Thank you for your response! You are correct in saying that GCSF doesn't download any data upfront. It constructs the file tree at mount time using only file metadata and downloads the actual file content only when it encounters a `read` call. This is an advantage if you're running low on local space (the file system essentially adds 15 GB of "free" additional storage).

Real time collaboration is indeed a shortcoming. I would still use the online interface of docs/sheets/slides for this purpose.

I haven't personally used Backup and Sync, as there is no Linux version of it. From what I gather, it seems that it uploads local files to a new category on Drive instead of the 'My Drive' directory. This can be useful for automatic backup. You simply set it up once and forget about it.

However, GCSF might be a better choice for the additional control it provides. Whereas with Backup and Sync you have to inspect a file manually in order to check whether it was synced or not, GCSF ensures that a pending write operation will only return once the file transfer is effectively complete. For instance, when copying a file to Drive, the execution of the command will take as long as the upload process itself. Once finished, an exit status of 0 will indicate precisely that the upload was successful and the file is certainly on Drive.

I imagine this sort of strategy is a better fit for use cases which require high confidence and predictable behavior.