You’re awfully curious, aren’t you?
Formerly of lemmy.sdf.org
That is a damn nice paper, thanks for sharing that! The comparison table is, if a little wacky-looking at first glance, a pretty great overview. I skimmed it for the abstract and conclusion but now I think it’s worth reading it in full.
Kinda depends on exactly how much you want it to sync, e.g. should it truly sync all files from the synced volume to all clients, or do you want the files to only sync/download to clients when requested, keeping a “master copy” on a server?
For me, I use Syncthing for a simple setup where I want multiple machines to have a few different shared folders that sync with each other, all files updated on-disk when connected to a network.
For a more Google Drive-style approach with only fetching/syncing files onto client disk when explicitly chosen, I believe Nextcloud has what you need, but I’ve not used it on any of my own machines. It can be self-hosted or paid for, depending on what you prefer.
I kinda agree, but I also think it’d be pushing it to have more than 5 categories/levels, and I feel like this gets close. Anything more than 5 would probably come off as too complicated for most people, imo.
I’d like there to be one between AI:R and AI:W to cover the more-involved end of the “AI assistance” spectrum, e.g. using it for general queries about the codebase vs. relying on it for making granular design decisions even when not having it generate code. However, I don’t know if that difference in use can be defined clearly and rigidly enough to justify a second level of AI:R or something.
Totally agree though, the AI:R and kinda the AI:W one feel a bit too broad/lacking in some useful specificity.