News for hackers. Stuff that matters.
Re:Emacs in the Age of Magic by jefurii on Wednesday November 11, 2026 @06:06PM (#50460233)
[delayed]
Re:Emacs in the Age of Magic by fock on Wednesday November 11, 2026 @07:51PM (#50461423)
good to see treesitter is now vibecoded too...
Re:Emacs in the Age of Magic by acedTrex on Wednesday November 11, 2026 @11:46PM (#50463165)
Woah since when, that one hurts
Re:Emacs in the Age of Magic by fock on Thursday November 12, 2026 @01:27AM (#50463739)
it's on the responsible side of vibecoding. Effectively since a person took over who decided that Claude helps them go brrrr...

The way we are accepting this is extremely unsettling...

Re:Emacs in the Age of Magic by mickeyp on Thursday November 12, 2026 @01:47AM (#50463848)
That's obviously the best use of an AI's time.

Have you ever written parsers for complex languages? It is a cumbersome slog.

Re:Emacs in the Age of Magic by hahn-kev on Thursday November 12, 2026 @02:02AM (#50463965)
Yeah it's also super easy to automate testing.
Re:Emacs in the Age of Magic by fock on Thursday November 12, 2026 @10:07AM (#50468583)
that's why they then proceeded to stop accepting PRs because of AI-slop. Funny, when the current dev appeared on stage mid 2025...

It seems quite cumbersome that this is, how this is supposed to work now?

Re:Emacs in the Age of Magic by bitwize on Thursday November 12, 2026 @12:25AM (#50463373)
Gonna get its milkshake drunk if GNU decides to allow AI.
Re:Emacs in the Age of Magic by vconnor on Thursday November 12, 2026 @12:46AM (#50463523)
How could they survive with fewer users and lower revenue?

Ah right, it’s free software. There is no milkshake, only people with similar ideology contributing to a common cause.

Re:Emacs in the Age of Magic by bitwize on Friday November 13, 2026 @04:23AM (#50480617)
The "milkshake" is developer capacity. Projects which ban AI lose doubly:

1. Ceteris paribus, the developers won't be able to deliver new features as quickly, nor fix bugs as reliably, as projects which use AI.

2. They will be unable to attract and retain a large number of developers because there are just not that many people who code by hand anymore; rational people can see the massive productivity gains AI gives you

And of course, the user base will be siphoned off entirely to the fork which gets more and better updates, leaving "Human Emacs", if it ever comes into existence, a largely abandoned project within short order. This has already happened, albeit not in the AI context: in the mid-90s, XEmacs was leaps and bounds ahead of GNU, and today it's pretty much abandonware.

There's just no future in coding by hand. It's like trying to keep punched cards alive. Forget it. It's done. Pencils down. Embrace the future, or face irrelevance.

Re:Emacs in the Age of Magic by brabel on Thursday November 12, 2026 @11:01AM (#50469567)
This will enter the history books as one of the greatest examples of a group of people who appear to love technology inexplicably becoming the most feverishly opposed group against a new technology that happens to replace the craft that they used to claim as their own. Never did they stop to think about the millions of people who lost their jobs because of computers, many because of programs written in emacs. Of all communities, this is the most surprising to me since Lisp was THE AI language for most of its existence, but since it doesn’t happen to be used by the new generation of AI technologies, I suppose AI is now the enemy. This is pathetic. I love emacs and all I can say is that emacs is actually making a comeback for me due to AI since now all I need is a simple text editor that lets me quickly type in some code when I am feeling nostalgic. And let my LLM finish the tedious parts I don’t want to do, thanks to gptel. There are many other LLM related emacs extensions which clearly shows a lot of emacs users are not prudes when it comes to LLM. This post absolutely does not represent my view as an emacs user, and that of most others, to the best of my knowledge.
Re:Emacs in the Age of Magic by dmz73 on Thursday November 12, 2026 @07:52PM (#50476688)
Computers made a lot of things easier, faster and more reliable until complexity became too great for people to contend with. They also created far more jobs than what they removed. LLMs are excellent bullshit artists. They can produce large quantity of very plausible output but you just can't trust it and since the volume is so great it is difficult to inspect or understand so it is essentially useless if you want quality. If you just want quantity, they are great.
Re:Emacs in the Age of Magic by bitwize on Friday November 13, 2026 @04:27AM (#50480662)
You can still work with an LLM and realize huge productivity gains with NO reduction in quality. Just review and vet the code you generate with an LLM, a little bit at a time.

This is not going to give you the velocity that a dark factory will give you, so it will not be much used in the corporate world except in Space Shuttle type scenarios, but for open source where they have a quality bar and manual review and signoff anyways, it's still a JATO bottle strapped to your car compared to what we were doing before.

And if you pass this speed boost up, you're just a damn fool.

Re:Emacs in the Age of Magic by brabel on Friday November 13, 2026 @05:48AM (#50481577)
If you think quality cannot be achieved with LLM, I recommend you try using AI to review your code. Use an existing skill to do it, all harnesses have one but you can customize it to tell the AI what to look for. I can guarantee you will change your mind.