As someone that uses AI at work because 1. this is a run-out-the-clock situation, and 2. I don’t give a fuck and would like to be at my desk as little as possible, it feels weird and hypocritical to expect no AI in all of the software around me as well.
With that said, if I still had any semblance of passion for writing code, which I do not, anything I worked on for the OSS community would be fully devoid of AI usage. But I’m mentally at a point where I’d rather move to a cabin and use computers less and less with each passing day.
I felt a lot like that too. I still do, but this helped me a lot:
Bookmarking this one for later. Thanks! It’s hard not to feel like my entire field has been cheapened. We spent so long perfecting our craft, only for it to to be completely devalued with speed as the main focus.
The bright side is that AI will never be as good as me when it comes to CSS. I’m still the go-to CSS person at work, despite folks having AI tools at their disposal.
I don’t give a fuck and would like to be at my desk as little as possible
I always had this stance and found it weird when people resisted. Your boss’s boss’s boss’s boss is telling you to work half as much? Whether you think that the output is half as good quality or not why the fuck aren’t you JUMPING at the chance to do that.
No, coding should be done to garner clout, so eventually when I start my shitcoin, I can rugpull more people! /j
It’s valid to feel that way working in software, but try to hold space also for the fact that computing is ‘the means of production’ in the information age, and how it is made and who controls it is inherently political.
EROEI in 1950: 50 to 1
EROEI today : closer to 10 to 1
What is the accelerating rise of m-dashes from approximately 2014-2019 about? That trend apparently continued until very recently, even slowed down a bit, before exploding at some point in 2025 or 2026 for obvious reasons.
Does anyone have a theory?
People slowly learned to use their keyboards for proper typography?
That was a good one.
If the graph is not just using made up data, then it also measures absolute em-dashes, not em-dash density, so possibly could be an increase in interest in contributing to the kernel?
By a factor of roughly 10-fold? Maybe. It could indeed be a measure of overal absolute contribution. But post 2025 that could be influenced by m-dash density especially.
No, the comment I responded to was about the ~2014 uptick, not the recent one.
The recent one is obviously AI one way or another. Absolute contribution rate is higher due to AI as well.
Yeah sure, no I am not surprised about post 2025. My question was indeed about the post 2014 uptick.
The OP graph is just absolute count on the y axis, so I also assumed it was basically measuring LoC, but looking at this chart, the em-dashes do appear to be definitely more of a modern phenomenon compared to raw LoC:
Maybe some policy/tooling change around 2014? A guess is perhaps something to do with allowing UTF-8 in source files, but that’s a pretty long shot.
Looking at that, the lines of code cannot really be the full story. M-dash count really just appears to rise from 2014 onwards.
I assume it’s due to phone keyboards.
I don’t think a lot of kernel devs are programming on their phone
That’s why the number was low before.
You can get a lot done when you’ve got 29 months in a year
Yeah sorry about that, was the default format in my spreadsheet software and I didn’t think to change it
?
The weird US American Date scheme is not used in most of the world.
I wish. Am European. I know one person that voluntarily uses their date and time formats. For reasons I will never comprehend.
He or she must enjoy the endless confusion and friction in the interaction with others. That said, some idiotic IT systems use the US format to. Causes a lot of problems.
D/M/Y > M/D/Y
There is only YYYYMMDD
sadly for everyone that’s factually incorrect, but the phrase DD/MM/YYYY > MM/DD/YYYY does not preclude YYYYMMDD > DD/MM/YYYY > MM/DD/YYYY
your pedantery is apreciated.
Because one lives the other should not long live Y-M-D
NO. You need a character per digit and you always want the leading 0 shown.
YYYY-MM-DD
At least it makes my folders really easy to sort through…
Only if you leave the leading zeros!
…or your file manager is smart enough to do numeric sorting
ISO 8601 is the only correct way.
OP’s graph uses slashes.
Dashes would remove all the confusion.
No because it’s the global format that typically uses periods.
D.M.Y (normal)
Y-M-D (superior)
M/D/Y (mental illness)Mixing up the formats and denominators makes this mess even worse.
you should also always specify YYYY-MM-DD because often M and D ignores leading zeros, YY means the 2 digit suffix, and Y… well, i don’t even know what that is but i’m sure it’s something less specific than YYYY
if we’re going to be picky about being clear about things, let’s be truly clear about things
The dates below the graph
What about them? They’re normal dates.
Pretty much only in the US.
Some BS AI “visionaries” keep blurting.
Emdashes on this chart are the Jaws theme of vibeshittification.
—
—
—
———
————
——
——
———— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— —— ——
Personally the only part I have seen is documentation is now vibe coded, where in the past, it was autism coded.
As long as it’s short, sweet and to the point and not prose.
“This function doesn’t return an
int*, it returns achar *.”deleted by creator
(Ftr; you’d also see this sort of thing if a new cohort, say ESL speakers from a particular country, was suddenly able to contribute when they couldn’t before.)
Man, I miss when Linux felt like a refuge. It’s always been flawed of course but now it feels like everyone in positions of influence are embracing the awful
And I remember the days when barely any hardware could run Linux. When lots of the interfaces looked like absolute shit and needed a mountain of configuration to even get running. Without commercial support Linux would have never gotten anywhere near the general purpose OS it is today. If you want a pure hobbyist OS run FreeBSD. If you haven’t moved over there, I’m pretty sure you’re just being a hypocrite.
It’s getting harder to justify the inconvenience of going off the beaten path if that path is turning into the same slop as the commercial ones.
I imagine there’s been quite an effort by commercial interests to influence the direction of this path
The Linux Foundation has unfortunately become a commercial shithole, run by a clusterfuck of Silicon Valley VCs. When I see the kinds of garbage AI projects that are being shoveled into it every day, I lose more and more respect for it.
Look for yourselves: https://www.linuxfoundation.org/projects
And it’s sad to see Linus falling into their lock and step. I respect him for a lot of things, but when it comes to AI, he’s sadly just wrong.
For me most of the projects at the bottom link to “Page not found” Confluence pages…
Holy fuck, that is a lot of AI hype projects.
The beaten path is sloppifying a lot faster than the alternatives and where management gets angry if you don’t use AI.
True, and for most things you get a choice, in contrast to Windows/MacOS. But if the kernel itself slides down this path there’s not a lot you can do unless you’re willing to sacrifice enormous amounts of time and effort.
I’m kind of done with this passion, I’ll be investing more of my time into yard maintenance and my boat.
Is that graph actually real/based on data?
Yep, I cloned the git repo and wrote a script to step through commits a week at a time and count the total em dashes in the repo at each point in time
git grep -I '—' master | wc -ljust did a quick check, my graph looks similar.
As an actual human who uses em dashes — or usually, two or three hyphens that turn into an em dash in Markdown — I feel personally attacked.
Found the bot
Microsoft Word also appears to autocorrect dashes into em-dashes. I’m not sure it’s because of copilot since I noticed around 2024, but a bunch of documents I use at work, drafted/revised in 2021, are full of them.
I will usually just use one (1) hyphen for that purpose - in the much same way that I will use "…" instead of “…” - because I like to upset typography nerds.
Have you tried using the dedicated triple dot character (…) instead of three separate dots (…)? It’s a life changer
They said they do it to ‘upset typography nerds’ — likely they tried…
Same, but I’m unlikely to use them in comments
aren’t those regular dashes? em dashes seem just a bit longer
N dashes. They are as wide as an n. M dashes are as wide as an m.
But what is em?
It’s a width same as point size of a font: https://en.wikipedia.org/wiki/Em_(typography)
deleted by creator
Sure. You can totally avoid them. Just by using very basic sentences. The same is true for a lot of puctuation.
But that doesn’t make your sentences planned better but is simply a sign of limiting yourself by avoiding any complexity.
Bless your — illiterate — heart.
Confidently incorrect. Read a book now and then.
Read a book ✨
No.
What does this say about how LLMs plan their sentences?
Important software that caught buggy code from Linus at review stage? Riight, we are all gonna die, sure
when using a monotype font, shouldnt n and m dashes look the same even?
They do, but I still use the correct dash in code comments. It’s easy to type with a compose key. I don’t know; I’ve just always done it. I think maybe because English is not my first language so I spent a while learning all the grammatical rules; it feels wrong to deliberately break them to indicate an informal tone, the way a lot of people do online.
The graphical representation does not matter, they are still different character
grep doesn’t care about font.s
I’ve been using linux for 18 years, but since reviewing the Open Slopware list I’ve been making a lot of changes.
FreeBSD ticks almost every box I need it to. It runs Plasma, it runs Steam/Proton (through a thin compatibility layer that enables a lot of other linux software), it can be set up immutably and has superior containerization natively via jails. It even supports XDG compliance.
But there is no support for changing voltage or clock rate on any amd gpu (maybe any gpu at all) and no driver support for my newer amd gpu at all.
It honestly feels like how linux used to feel back in the day 😅
I don’t think I’m gonna jump ship just yet. The recent ai code fiasco that happened in the linux kernel (race condition, regression, vacation) was averted by the team of talented human developers who reviewed Linus’ code. I am choosing to have faith that it will be ok for a while… given there are few alternatives.
That being said, I think I might finally be ready to hop on the fuck-sysyemd train. I was able to look past the megalomaniacal tendencies of the dev because the tools were so convenient, but the AI PRs and code review pushes me over the edge. I am gonna configure a daily driver in Artix soon.
Edit: I should note, FreeBSD also has some AI code issues. But the linux-libre kernel might not. Something like Hyperbola (GNU/Linux-libre) might be more appropriate for someone who is trying to avoid ai code strictly.
I recommend you looking at RedoxOS. It’s still not stable, but it’s evolving fast and has an anti-AI instance. Also it’s insanely efficient and optimized.
Last I tried, BSD didn’t like the WiFi on my laptop, so it’s currently a no-go for me.
Gonna survive on Linux for now, but at least I know I have a potential refuge elsewhere.
I use Gentoo with OpenRC btw.
can you elaborate on the recent fiasco you mentioned? I totally missed the news on that
Linus admitted he uses LLMs sometimes, so now the most trustworthy developer in the history of computers is now being shat on and raged about by people who probably couldn’t write a hello world app in any programming language. They’re just parroting the idea that Linux is vibecoded now, but it isn’t, and if anything they should be heartened by the recent decision to raise the bar for kernel contributors. Instead, they’re pretending Linus and other contributors aren’t trustworthy at all, based solely on a desire to feel outrage. When people correct them on these lies they’re spreading, no one ever admits they are wrong and stops spreading it, they just don’t respond to comments like that and continue writing them elsewhere.
Tldr; people are stupid.
You didn’t answer the question at all.
I take it you’re another outrage junkie who cannot help but tell everyone the kernel is vibecoded. Your feedback is only valuable in that it communicated that
Linus introduced a race condition and a regression with a AI generated code, then handed off fixing the problem to someone else because he was going on vacation.
Fiasco might have been an exaggeration, but it’s still pretty lame.
Race conditions are not unusual, and it was caught during review, so there’s really no problem here.
This was caught during review, so nothing was “introduced” quite yet. And there’s no indication that Linus wouldn’t have produced a race condition (a notoriously more easy to miss kind of bug) without the use of AI tools.
It’s only being exaggerated so much because he used AI.
Yea he basically said “This seems to work but is slop, lead Dev, make it real.”
That is something I am surprisingly okay with. Quickly figuring out if something is feasible is good, and doing it in a back of the napkin style is the norm, even without LLMs. Of course then you need to do it properly, without falling into the trap of reimplementing the slop.
I could attribute the race condition just as much to their use of the C language as to AI.
Race conditions I’m pretty sure aren’t language specific. Rust can have them too.
Yes, but the tools provided by Rust:
- Prevent most common ones via aliasing^mutability, Send and Sync marker traits, etc.
- Require explicitly marking the cases not handled by this, e.g. unsafe features require an
unsafe {}block to be used.
Great
This doesn’t make me nervous at all…
It shouldn’t, because it’s a lie.
pretty lame
lame /lām/
adjective
Disabled so that movement, especially walking, is difficult or impossible.
:/
This is why dictionaries list more than 1 definition and you shouldn’t stop at the first one:
- weak; inadequate; unsatisfactory; clumsy.
a lame excuse.
- Slang: Sometimes Disparaging and Offensive. awkward, dull, stupid, or uninteresting.
- Slang: Sometimes Disparaging and Offensive. out of touch with modern fads or trends; unsophisticated.
People who write this dictionary should listen to the public more maybe. Lame doesn’t mean what you wrote for at least 25 years.
Not what happened. I’m not digging it up but this story has been debunked repeatedly
(you’re welcome to just point folks to a debunking)
People are welcome to stop spreading the falsehoods. In every thread where this is said (including this one), the truth is pointed out but those spreading it couldn’t care less. Just scroll. Others are pointing out the details here just like every other place.
Literally you got hit with an linux mailing list saying this is what happened?
Those words don’t make sense
I really love the emdash. I’m very—very—upset ai ruined it for me—cunts.
Rebel. Start using triple em dashes⸻see?
Fu———ck yeah.
This reads like Haiku
It could be a haiku. They don’t need to be 5-75 syllables. It’s more about an observation or a moment of clarity/realisation. The first two lines set up a concept or image, and the third line closes out the thought, usually with contrast. Just have to structure it properly.
I really love the emdash,
I’m very—very—upset ai ruined it for me.
Cunts.
I once tried to write a haiku,
Whose syllables were out of skew.
The words started flowing;
I just kept on going,
A Limerick will just have to do.5-75
You have been reading much longer haikus than I have 😜
That’s how you know that comment isn’t ai. No dash.
timetoproveourhumanitybyrejectingallpunctuation
Oh, that reminds me: is Haiku still a slop-free OS?
Yeah, they’re based
Slop and driver free
Turns out AI learned from humans, who learn from the writings they’re exposed to
It’s an endless cycle, and we’re only in the first revolution
And one solution would be to just turn off all grammar and spell check and let people’s fat fingers and inability to spell let them stand out as the human. Oh boy.
Ya except the internet called everyone grammar nazis for caring.
I did that over a year ago. It was always misinterpreting what I was trying to say and fucking it up, so this is honestly way better.
On the plus side, drunk people would have a MUCH harder time starting arguments…
So, I’m guessing you don’t meet too many drunk people…
Ypu think I cant spell drunk amd anget huh bitcj!?
Greta jda
It’s not just Linux, it’s GNU + Linux
I’d just like to interject for a moment. What you’re refering to as GNU + Linux, is in fact, Linux
It says Linux Kernel
It’s a joke about common LLM-gen-AI text patterns
🟥 Linux Kernel — CORRECTED: NOT GNU/Linux
