I have worked in monorepos for large codebases. In those, file trees are completely useless because you can't fit the repo layout to your mind. Also, the file trees are so large that when I'd ask a coworker using VSCode to open a file, they'd usually take a few seconds trying to locate it (on remote ssh, it even takes a bit to load files to expand directory as well).

For smaller code bases, I usually read the repo first and I'm able to learn what exists while working through it.