Obviously, a certain amount of data will inevitably be shared (DNS lookups, Google searches) but that kind of information should be communicated to the end-user as much as possible.
I'd love to see more discussion around this, because I think that this is a topic that will become ever so critical with the increase in "telemetry data" being collected.
Your browser alone is probably 10 fold the threat to your privacy the operating system is, and a share of the telemetry is legitimate and very much directly useful for improving the product and responding to incidents.
Any linux distro probably won't hash/index all your files into a voice assistant, won't virus scan, and won't backup to a cloud, which is an immediate privacy and security win.
If you want to engage in the privacy mindset, you must first focus on observability. Specifically you should be thinking about how to set up a proxy/firewall between your daily driver and the raw internet to see what is accessed.
Without observability, you don't know what is doing what no matter what you use. You are forced to trust without understanding, and you wouldn't be able to answer your own question empirically. Without an off machine firewall, you can't prevent the behavior you want to prevent.
For a long time OpenBSD had the reputation for being most secure by default. I don't know if that's still true. Running some type of observability/firewall platform on that would probably be useful. I've never used mitmproxy, but I imagine that would be useful.