As seen on GitHub's "Get up to speed with partial clone and shallow clone" page, you can do a clone of a project in such a way where not all "blobs" are downloaded upfront, but instead on-demand, creating a much smaller/faster initial download. From their description, the variant I'd usually want is probably:
git clone --filter=blob:none <url>
They call this a "blobless clone" and recommend it for if your use "spans multiple builds", which is usually the case for me.
If "git stash" is failing with an error message mentioning a file being "not uptodate", a possible workaround is staging/adding all files and then stashing.
git checkout HEAD -- myfile.txt
(as described https://stackoverflow.com/questions/7147270/hard-reset-of-a-single-file)
git reset -- myfile.txt
Do NOT use git rm, as that actually stages the DELETION of the file!
(https://stackoverflow.com/questions/6919121/why-there-are-two-ways-to-unstage-a-file-in-git)
Let's say you have some notes or something that you don't want to add to the git repo but that you don't want to "pollute" the .gitignore by mentioning, since nobody else will probably need to worry about ignoring files with that name/path/pattern.
What you can then do is to edit <your project dir>/.git/info/exclude file ("exclude" is a text file without an extension, not a directory name) with the names of the files that you want to exclude!
This comes from https://docs.github.com/en/get-started/getting-started-with-git/ignoring-files. The link also mentions a global computer-wide option, but that sounds more error-prone and more likely to be unnecessary.
https://stackoverflow.com/questions/2364147/how-to-get-just-one-file-from-another-branch
git switch dest_branch
git restore --source source_branch -- thing.txt