Bozhidar Batsov (a.k.a. Bug)I jotted down some thoughts on the future of editors like #Emacs and #Vim in the age of AI https://batsov.com/articles/2026/03/09/emacs-and-vim-in-the-age-of-ai/
I'd be very curious to hear your thoughts on the subject as well.
nthcdr@bbatsov I'm worried about the future. AI is getting pretty darn good and I primarily worry what that means for my own wellbeing both when it comes to keeping my job long term, and what it means for keeping my sanity short term. I've only been full on ai-assistedly coding for a few weeks and I feel like my attention span is shrinking and I'm getting more stupid from it.
I have a feeling that an AI-induced problem going forward will be tackled soley by an AI-solutions
Torstein Krause JohansenI believe it's a slippery slope: Once you start using AI assistance for coding, you soon become lazy. After a while (half a year? Two years?) you're in practice dependent on AI to do any coding. You cannot code without it. In the beginning, you review each line, are careful you understand, really understand the code AI writes. As time goes on, you let it slip, let it fly. The vibing colleagues and AI agents throws so many PRs at you that you can't possibly review each line. Onwards.
At one point, why bother to review the code at all? Why not have the AI write a bunch of tests and trust that that's good enough? How about vibing some AI agents to scrutiny the code from different angles (security, performance++)? That surely fixes the problem of the annoying feeling that humans should review and understand this, surely?
Then, later, you might ask: why does the language, libraries and frameworks matter if humans are never to understand or touch the code? Why not just go for foolang, since that works the best with AI agents?
My day as a coder consists of 10-30% coding. The rest are meetings, meetings, meetings, emails, reports, code reviews and meeting preparations. I wish AI would help me with the 70-90% of the non-coding work, so that I can do the fun bit. Instead, we're investing SO much money to solve the 10-30% part, so that we can spend that time reviewing the code the bot generated.
As with all challenges at work, I believe the key is to do what Beatles suggested and "get by with a little help from my friends":, i.e. focus on the people you work with and ensure you are in a place where you enjoy the company you're keeping.
Alfred M. Szmidt
Dawid@nthcdr @bbatsov don't use it to do the work for you. You should do the work and then ask it for advices and suggestions how to improve it. According to MIT research that showed that people are in fact getting more stupid from using AI it was the only path that showed that people were using brain more than before using AI
Christian Tietze@bbatsov Emacs as the text-buffer-based abstraction for personal computing is *amazing* today.
I wrote an MTP TRAMP extension to copy files via dired onto Android devices.
I now have a mise-en-place manager to interactively remove, install, upgrade, ... CLI tools per project and globally.
A PARA filing scheme helper, too, with org agenda integration.
The integrations you listed makes _acting_ on Emacs buffer contents as input from the 'IDE' trivial.
Without it, we'd only have .md files!
Omar Antolín@bbatsov Good piece! One minor correction: I think you completely misunderstood the purpose of the Evi fork!
You wrote: "The Vim community recently saw the creation of EVi, a fork of Vim whose entire raison d’etre is to provide a text editor free from AI integration."
I don't think the main goal of the fork is to keep users from integrating AI into their workflows, I think the main purpose is to avoid having LLM-generated code in the editor's source itself. You'll notice that they forked right before the first publicly recognized LLM-generated commit. And it's not like they have added any code that would keep you from using any of the AI integration plugins you describe in the post.
@bbatsov Oh, now that I've seen your other post¹, I think I just misunderstood how you use the term "AI integration" and you have the same understanding that I do about why EVi was created. Sorry. But I would posit that your wording is confusing and others are likely to misread it the way I did.
Bozhidar Batsov (a.k.a. Bug) (@[email protected])
Seems someone's so unhappy that #vim and #neovim accept AI-assisted contributions that they forked vim https://codeberg.org/NerdNextDoor/evi/issues/1 Seems like a somewhat pointless uphill battle to me. If the contributions are of high quality and solve real problems they are perfectly fine in my book. It's not like the number of C, VimScript and Lua experts has been growing a lot in recent years. ;-) AI agents might help make some niche technologies less niche, which is not a bad thing...
@oantolin Yeah, that's a fair point. Poor choice of words on my part - I'll update that.
NerdNextDoor 
@bbatsov @oantolin Hi, I'm the creator of the EVi fork! Not sure what this said originally, but what it says is correct.
"The Vim community recently saw the creation of EVi, a fork of Vim whose entire raison d’etre is to provide a text editor free from AI-assisted (generated?) code contributions."
I hard-forked the last non-AI commit of VIm to make an AI free VIm, no slop code.
As a result, we've been rebranding away from VIm too.
Happy to see EVi get coverage though, so I thank you greatly!
P.S: If you want a bit more on what we mean by AI free, you can read our CONTRIBUTING.md here: https://codeberg.org/NerdNextDoor/evi/src/branch/master/CONTRIBUTING.md
Hey @bbatsov, do you think you could modify your RSS so that it has the full content of each post, please? (I keep meaning to ask you that, and keep forgetting.)
@oantolin Hmm, I thought that was the case already. Which RSS feed exactly are you referring to? (as I have 3 different blogs :D )
@bbatsov This feed only has summaries, I believe: https://batsov.com/feed.xml
(For or a minute there I thought I was wrong because my feed reader was now showing me the entire post —it has a feature to download the entire post when all you get is a summary and apparently it caches the answer. So I had to check the feed URL directly.)
@oantolin Hmm, I'll investigate!
@bbatsov This other feed does have full posts: https://emacsredux.com/atom.xml
What's your third blog? I like the other two so I'll probably like the third one too.
@oantolin metaredux.com (it was meant to replace batsov.com, but after a while I changed my mind and tried to find different purposes for both)
I overdid the blogs back in the day. :D
Steven G. Harms@bbatsov enjoyed this! After doing some editor hopping from vi, to emacs , to textmate, back to vim, and having recently upgraded to neovim in the last month, I think your observation that AI can make these esoteric internal languages more accessible is spot on.
On top of that, lately I’ve been prompting so well with Claude code that I’ve just not needed vim to put in code. I use vim to edit prompts and configuration files, but I’ve lately taken back to using raw EX editor instead. It’s shockingly good.
Frankly, all of the noise that VS code shoves in my face between terminal, prompt, and code is overwhelming.
Ex at the top seems to be enough. More experimentation is warranted though.
And finally, I completely and with you that it seems baffling that so many techies maintain that AI can’t do anything. I have tons of things that has helped me do, how is that a dependable position? I don’t get it.
I was roundly mocked for suggestion FreeBSD ought try to get Claude support. Baffling
Svante@bbatsov I'd mainly want to discuss on the basis of the chapter »Do you even need a “power tool” anymore?«.
You write »all of it is in service of making the human more efficient at the mechanical act of coding« — but that's not the point: it's to give the human a faster way to put their thoughts to canvas, thus reducing interruption of those thoughts, and to give them more time to /have/ thoughts.
Put that way, I don't see how LLMs help at all, and frankly, I actually do believe they don't.
@bbatsov You write: »Cursor can scaffold an entire application in an afternoon« — but it can't. It can throw together random existing snippets that look like they might plausibly work, and iterate on them until it compiles and, if it is nothing innovative, might even resemble something you expected too.
Imagine a random PO send you some code with the note »hey, I just made this, can you put it into prod?« — aren't you shuddering? This is just automated script kiddies with even less props.
@bbatsov These two assumptions — about the goal of an editor, and about the quality of LLM output — are just two facets of the misconceptions that lead some people to actively drown the world in slop.
No, the job of actually understanding will not go away. But a lot of people will get hurt on the way to understanding that.
I believe that it is our job not to casually spread or tolerate such misconceptions, but to try to mitigate the damage.
@Ardubal I don’t think this is as clear cut as you portray it to be. I understand very well how LLMs work and what are the caveats related to their usage, and this doesn't change the fact you can leverage them to produce useful (and good quality code). AI is still just a tool, and the outcomes depend vastly on how you're using the tool. I agree that AI agents can't innovate, but in practice for most tasks you don't really need to innovate anyways.
But I don't really want to discuss AI's role in programming, but the role of Emacs and vim in a world that's increasingly dominated by AI-assisted coding.
@Ardubal I get your point, but how is this related to power editors like Emacs/Vim? The point of my article is the future of programming and its relationship to the future of editors used primarily for programming. Perhaps that wasn't clear, though.
Jess Robinson@bbatsov "like everyone else in the industry" nope not everyone