Here are some:

• Work on multiple changes at once, easily. No need for multiple checkouts, workspaces, constant rebasing, etc.
• No need to make or name branches. Branches are just in the eye of the beholder.
• Effortlessly rearrange, split, and squash your changes without ever abandoning a rebase -i halfway through because you’ve lost track of what’s happening.
• Deal with conflicts whenever you feel like it, not right in the middle of your train of thought.
• The oplog lets you globally undo and redo by just doing undo and redo, no need to decode a reflog. It’s version management for your version management.
• Totally interoperable with a shared git repo. No one else has to adopt it for you to get the benefits.
• gg: Makes things that are a slog in git ridiculously simple. You have commits A, B, C and wish it was C, B, A or even C, B+A? You made a change and now wish you had broken it out to a different branch? Just drag and drop.

Random testimonial, but I was in the same boat as you. I used stacked branches a lot (thus I rebased often), and Git mostly just worked for me. I still tried jj out of curiosity a few months ago. I haven’t used git since then.

There’s something about jj that just clicks for me. It seems my mental model matches closely the API offered by jj. I can’t name one specific killer feature, or one thing you can’t do in git and only in jj; it’s just that there is never any friction in whatever operation I’m doing. Everything seems effortless.

Yes. Read a tutorial or two, and give it an actual serious try for a day or two. That should be enough.

jj has a builtin diff editor that does some of what magit does. It’s still very much a prototype though:

https://www.pauladamsmith.com/blog/2025/01/cheatsheet-for-jjs-builtin-diff-editor.html