EVi, a Hard-Fork of Vim (codeberg.org)

I was going to comment how it might be ironic to call the project evil instead, but remembered that's the name for the vim emulation on emacs.

What is AI-taint and why is regular vim tainted?

It would be nice if specific offending portions of the codebase were highlighted. As of now, it’s hard to see why one should use this fork. Also, since the source is available, anyone can just compile a past version of vim.

I tried to find any recent issues related to AI in the Vim repo, but did not find any.

Offending commit https://github.com/vim/vim/commit/fc00006777594f969ba8fcff67...

Just Claude as a co-author.

> while avoiding AI taint

Don't be shy. Tell us what you really think.

Why am I entirely unsurprised that this anti-LLM hard fork is hosted on Codeberg? At this rate, how likely is it that Codeberg is just going to become a wasteland of abandoned ideological forks (with the exception of one or two major projects) by next year?

> to build further upon the foundations of Vim, while avoiding AI taint.

Did I miss something ? Where is the AI Taint coming from ?

This kind of news are about nothing. Tell me in a year, even in 3 months, how your fork has been doing. Clicking Fork on gh and writing a blog post is not a fork. A fork is a lot of work. Color me surprised if this will even keep up with the upstream just filtering out the AI commits.

> https://codeberg.org/NerdNextDoor/evi/issues/19

This does not inspire confidence in the maintainers.

Honest question: why would anyone use Vim and not NeoVim nowadays? I've switched what, 12 years ago? And I've never had to look back. Just curious, to be honest. Especially since neovim is full of new features, while the Vim9 scripting language kind of tanked

Forker’s bio:

https://codeberg.org/NerdNextDoor

”I am NerdNextDoor, an autistic OS developer from Scotland who is heading to College soon with the end goal of doing Computing Science in University with some experience in (not much of any, but a bit of) Kuroko, C and Assembly in some architectures.”

Indeed.

Same motivation, different generation. I carry my own fork of vim 5.7, from around 2000. It did what I needed, and did it well, and I could already see where it was going.

SMH at what I see in it now!

Fabulous! Compiled & installed on my Fedora dev VM, no issues. Glad to see the effort made, and I expect we'll see a lot more projects like this going forward as slop unfortunately pervades more OSS.

"AI Taint" you mean code that is written better and faster than humans? I've been programming since 2002 and I consider myself a better than average programmer and I have no problem admitting that using Claude Code has dramatically improved the quality and speed of my output.

Very nice, will look into it myself. Now if only they can set the defaults back to what vim 6.x used I will be happy.

The new defaults drive me crazy and I had to update ~/.vimrc to disable them.

This is what I had to add, but once in great while a default will be reset which I have yet to figure out why that happens.

    :let no_man_maps=1
    :let loaded_matchparen = 1
    :set comments=""
    :set matchpairs=""
    :set mouse=
    :set nocin
    :set nohlsearch
    :set noincsearch
    :set nois
    :set paste
    :syntax off
    """ disable autocomments
    :au FileType * setl fo-=cro
    """ turn off brace match comment matching
    set noshowmatch
    """ stop autoindent
    set noautoindent
    :filetype indent off
    ""set columns=80
    set noundofile
    set showmode ruler
    " turn off a new vim 8 default
    set scrolloff=0

The reason for the fork:

> EVi is a hard fork from Vim v9.1.2073 (Jan 2026) to build further upon the foundations of Vim, while avoiding AI taint.

I wish they would have included at least one example of the "taint" on vim. If it's "tainted" enough to justify a hard fork, then surely there's some evidence. Without that, it just rings like an ideological thing.

Does anyone have some example(s) they can point to?

The AI genie is already out of the bottle; there is nothing you can do to avoid it.

I have been slow to accept it, but here we are. I think the best way forward is to write better, more opinionated linting and testing for your project. If there are certain stylistic choices you prefer, explicitly test for them, and automatically fail PRs that don't pass. If something gets through that you don't like, update your linting.

Which ironically, would be more difficult for humans, but at least your incoming contributions will be better.