Vim has long been the standard edition of the venerable vi editor, which turns 50 this year.
Both Vim and its successor, NeoVim, explicitly allow LLM contributions. Vim’s policy says:
When using AI for contributions, please disclose this.
And NeoVim’s policy says:
Using AI for contributions is acceptable
Both policies give the usual parameters about reviewing one’s work and not submitting AI contributions blindly, but fundamentally, both projects say “AI is OK”.
There are those, however, who don’t like that.
The first is the politicized Vim Classic, which peppers its home page with various national and movement flags. I always find people who drag politics into every facet of life extremely distasteful (sure, war is bad – but what does that have to do with text editors?) They state:
I’m proud to announce the release of the first version of Vim Classic, Vim Classic 8.3.0, a stable, long-term support fork of the ubiquitous text editor, maintained without the assistance of generative AI tools.
less politicized but more foam-mouthed about AI is EVi, which runs on Codeberg. It took me about a half hour to get Codeberg to return a git clone, but it finally did, after numerous 502s indicating that the system was nonfunctional. Perhaps EVi is on that platform because they don’t like GitHub’s AI. But whatever the reason, while they are less concerned about fighting the world’s wars, they are very upset about AI:
Using generative AI for patches is not accepted in EVi. Attempting to send in AI-written patches or AI-created PRs will result in a permanent ban from contributing with no chance of appealing.
No chance! You are CANCELED!
So…will either project go anywhere? I doubt it.
While the vi editor is widely used, it’s not the sort of project that has thousands of contributors. Of the small population of people who write code for the vi family of editors, how many are up in arms about AI? A small fraction, which is now split among at least two different groups.
You already have Vim vs. NeoVim. That doesn’t leave many minds left over for Vim Classic and EVi. The latter two projects offer nothing besides “we did not use AI”. For users, this isn’t much of an advantage. In fact, considering that these products’ progress will be slower, it’s a disadvantage.
Does this herald a new era of “pure” non-LLM products and fractured LLM-OK vs LLN-not-OK codebases? First, how does anyone know if a dev is using AI? If I’m working in VS code and ask it to analyze a block of code or ask it to generate a function which I then review and insert into my codebase, there is no dark immutable stain that is attached to the code. You’re relying on the honesty of the contributor – and yes, contributors should be honest. But as workflows evolve, these AI tools will become more prevalent. VS Code and other IDEs will be radically different in years to come, as collaborating with AIs becomes standard.
So I guess if you want to contribute to EVi or Vim Classic, you’ll need to work exclusively in those editors. Sure, there are grognards who’ll do just that, but I suspect these projects will fade while Vim / NeoVim continues to flourish.


















Leave a Reply