It's primarily aimed at creative workflows. I mainly use it for creative coding, but I realize now it could also serve the same purpose for tracking changes made by AI agents.
https://github.com/zed-industries/zed/discussions/54150 failing to show newly created files and declining to provide a refresh button, instead adding a polling backend, breaks Zed on WSL.
Why a new version control system? Why not git, jj, or another existing system?
- Vertical activity bar (https://github.com/zed-industries/zed/pull/58868#issuecommen... - rejected because it "doesn’t match our design vision for Zed")
This is my #1 annoyance: clicking microscopic icons at the very bottom of the screen to switch views.
- Don't crash my entire PC when there's a large (100+ MB) JSON file in the working directory
- Don't freeze when resuming from suspension (I suspect this is GPU-driver related on linux) forcing me to `killall zed-editor`
Marketing aside, this is the kind of thing (advancements in VCSes) that I’d usually be excited about, but which I’m less excited about coming from Zed because it makes me start to worry that they’re going to continue focussing less on the core editor experience.
I love Zed and I hope that it doesn’t start to go flakey.
Some issues I've encountered recently at work:
- Routinely encounter bugs with the python language server, it's particularly slow when compared to using VSCode. All that snappiest disappears when finding symbols after the '.' is considerably slower than my typing speed.
- I found similar issues with C++ code bases.
- It's also so hard to debug when something goes wrong. I was trying to find out why pyright keep spinning on startup, but the "LSP logs" never show anything.
- Connecting via SSH always opens the repo in restricted mode, even though I check the "Trust all projects in" checkbox. Never happened before.
- Debugger UI is a mess. Make a run, stop the run, edit the task.json, try running again, and 2 tasks show up. One for the old and one for the new.
--
I stuck with it because I really do like the UI, the performance, and initially it seemed to be an improvement from VS Code for my workflows, but lately I've been moving back to code as it seems more stable. I don't really use agent features, so I don't think I'm the target customer for this or other priority Zed features.
I do see the usefulness of something like this, I'm interested to see if the new agentic style of programming will result in the "code" part of editors becoming second class (for the better) and will allow for more rapid development. Although I'm not looking forward to read my coworker's "threads" of how a feature was made, I have a feeling that it will be more difficult to read than the raw code, as from my experience, LLMs don't really employ terse and effective communication (might not be entirely fair, I found ChatGPT /study mode pretty good.)
--
Anyone know of a good replacement editor that might meet my usecase? I'm really a batteries included type of developer.
It’s true that they could focus on polishing the edit experience, but given that I rarely code anymore by hand, I have some doubts that it’s future proof to do that.
I think is not a bad bet trying to innovate in the LLM-coding space. Looking forward to try deltaDB and I hope it works outside of Zed !
I'd much prefer some more time spent building a solid editor which is why I switched to Zed early on in it's development cause I was frustrated and tired of VSCode. But saying that ever since Zed AI I've felt more and more frustrated with Zed and it's quality is starting to lack?
I can write a bunch of checkpoint commits and then squash them, I suppose, but the interface for doing it isn't super inviting and there are a lot of things I need to consider in terms of tracked and untracked files.
If I understand this correctly, I can deltaDB all my untracked and ignored files and git push my thought-out commits, fully respecting things ignored. And If I realise that I should have git-commited 10 minutes ago, I can rewind, commit, fast forward and continue.
Is it available now? Still seems to be a waitlist?
Now this DeltaDB goes way beyond that
simultaneously, i find it curious that the ide's git integration feels incomplete in front of vs code and its forks.
git is not perfect but for a lot of users, a better ux can help solve most of their grievances. regardless, network effect exists and it would take more than a killer app/site to get this going.
i would not outright dismiss this attempt, just like svn guys may have treated git back in the day.
A total nightmare and imo every dev should push back using this stuff. You'd otherwise have management brag about you not being a good communicator with the ai agent, the same management that have no idea how to steer these tools.
I really like Zed the editor, I'm not sure I like the path they getting into
Once I've tried one SCM to track every single change to my files, and the other for bulk commits. The result was quite satisfying.
It was some 15 years ago, we just switched from CVS to SVN and everyone thought it was the end of the history, at least for SCMs. Main branch was called master, and we were locking files by simply telling our teammates not to touch them.
At one time I decided I really missed the Eclipse feature which was tracking all local changes to the files, even without an active SCM. Making tiny commits was not an option, because I want it to track every file save, and we haven't yet figured out working with branches.
So I decided to install mercurial and made a hook which made a dummy commit every time I saved a file. Everything mercurial related was ignored by SVN.
Suddenly I was able to teleport myself to a version from before the lunch while SVN history was crystal clear.
Somehow I still miss integrated, per file, timeline in IDE.
Maybe I’m pessimistic, but I think this just gives micro-managers the data they need to micro-manage you. It invites people to scrutinize the process, instead of the results.
Not once have I wished to review a colleague’s conversation with an AI. I want to review finished work.
I look at this and I can only imagine layoffs being justified with “bad prompt quality”. An incident retrospective broadcasting your agent conversations to the entire company, with people speculating where you should have put the “make no mistakes”.
It’s a shame, really, because I like Zed as an editor, but I can’t bring myself to support them if this is the direction they’re choosing.
What actually is it and what value do I get out of it?
Reading the linked blog posts: It's a layer on top of git for CRDT-based collaborative editing that captures addressable code deltas + agent conversations. I could have real-time collaborative humans + agents development and can resume any agent conversation from any point in history, because it is all stored (powered by DeltaDB). I guess?
I actually might want this, but wouldn't know it from the landing page.
Personal use solo repo. I don't know id want people seeing my cc sessions lol.
Regarding DeltaDB’s request to 'log conversations and changes together', I think the Linux kernel's development model — which combines mailing lists, patch exchanges and Git — provides a good solution.
I suspect that, to secure funding or land enterprise contracts, they need to showcase a “showcase” project like DeltaDB, but I believe they also want to focus on the core parts of the editor.
--- I wrote the original text, but I had an LLM polish it, so there are em dashes in it
Jujutsu already does this. I had agent screwing up changeset over couple hour long session and it took one shell command to split on pieces with longer pause.
The only downside is that because of Jujutsu snapshots it can't be used inside sandbox (since all jj commands are mutating)
I am convinced that the conversation is the primary artifact produce while coding, which is not a common/popular opinion. Still, i love to see someone pushing forward on this.
(It's also why you never need to manually save a file in jetbrains. It just assumes you want to save, because you can always safely go back if needed. Editors with manual saving feel cumbersome now)
https://www.eng.auburn.edu/~kchang/comp6710/readings/They%20...
"There is the software. And then there are the databases beneath the software, two enormous databases, encyclopedic in their comprehensiveness. One is the history of the code itself -- with every line annotated, showing every time it was changed, why it was changed, when it was changed, what the purpose of the change was, what specifications documents detail the change. Everything that happens to the program is recorded in its master history. The genealogy of every line of code -- the reason it is the way it is -- is instantly available to everyone."
And the workflow they are describing in the post sounds amazing, I‘ve been wishing for that. It‘s orthogonal to CVS, no? You can still make PRs. But, crucially, you can loop people in early on, when it matters the most.
I see some companies do things that are widely out of their core product and engineers there just having fun with a bunch of different stuff with seemingly no commitment to bring in revenue. They'll re-invent/re-write things constantly, write a new compiler/package manager for their chosen language, create new frameworks, new open source projects unrelated to their companies, etc.
It must be awesome to be able to work in these conditions and still get paid. I'm not hating on it, it's how we get a lot of the good stuff we use.
Having said that, I love the idea of correlating the agent sessions to the changes. I'm glad someone is working on it.
If it’s targeted at working alongside git, and competing with the “local history” provided by like vscode, jetbrains and even vim where they track all saves to files it might serve a purpose?
The latter could be “useful” as an offline record. But the way they’re talking about sharing that, it feels like this is targeted at replacing git.
Most projects are not created under those conditions.
Completely ditched VS Code and never looked back.
Maybe some Ubuntu-specific issue?
I guess it depends? You could certainly start with another repo clone that's separate from the one you work on yourself?
As anyone who's ever tried to split a days worth of work apart with `git add -p` is well aware.
It may be feasible with a million agents, but they could remember what they typed three hours ago anyhow.
Revolution is in the air.
2. There is money in AI products, and a huge demand for solid AI workflows from many developers. Lack of AI tooling is absolutely a dealbreaker for many developers, whether you like it or not.
3. The core primitive here is just as much about enabling live collaboration which has been on Zed’s roadmap for many years before agents took over the industry.
Looks like omp supports ACP, so all you have to do is specify a Custom Agent in Zed and it should just work.
At least Excel has the courtesy to telling you "Hey the copy-paste buffer said it didn't catch that, can you try again?"
Linux Wayland (Ubuntu 26.04)
Zed is my daily driver and I'm pretty happy with it. I often switch over to other editors to do things Zed can't, but it's rarer than you'd think and really isn't that much of a hassle.
I've written my fair share of landing pages and copy lately. I start out with Claude for ideation and headlines it produces are almost always low effort word salads. Case in point:
"Software is made between commits"
It never fails to emphasize the last part, even if it bears no weight.
The only response to that headline is "Sure I guess".
It doesn't give you an idea of what problem it solves, doesn't sell you a vision or get you excited about the future. It's not even that insightful and is borderline factually incorrect.
1. Commits are really just manual checkpoints while you’re working. Why not automatically checkpoint everything so you can actually go back to an older implementation without needing to remember to commit it?
2. Agentic workflows are so deeply single player. It’s very annoying, I want to collaborate with my team with agents, not review the output of one dev’s agent which they’ll just shove back into their existing agent.
It’s crazy that is still an issue. I really tried to use Zed because I was trying to move away from VSCode and VSCodium did not have a featureful python lsp. I just couldn’t stand little things like that that. Also, the font was really blurry. VSCodium now has Pyfly which is decent, so I just switched back.
Have you considered using a keybind? They even show which keybind to use when you hover them.
Amazing that someone would prefer a vertical bar that is 85% empty, wasted space instead of pressing a few keys on the keyboard - for a text editor of all things.
I don't think you can put this on Zed. An unprivileged userspace application, even if coded poorly, shouldn't crash the machine.
I can't use an editor where the file contents I'm being shown don't match the reality of the file. I guess this issue doesn't affect most people somehow, but in the age of CLI agents I'm not sure how more people aren't running into it.
I was really enjoying Zed for a short while before I ran into this fundamental issue.
I have never used snippets so unsure about that too, going from command bar seemed to work.
I think it might be bad, but could also be it's not correctly configured maybe try asking codex/claude or something to check if it can fix/fine tune the config, if that doesn't work file issues?
I do suspect it might be working for me cause I got lucky I do see an open issue[1], but given Zed is used a bunch the lack of comments does suggest maybe it's not that wide spread either way not a good look.
[1]: https://github.com/zed-industries/zed/issues/47004
Maybe I got lucky?? Does it not auto update on linux maybe I am on an old build... Very suspicious.
For me, its a massive improvement over IntelliJ for my daily workflow.
Terrible Linux support.
If it's really so for some people, a case could be made that their problems can already be half-solved by changing the granularity of their commits: just commit more often.
VS Code can slow down as you add functionality, but Zed often just doesn’t have that functionality at all. So the proper comparison is “VS Code minus all extension-provided features” which is actually pretty fast.
Importantly, the group avoids blaming people for errors.
The process assumes blame - and it's the process that is
analyzed to discover why and how an error got through.
At the same time, accountability is a team concept:
no one person is ever solely responsible for writing or
inspecting code. "You don't get punished for making errors,"
says Marjorie Seiter, a senior member of the technical staff.
"If I make a mistake, and others reviewed my work, then I'm
not alone. I'm not being blamed for this."I still want to. I will give it another try when I am free.
But the whole environment has something that feels "invasion of privacy". Maybe its me.
I want an installation where typing latency must be invisible, literal invisible to the eye. Heard zed is great at that.
But I don't want my sensitive code to be put in a state that might lead it to exist in a cloud environment without my explicit permission.
Besides, the concept is good. Many a times I have wanted this exact feature - where what happens between commits can in principle be snapshoted with full checkout to any point - without of course any conscious effort from my side.
The neat thing is that they thing decide to make your life miserable based on their made up findings!
All my code lives on the other end of a Samba share. Guess what doesn't work reliably.
I fixed this and a couple related issues in a PR but it hasn't gotten any attention yet. I guess they are a bit swamped. https://github.com/zed-industries/zed/pull/59937
It's surely a lot of work but you get basically the discovery phase til market fit with your ass covered. Kudos to the engineers that pull that out, I really mean it!
This is how Zed leapfrogs the field and becomes a centicorn.
If you can train a model to operate at a higher level and with looser requirements, you've gotten rid of engineers in the loop and built a system for the automated synthesis of high quality software.
You're training systems to produce higher quality chain of thought across the entire software development lifecycle.
This is training data for the next stage.
The same logic also applies for software that is used internally. It's not always that developers want to rebuild stuff, but that they need to rebuild or alter software to fit their / company needs.
We have not even talked about cost or licensing or other issues.
I am a paying user, and have no plans to stop supporting them financially, but I wish they locked in and went through GH Issues queue, starting from the oldest.
you must not use it much. try pasting in an image, quitting zed and restarting (ie: upgrading the app). it shows you the base64 encoded data instead of just showing the image. heck, it doesn't even show you the image when you paste it in. it just shows the filename.
Re-invent is fine, that's how we actually get the good stuff we use.
Re-write (frequently) is usually a sign of bad power dynamics in the team, code rot and generally bad engineering management, in my experience.
That's why most React shops end up in this vicious cycle of constant re-writes of a perfectly fine working system for this reason. Yeah, must be really nice getting paid for it too.
I want "VScode without the bullshit". That's what I thought they were building.
(2) hence why I call them “distractions”. And yet Sublime Text hasn’t gotten greedy..
(3) Nobody cares about the collaborative features either.
You’ll pry Sublime Text from my cold dead hands. I tried Zed and it really was close but ever since they got greedy the editor has gone to crap. So for now I will continue to happily renew my Subl license every few years. For less than what it would cost for an ongoing subscription to Zed’s AI plans to “support them”.
I suppose companies could do this anyway just by installing spyware in your work computer, but I would very much prefer if that spyware was not my editor of choice.
It will be interesting to see how they handle balancing both an editor and a VCS without Big Tech resources like Microsoft's. Considering how many contributions and years it took for Git to reach its current level of maturity, taking on both projects seems like a massive undertaking.
It's 25 years old, and the patch files that make up the core of the decentralized collaboration system are fully 50 years old. Have the design requirements and constraints perhaps changed over the last 50 years?
Yeah I’m an avid user of the editor. I don’t care for their AI stuff I bring my own but I just want to see the editor become awesome. A more detailed object plugin ecosystem would be nice.
When do you expect them to become a ‘centicorn’ based on this fabulous technology?
It may seem weird nowadays but at the time it made sense since wasabi was "transpiling" VBScript (on which their existing codebase was written since they were all into Microsoft products at the time) to other languages in environments where VBScript didn't exist (e.g. IIRC they could transpile to PHP - and there isn't "one" PHP).
They were basically making shrink-wrap software running on an environment (the web) that wasn't made for shrink-wrap software (especially at the time) and came up with a technical solution to it.
FWIW they weren't the only to attempt something like that. HaXe is basically the same idea and was used to port a ton of games from Flash to all sorts of platforms and Monkey was another similar attempt (i think most people aren't aware of these two because they were gaming related and really limited to indie games).
The only thing that doesn't work well is inertial scrolling[0]. But scrolling is a fucking shit show on Linux no matter what... so I don't necessarily fault Zed for this. The same issue occurs in other apps.
Why?
I used Zed for about 6 months but I lost motivation.
I have the full jetbrains subscription and use it for some debugging etc. I got vim for quick edits and cli tooling for ai. I got no use case for Zed.
That's a bit of a low bar. Perhaps that's the disconnect?
For example, a lot of disable buttons do not actually disable stuff. They just hide it from the UI.
For any problem they just start chatting with LLM for 4 straight hours.
Can’t believe people hem and haw about paying for a license for them. $168 is nothing. That’s less than a half a day of salary for me and the amount of time saved using them easily adds up to multiples of that.
How would you prevent that? It can write until root disk is full, can take all your memory, saturate your CPU and a whole bunch of other stuff we (users) like to be able to do for the things that need them.
Unless you're suggesting something like Qubes or similar UX?
I'm also starting to wonder how much real, human-made copy is mimicking AI-created copy that has been filling marketing pages as of late. When people make marketing websites, they often riff and rip off other websites' (à la "Good artists copy; great artists steal"[0]) design ideas and language -- I wouldn't be surprised if there are humans now inadvertently adopting LLM's writing style like that.
[0] - https://en.wikiquote.org/wiki/Steve_Jobs#:~:text=Good%20arti...
Based on what exactly? I think that git is more useful than ever in this new "age of ai [slop]".
Thankfully, something being old doesn't automatically make it not useful anymore.
Be among the first to try DeltaDB, a version control system that records the work as it unfolds and keeps every change connected to the conversation that shaped it.
Rewind to any edit
DeltaDB captures every operation in between commits and gives each one a stable identity, so you can point to the code at any moment in its evolution.
Trace code to conversation
Every change is linked to the agent conversation that produced it. From any line of code, find the conversation. From any message, jump to the code it touched.
Branch at any moment
DeltaDB virtualizes the worktree, so spinning up a new agent branch is effectively free. Any point in history is a valid branch point, including mid-run.
Share the thread, not the PR
A teammate can join while the work is still happening, talk to the agent that did the work, and annotate as they go, without waiting for you to commit and push first.
Just need to see reality for how it is in the tech industry, products aren't made for users anymore they're mostly a ritual to please the VC gods. Bonus points if you're disruptive and a net-negative toward society.
It's like taxis and Uber. Yes it's probably bad long-term that local taxi companies have been replaced by a mega-corporation. But there were so many bad taxi companies with rude dispatchers, lazy drivers, cars that show up 30 minutes late, drivers that don't run the meter, artificial supply restriction in NYC (medallions), etc. If taxis didn't suck so much, they might have been able to fight back better, and got more sympathy from the public along the way.
You see similar claims leveled at certain unions like the ILA and police unions.
You just gave an example of 3 unprofitable projects.
That's pretty much my point: great that we got all of this stuff out of it but in any company not propped up by massive amounts of VC money you can't pay tens of millions of dollars in engineering salary for no immediate gain.
I know people disagree over definitions of "hard fork". Gram does not send patches to Zed but does occasionally merge fixes from Zed. There is not any specific effort to keep (non-AI) feature parity.
1. Git breaks down if files are too large or lines within files are too large.
2. Add/Rename conflicts will always burn you because git does not protect the intent behind changes.
And really the third is that IDE's are shit because there's no document object model (like the browser has). You have to make an error recovering parser to try to deal with the fact that an IDE is editing code half the time and text half the time. You're stuck playing a game you can't possibly win. Can you imagine if the HTML DOM kept blinking into and out of existence? It'd be a nightmare...
The vast majority of people don't like using poorly made things, especially when it is forced upon them against their will.
Eg Pi has its own tools for editing files and ignores the ones Zed sends over ACP. If I used Zeds AI pane to have Pi change a file, I can see the edits in the ACP window but the editor view doesn’t update.
I end up having to use something from the command palette to force refresh it, something like “Editor: Reload File”.
It happens to me across Windows, OSX, native ArchLinux and ArchLinux via WSL.
Granted, a bit of AI-influencing phrasing in human writing is common, but the idea that someone sat down, worked hard, and came back with an entire page of a pitch-perfect AI slop is simply incredible.
It shows the engineering effort that the team puts into it, which is why it is one of the few Electron based apps that is bearable to use.
I think Zed doesn't need to build this though.
Zed was a promising new direction I.e. VSCode minus the bullshit. But it's not getting the polish / productionisation it would need to actually be a better editor than the others. And now it doesn't look like it will.
it is a code editor after all. how about integrating the LSP so that it doesn't have to resort to rg/grep/sed/scripts to edit code?
In my opinion / experience, AI copy is a generally more cumbersome to read and process. So I would assume a company at Series B stage would have a human (with marketing guidelines) go over the language used on a one-pager to announce a new product.
But -- evidently -- anyone's guess!
<Accept> [Decline]
Also I think engineering team quality is bimodal, essentially a mixture of "good" teams "bad" teams. As with taxi drivers, only the "bad" ones really deserve to get disrupted.
My point yet again: only companies with money to burn can take that gamble.
I just checked my workstation and found Watchman alone holds ~556K watches across 34 roots. It keeps those watches even after the tool that spawned it disconnects. Notably, I've never invoked Watchman manually -- other tools do. And the 8k floor isn't much of a floor. My nixpkgs checkout alone is 37,686 directories, 4.6x that.
Truthfully, I had no idea Watchman stays resident and accumulates watched roots. But, that's kind of the point. It's a very easy thing to miss.
my favorite ux goof right now is the fastmode switch button right next to the stop button and how much space it all takes up.
Yes, it is, and for good reasons. That's why I disagreed with the original "An unprivileged userspace application, even if coded poorly, shouldn't crash the machine" part.