Linux

I've been running Debian Testing for decades, but I've recently switched back to Stable on my homelab server and started using Arch on my laptop.

Testing is not as stable as you want. It does occasionally suffer from problems that have to be fixed manually, just like in Unstable. In reality my system was mostly running Testing, with some manually pulled packages from Unstable or even manually pinned on Stable. This was sometimes necessary to solve specific problems, particularly security fixes that weren't available in Testing yet. I followed all the security mailing lists and acted upon them manually to stay somewhat safe, which is a decent chunk of extra work.

In recent years it has become less and less important what software versions your distro ships with, since you can easily run software from Flatpak, homebrew, or even directly compiled Go or Rust software. Tools exist to keep all of that up-to-date automatically. In practice that solves all my cases where I wanted to run a different version than Stable offered, so it's more secure that way.

If you want to run a different version of something as foundational as Gnome you've basically picked the wrong distro. It would be such a PITA to change that manually, with tons of dependencies on other stuff. Switching to a rolling distro would be a far more logical step, which is why I run Arch on my laptop.

In short: don't use Testing. It wasn't meant for what you're trying to accomplish and you're going to have a bad time (security + more manual intervention needed than you want). Either accept older software and the peace of mind that brings (running Stable is REALLY easy, everything is sorted automatically) or go for something completely different, like a rolling distro.

Personally I'm pretty wary of the new wave of distros basing themselves on Debian Testing. I have a feeling that they don't realise what they're getting themselves in to, and that they can't offer the same level of support as Debian does for their Stable branch.