I've noticed that many browsers default to starting a download as soon as a user clicks on a download link and before the user selects a destination. This has the advantage of speeding up the download effectively: the time the user browses for a destination can be used to begin the download.
The disadvantage is that if the download contains sensitive information and the intended destination is inside an encrypted device, the unencrypted data begins to be written to what is probably an insecure location.
Basically, I'm looking for a way to get the old behavior back: don't download a single byte until the final destination has been selected.
Edit: to clarify, I have tried the "Ask where to save each file before downloading" setting (in fact, I always use that setting). It does not achieve the goal. Observe your default download location when beginning a download, and you will see this:
This grew until it was 438MB (the latest Raspberry Pi OS Lite image). You can also observe the Chrome progress bar, etc. The file was written out completely as a
.tmp file before I hit "Save" in the dialog box.