Once running, it's just a case of copying or editing files in a folder and they're everywhere else. No manual intervention, no trying to remember which computer has the latest copy, it doesn't get upset if I throw terabytes of media at it, it doesn't get upset if I `git init` inside multiple folders and suddenly I have my coding projects ready to go on whatever laptop I pick up.
It also does some nice additional bits, like making it possible to have machines that store a copy of the data but never have decryption keys, so I can use a cheap VPS without worrying about it having a plain-text copy of my data. There's also the ability to store multiple revisions of changed files, should you want it. It's also, at the end of the day, actual plain files on your filesystem which can save a lot of headaches.
It also does a pretty good job of finding the best path to a different box. It shares IPs (that you run the service on), so same lan, different vlan, via a VPN, via a double nat'd someone else's wifi; it doesn't care and usually does the right thing (and is configurable if you wish/need).
The comparison isn't rsync. SyncThing is a host-your-own dropbox/onedrive/other alternative without a requirement for a provider, accounts, fees or your data belonging to someone else, and stunningly low barriers to entry/usage by non technical people.
So, don't think I need this but will keep it in mind for NAT traversal or something like that.
> without a requirement for a provider, accounts, fees or your data belonging to someone else
Sounds like it is not completely true if you need a server in the cloud as a prerequisite, though maybe I misunderstood.
Rsync is great as there is a clear master copy. But if changes may happen at any copy of the data, it can become a chore.
Also SyncThing is bidirectional whereas rsync is one way (obvs you could also push both ways).