what if that setting were an enum every operator started with AiMode::Off so we can consciously opt-in, plus then Zed could have more AI Mode levels besides just on or off?
Maybe there could be a level for experienced programmers and another level for vibe coders, and prompts / permissions could be set accordingly?
What I'd really like is an option to disable the automatic multi-buffer behavior for things like git diffs. It just doesn't work for me, and the Zed UX seems to not even realize that not everyone will love it.
I am really, really starting to like Zed. I noticed recently that it now respects my EditorConfig files, which was one of my biggest gripes before recently. In general almost everything that I didn't love about Zed has been fixed.
I have only tried it a couple of times, but Zed's AI assistant view indeed works as promised, doing about as well as Cursor does.
Now there's just one thing that is a bit annoying... It's kind of heavy on the GPU. I mean it looks and runs great, but it really loves to suck the battery out of a device that's running on battery. But frankly, I've been ready to ditch text editors based on web browsers basically since they became "in vogue" and have just been waiting for the right thing to show up. As time goes on, I think it becomes more and more obvious that Zed is "the right thing".
I don't know if there's anything we can do to make sure Zed succeeds sustainably and doesn't eventually get enshittified, but I'm crossing my fingers that for once we can have nice things.
This is why I still use VS Code with AI features off and none of the AI integrated IDEs. It's not that I don't use AI. It's just that having the AI de-coupled from editor makes it much easier to separate concerns. Some days I don't feel like using AI and just need to edit one line. Other days I want to do a major sprint and test the latest AI and see if it will accomplish the task.
I mean all good but I still wonder if building an editor is really sustainable you basically only fight for new coders everybody else will just stay with vs or IntelliJ because of muscle memory
I know you can share keymaps but it never works 100%
Like I must’ve been tremendously more disappointed with my current ide to even consider switching
Sadly, I came back to using VS Code recently. There's a lot to like in Zed but imo that decision to write their own rendering framework is unfortunate, because of ridiculous problems in Linux still not resolved like poor font rendering, especially on low-DPI screens, or visible lags of UI which is being developed to be blazingly fast. So far, VS Code is faster for me.
I tried Zed many times. In fact I try to use it at least once a month because it is indeed fast. However, I just think it’s a half-baked product and I don’t want to invest too much time into it. Wish they didn’t get distracted by AI and focused on making a viable VSCode replacement as an IDE/Editor.
I went from VS Code to Cursor, then got frustrated with Cursor breaking keybindings and other things, tried to go back to VS Code but missed the superior tab completion. Then I gave Zed a long hard try, but after over a month of daily usage I went back to Cursor again, just for the tab completion quality.
I don't use any of the chat or agent features, but for me Cursor's tab completion is a step forward in work efficiency that Zeta and Copilot were not. Sometimes it's subtle, and sometimes it is very obvious. Cursor seems to have sources of context that the others don't, like file names from the directory tree, and maybe even the relevant .pyi type annotations and docs for python modules. It also jumps to the next relevant problem site very effectively. It feels like the Cursor devs have done a ton of practical work that will be hard to match with anything other than a full-on competitive effort.
I want to see Zed succeed. I think it's very important that VS Code and its ultra-funded derivatives not dominate the modern editor landscape too thoroughly. Tab completion used to seem like a straightforward thing, but if the state of the art requires a very elaborate, whole-workspace-as-context environment to operate in, then I wonder if it's going to become a go big or go home kind of feature.
I can't help wonder what the actual internal API for this kind of thing is going to look like in the future. It used to be something like, what's the current token behind the cursor, and look in a big prefix tree of indexed words. Then maybe it got more elaborate with things like tree-sitter, like what's the incomplete parse tree up to this point. Then when editors started using AI, I stopped having any idea of what the actual inputs are. I'd love to hear about real implementation experience at any stage of this evolution.
Zed seems really nice, and the usability has come a really long way in the last couple of months. That being said, I have one issue, and I hesitate to even bring it up because it seems so shallow: I hate all the themes! They all look so ugly and amateur. I know this is a really trivial thing, it's like complaining I don't like Python because it uses whitespace for indent or something, but I just can't get over it. VSCode/Cursor is beautiful in comparison.
I recently found Github Dark Default, which is probably the okay-est of the 10-15 I tried, but there is so much that still looks bad. The autocomplete popover looks far worse than VSCode, the file tree looks much worse, tabs look ugly, etc.
My problem with AI in Zed is not that it's there, but that it feels like it's always behind in AI code editor paradigms. They were pretty late to the party to add edit predictions, and their agent UX is pretty behind the game. Recently, Cursor added background agents which I feel is a game changer and I now feel it's a deal breaker when choosing an editor.
It makes me wonder if choosing to build their own GUI framework in Rust was the right move. Zed is a great code editor, but for me, it's not a great AI code editor.
I did 4 months of Cursor. Figured I’d try 4 months of Zed, and see from there. I’m about 3 months in and so far might stick with it. I’ve generally been pleased with it. My only issues really have been
a) all the settings are just in a file, not even a commented skeleton, and with a decent change rate, hard to invest in learning them; I’d love something like the jetbrains settings dialog
b) the stupid placement of the burn mode button; it’s so easy to accidentally turn this on and not realize (done it twice now). Would love an option to just remove said button
c) I wish there was more panel control; I’d like to place outline and project views BOTH in a panel, but you only get two full height tools in the left/right panels.
Zed has everything I want out of an editor, the AI stuff wasn't really a concern to me since its always been optional, but I fully appreciate where they're coming from. I think Zed is one of the good gems of our time. I have been looking for a decent spiritual successor to Sublime, and Zed to me is really good at it.
This gives me hope they might fix collaboration too. Unfortunately had to switch to VSCode Liveshare (which has its own problems) a few times because Zeds collaboration was so broken.
I moved to Zed last week and its been amazing so far! Something about using it feels a lot more crisp and fast compared to vscode. Their AI integration is pretty decent too. But I am very glad I can disable AI now. I find AI sometimes takes me out of the flow because I am constantly reviewing its code.
I appreciate this option since I use Zed without AI. However, their overwhelming focus on cramming AI in the editor just disappoints me. The core experience is still very raw and there are small things that build over to make it annoying.
I still have Sublime Text as the backup editor. It's lack of a powerful sidebar and search are the only things that stopped me from using it regularly.
I do use VSCode at times and the CoPilor powered tab completion (which mostly hallucinates and spits out nonsense) is just obnoxious. I've found JetBrains' implementation of full line completion and block completions a lot more thoughtful and reliable.
53 comments
[ 3.6 ms ] story [ 49.3 ms ] threadBut the reality is like for JavaScript: it's required by default :D.
Good that Zed fixes this!
what if that setting were an enum every operator started with AiMode::Off so we can consciously opt-in, plus then Zed could have more AI Mode levels besides just on or off?
Maybe there could be a level for experienced programmers and another level for vibe coders, and prompts / permissions could be set accordingly?
I have only tried it a couple of times, but Zed's AI assistant view indeed works as promised, doing about as well as Cursor does.
Now there's just one thing that is a bit annoying... It's kind of heavy on the GPU. I mean it looks and runs great, but it really loves to suck the battery out of a device that's running on battery. But frankly, I've been ready to ditch text editors based on web browsers basically since they became "in vogue" and have just been waiting for the right thing to show up. As time goes on, I think it becomes more and more obvious that Zed is "the right thing".
I don't know if there's anything we can do to make sure Zed succeeds sustainably and doesn't eventually get enshittified, but I'm crossing my fingers that for once we can have nice things.
I know you can share keymaps but it never works 100%
Like I must’ve been tremendously more disappointed with my current ide to even consider switching
I don't use any of the chat or agent features, but for me Cursor's tab completion is a step forward in work efficiency that Zeta and Copilot were not. Sometimes it's subtle, and sometimes it is very obvious. Cursor seems to have sources of context that the others don't, like file names from the directory tree, and maybe even the relevant .pyi type annotations and docs for python modules. It also jumps to the next relevant problem site very effectively. It feels like the Cursor devs have done a ton of practical work that will be hard to match with anything other than a full-on competitive effort.
I want to see Zed succeed. I think it's very important that VS Code and its ultra-funded derivatives not dominate the modern editor landscape too thoroughly. Tab completion used to seem like a straightforward thing, but if the state of the art requires a very elaborate, whole-workspace-as-context environment to operate in, then I wonder if it's going to become a go big or go home kind of feature.
I can't help wonder what the actual internal API for this kind of thing is going to look like in the future. It used to be something like, what's the current token behind the cursor, and look in a big prefix tree of indexed words. Then maybe it got more elaborate with things like tree-sitter, like what's the incomplete parse tree up to this point. Then when editors started using AI, I stopped having any idea of what the actual inputs are. I'd love to hear about real implementation experience at any stage of this evolution.
I recently found Github Dark Default, which is probably the okay-est of the 10-15 I tried, but there is so much that still looks bad. The autocomplete popover looks far worse than VSCode, the file tree looks much worse, tabs look ugly, etc.
Does anyone have any suggestions here?
a) all the settings are just in a file, not even a commented skeleton, and with a decent change rate, hard to invest in learning them; I’d love something like the jetbrains settings dialog
b) the stupid placement of the burn mode button; it’s so easy to accidentally turn this on and not realize (done it twice now). Would love an option to just remove said button
c) I wish there was more panel control; I’d like to place outline and project views BOTH in a panel, but you only get two full height tools in the left/right panels.
That creates a system with a potential double negative (disable is false) which is unclear.
It's cleaner to do:
{ "enable_ai": false }
I still have Sublime Text as the backup editor. It's lack of a powerful sidebar and search are the only things that stopped me from using it regularly.
I do use VSCode at times and the CoPilor powered tab completion (which mostly hallucinates and spits out nonsense) is just obnoxious. I've found JetBrains' implementation of full line completion and block completions a lot more thoughtful and reliable.