I made a vim extension where you describe the edit/action you want in natural language, and my ollama model thats trained on books like Practical Vim returns the key sequence and you can press e to execute without leaving vim. So you get automation help but also learn the syntax.
I used mcp-neovim-server to let my LLM control my vim session. that way it can debug and test and poke around. It's crazy good at debugging plugins. It's absurd how little these things need to look-up docs; half these models are just out the box wildly good at vim. "Open these files on these lines" and now there are four splits with me at the relevant line numbers. Awesome. And it'll explain how to do things & test it out & validate! https://github.com/bigcodegen/mcp-neovim-server
That was a little tricky to set-up. I ended up writing nvim-auto-listen, which uses some heuristics to find your project root, and starts a .nvim.socket in that directory. That makes it easy for mcp-neovim-server instances to find. https://github.com/rektide/nvim-auto-listen/
I'm only somewhat getting started, but the workmanship, fit and finish is just outstanding on Codecompanion, for a fantastically well put together in vim agentic experience. Works really well driving a headless opencode mcp. Being able to stay in vim but still get a great opencode powered experience has been mind blowingly sick.
https://github.com/olimorris/codecompanion.nvim
> But where are the AI features?? Gonna get left behind!
Obviously vim doesn't need AI, but one feature I really wish vim had was native support for multiple cursors.
It's the feature that lured me away to Sublime Text in the first place many years ago, and it's a pre-requisite for pretty much every editor I use these days, from VSCode to Zed.
There are plugins, but multicursor is such a powerful force-multiplier that I think a native implementation would benefit.
The canonical answer to this request is as follows: if you need multi-cursor (or, worse, multi-cursor with mouse support) then you are doing something non-Vim way (aka: wrong way) and there is a better way to do it.
If you need multi-cursor to do manual search and replace in text, then don't, just do automatic search and replace, maybe scoped to a block.
If you need multi-cursor for refactoring or renaming a variable across entire source file, then don't, use LSP plugin (or switch to Neovim) and do the proper refactoring action.
Sure, there are legit cases of using multi-cursor in Vim, but they are rare. So it's not worth to put it into Vim itself.
I agree and I know what you're saying, but I'm pretty curious: how are people using AI with vim? I've seen some scripts for ollama but what are most people doing?
I imagine with vim, from the document you're editing, you'd go:
:ter
to get a terminal. Fire up aider with --watch-files in the terminal. Hop back up to the file and start telling it what to do. Hit L when it's done to see the changes.
That's just a guess but after writing it out I kinda want to try it.
When I use aider it's via its chat interface and then I load the file with vim in another terminal tab to follow along but I think --watch-files with vim would be fun.
Does it count if I share my experience with AI and nvim? I use it to update my configuration, discover new plugins, write custom lua code (I don't know lua) and inquire about motions that would help me in specific workflows. I started learning vim motions last summer and AI really lowered the entry barrier and allowed me to focus on the motions rather than the setup.
Also related to my nvim workflow but not strictly vim related: I use AI to write and update a bash script that handles tmux windows. Again, it lowered the barrier to entry and it made switching to nvim as my primary editor easier.
This. With so much of my work being done with Claude Code via terminal, I’ve used vim and tmux more than I have in the 20 years since I was first introduced.
With all the buzz about orchestrating in the age of CLI agents there doesn't seem to be much talk about vim + tmux with send-keys (a blessing). You can run as many windows and panes doing so many different things across multiple projects.
The way I see it using tmux to orchestrate multiple agents is an intermediate step until we get a UI that can be a product offering. Assuming we get orchestration to the level it has been touted, there is a world where tmux is unnecessary for the user. You would just type something to one panel in which the "overlord" agent is running (the "mayor" if we talking gas town lingo) and that agent will handle all the rest. I doubt jumping between panes is going to stick around as the product offering evolves.
Nearly this, but using ghostty instead of tmux. You don’t get the remote connection aspect of tmux, but for splitting/zooming/preserving windows it is fantastic. The best part is you can configure natural shortcuts rather than using a leader for everything.
AI makes advanced IDE features less relevant (or, more precisely, much easier to ignore or work without.)
I still have PyCharm, especially for working with data which I do a lot it helps quite a bit, but by default I'm back to a very vanilla Vim setup. Others have mentioned tmux which is great and I'd use anyway especially over ssh, but even just terminal tabs for instances of agents are fine frankly.
The announcement itself looks potentially AI-assisted, judging by the bulleted list style and redundant text under the "Charity: Transition to Kuwasha" section. But maybe some people just write that way.
I admit I am guilty of that, although I ran it via several iterations to make sure it covers everything I needed and created a nice style that i liked. It just saves me so much time.
I was happy with VSCode after decades of Vim because it felt light enough out of the box until Copilot starting showing up in every nook and cranny of the damn thing. I switch back to Vim last year.
I have not had that experience, most of the time the duplicate question was answered, but to address the argument, it seems like it would be correct to mark a question as duplicate even if the original isn't answered. Why should there be two instances of the same question with no answer as opposed to one instance with no answer?
Boeing has the contract for SLS, not ULA. Boeing owns 50% of ULA, with Lockheed Martin owning the other 50%. But SLS is a Boeing product not a ULA one. ULA's main rocket now is the Vulcan, with a few more Atlas V launches left.
OK, but what will they do next quarter? Loan out another 20 billion and get it back? And the quarter after that? Eventually you run out of people will to take loans from you to buy your chips and what do you think happens then?
I think you're looking too deeply at this. It's generally well written. I feel like you could take almost any sentence and say "look like AI" if you squint hard enough.
Regardless of it is was fully or partially written by AI, do you agree with the main points? Do you disagree?
Genuinely curious, what did you find painful about it? A while back I found it annoying that I'd get errors when cleaning out my branches because they were checked out in a worktree I'd forgotten about, but git now highlights branches checked out in worktrees and has done so for a while.
Only joking of course, actually quite refreshing to see a new version announcement of something this major without any AI nonsense.
reply