jj is in the training data for at least Gemini these days. My experience is it uses less tokens to do similar workflows with jj vs git because they are simpler.
Also, remember that humans aren't going away. The touch points are changing but they still exist. The VCS is a common touch point for many and my preferred one. For others I'm sure it's online code review tools. If an agent is operating under my name in my workspace I want to manage it within the confines of my workspace. Once it becomes more agentic and operates as it's own entity I'm sure I'll change my mind but that's not the case today.
If you have ever TL'd a team, it doesn't sound too crazy. I have 8 folks I generally talk to very consistently throughout the day. If I'm not in 1:1s with them I'm usually reviewing their changes or chatting with them over chat. I don't think I can do all of that and work with a bunch of AI windows, but I do think they could likely do something similar to me with several agents running in parallel.
The problem is that trust shouldn't be so binary. We should have ways to increase trust without needing to resort to the standard library. There was an effort to do this at some point in rust but the idea was sadly not well received. Maybe it'll end up reviving itself with modern supply chain concerns.
The idea is that there could form some groups of well maintained crates that only depend on each other and have a similar amount of oversight. This actually naturally happens in c++ because grabbing dependencies is so painful, but it makes dependencies more trustworthy. For instance boost, absl, folly, etc.
A lot of humans don't currently trust agents to touch VCS today. I also find that my agent tends to be much better about dealing with jj than it is with git.
I think it comes down to your ability to plan and understand how the work can be broken down before you try solving it. I often know what every commit will look like before I ever touch the code. I do sometimes learn things and change my mind as I make changes but it doesn't often change my commit structure. I tend to work on a codebase I have 8+ years experience in though. I'm sure it doesn't work in a variety of situations though.
The difference in its ability is immense. Even with less features it makes a lot of sense to switch. It really shows how much the harness matters almost equally to the model.
There is little that they can do short of running the programs in a VM. Linux distros aren't engineered to consider applications as something different from the user running them. You need a completely different security model to achieve that and the Python runtime isn't tackle that.
Bold claim. I have the opposite experience. On this site I imagine most folks will agree with you, but there are a lot of folks who choose to work at larger companies over small ones.
Notably the rewrite managed to also both reduce binary size and reduce compile time. Given it's largely a straight non idiomatic migration, that's surprising to me.
I use AI prolifically for coding, but I haven't crossed the chasm of letting it do mutable operations on my vcs. It tends to revert changes too eagerly. Commit messages written by it usually miss the point of the commit and end up too verbose and need to be modified or completely rewritten in many cases. I feel like prose needs to capture style which is unique per individual which makes it difficult to do well.
If you like a thing, it feels natural to tell others about it so that they can also benefit from it. There is no ulterior motive here. Sure everyone in the community will benefit if jj becomes more popular, but it's not like they are trying to sell you something for money. If you don't like it, that's fine, but if someone is facing similar problems to the OP then they will benefit. Almost everyone I've shown jj to at work has converted because it solves the same problems for them as it did for myself. Of course, not everyone has those problems.
I work primarily in c++ and fully agree with the author. I think it makes more sense in the context of larger teams, possibly also monorepos. Review speed (both latency of feedback and how thorough your reviewers are), presubmit test runtime, flake rate, etc can lead to it taking a few days to get work submitted. In some cases you don't want to land work until you've finished a milestone of some sort. If others work on similar files to you, you will end up with several rounds of rebase conflicts to deal with too.
If most of the glaring problems are addressed (massive unsafe usage), and metrics show improvement (less crashes), then did it really go wrong? The fact the code is not idiomatic is less interesting, because that can be addressed incrementally. Let's wait 3 months and reflect.
There is enough flow of engineers between the two companies that many of the things Facebook has done better have been copied by Google. I think in many ways it goes both ways.
There is in fact a version of cider that does work for folks working outside of google3, such as android. It's not as seamless, but it has been getting better and tries to recreate the google3 experience.
The internal version is just slightly optimized for Google's unique developer environment (auth + cloud vm + google3) and is a dogfood version, meaning we get unstable features first. It's mostly the same thing though.