The LLM part is bad news but it isn't the sign.
-
RE: https://woof.tech/@techokami/117439226690379216
The LLM part is bad news but it isn't the sign. It's the talking about planning succession while simultaneously complaining that nobody new wants to learn, that's the sign.
Y'all the kernel is in deep, deep trouble. We saw what happened to Perl and the kernel is in deep trouble.
@mhoye tldr; #LLMs are here to stay. People who actually deliver working solutions that shape the world have spoken. Measures to limit slop are being developed. #Linux becomes even more commercial.
I personally hope more open models will arrive, GPUs will become cheaper and universal income will help people who wont be able to find job in such toxic and draining field as computer science.2026 welcome in ultimate hell everyone! Dont forget to enjoy the ride!
-
@cohentheblue I've heard it discussed on Linux Dev Time once or twice, but I had it jumbled up with Zigbee in my head.
-
RE: https://woof.tech/@techokami/117439226690379216
The LLM part is bad news but it isn't the sign. It's the talking about planning succession while simultaneously complaining that nobody new wants to learn, that's the sign.
Y'all the kernel is in deep, deep trouble. We saw what happened to Perl and the kernel is in deep trouble.
@mhoye reading the actual article I am _almost_ sympathetic: the team is clearly inundated with thousands of slop patches, and circling the wagons around just "people we already trust who have write access" is a strategy to protect their limited remaining sanity. Fine. But of course, Linus insists on undercutting all of these arguments by throwing in his personal feelings about responsible LLM usage, strong reliance on the agent triage bot, and general "we want only 10x leet coders" mentality that it easily overshadows anything the other participants may have said or believe.
-
@mhoye reading the actual article I am _almost_ sympathetic: the team is clearly inundated with thousands of slop patches, and circling the wagons around just "people we already trust who have write access" is a strategy to protect their limited remaining sanity. Fine. But of course, Linus insists on undercutting all of these arguments by throwing in his personal feelings about responsible LLM usage, strong reliance on the agent triage bot, and general "we want only 10x leet coders" mentality that it easily overshadows anything the other participants may have said or believe.
@mhoye In other words I strongly agree that: it is NOT possible to turn drive-by vibecoders into real contributors. Especially if they also respond to questions with more Claudespeak emails, don't try to engage, and violate community rules about turnaround time. It's just not worth trying to do that - the time should be spent elsewhere. Hence the search for signals beyond "sending a patch and answering an email", like "going to a conference" to indicate an actual potential contributor
But then th team captain is also dancing in the endzone using the exact same tools and praising their utility to the stars, so like, what can you really say???
-
"If your community isn't accessible and welcoming to new, learning participants, the clock's ticking."
Quoted for truth.
@unchartedworlds @mhoye Back in the early 90s BBS/Fidonet era I tried joining a C++ board and mailing list because I wanted to learn more about the language. Got rejected because I didn't know the C++ black arts. I began to resent programmers for a long time after that.
But social media did what was best-suited for: exclusion of voices that constitute noise to an in-group. Nobody is obliged to be welcoming to anyone. Rejecting outsiders did not stop C++ from becoming a central technology.
-
@hipsterelectron @froztbyte @mhoye This shit right here is why I am not working on my seL4/Idris crap in public. Giving your shit away to comrades is fundamentally different from giving your shit away to the people in the power structures above you, and we need a way to encode and defend that distinction before we can have free software that doesn't feed the monsters.
@lykso @hipsterelectron @froztbyte @mhoye > Giving your shit away to comrades
Beware shutting the door for any stranger yearning for liberation yet not already part of the community. Bootstrapping issues are a problem (and overenthusiastic gatekeeping magnifies it). -
@corbet @mhoye @gvwilson I do also wonder if these problems are in some way an unexpected social consequence of a monolithic kernel with in-tree drivers etc. It seemed like a brilliant idea for years/decades, because internal APIs could be refined without piles of compatibility layers, etc - but it also ends up creating a monolithic community.
I've been wondering for a while if a less performant microkernel(ish) approach with defined APIs would make for a healthier social ecosystem. But I guess industry might not have tolerated that performance hit.
@srtcd424@mas.to @corbet@social.kernel.org @mhoye@cosocial.ca @gvwilson@mastodon.social
It seemed like a brilliant idea for years/decades
It was decried as an awful idea by literally anyone caring about security or maintainability for just as long.
Even before considering many of the corposcum-contributed drivers are so obfuscated and underdocumented they might as well be blobs (some literally do include mystery blobs as variables).
I’ve been wondering for a while if a less performant microkernel(ish) approach with defined APIs would make for a healthier social ecosystem.
It would.
-
@srtcd424 @mhoye @gvwilson There is *one* Rust developer who left after some frustrating experiences. We were sad to lose him. Meanwhile Rust is going from strength to strength in the kernel.
If you look at my Kernel Recipes talk last month, you'll see that, among many other things, I said that the LLM onslaught was likely to cost us developers who don't want to deal with it. To a lesser extent, the advent of Rust may have the same effect. It is not desirable, but I'm not sure what you do about it.
Those who know me know that I am not a big LLM fan. I don't like what they are doing to our community, I fear for the long-term maintainability of a kernel filled with code that its authors don't understand, I detest the time I have had to put into defending LWN against scraper DDoS attacks, I am terrified by the prospect of making our free-software project dependent on proprietary tools owned and developed by psycopaths, I don't believe we can afford the environmental costs, and so on. But the community has to respond to all of this somehow; we're all still trying to figure out what that is.@corbet@social.kernel.org @mhoye@cosocial.ca @srtcd424@mas.to @gvwilson@mastodon.social
I fear for the long-term maintainability of a kernel filled with code that its authors don’t understand
I don’t want to run any such kernel or code, either.
-
@lykso @hipsterelectron @froztbyte @mhoye > Giving your shit away to comrades
Beware shutting the door for any stranger yearning for liberation yet not already part of the community. Bootstrapping issues are a problem (and overenthusiastic gatekeeping magnifies it).@lispi314 @hipsterelectron @mhoye @froztbyte Yes, that's a worry I'm keenly aware of. At the moment, the door is open to anyone who passes the "smell test," so to speak. Ask for a Wireguard config and a username and you'll receive.
I like the idea of extending the federation concept to the Internet more broadly, TBH. So if you set up a Wireguard subnet on a private IPv6 as well so you can independently invite people in and you want to connect networks, LMK.
-
@lispi314 @hipsterelectron @mhoye @froztbyte Yes, that's a worry I'm keenly aware of. At the moment, the door is open to anyone who passes the "smell test," so to speak. Ask for a Wireguard config and a username and you'll receive.
I like the idea of extending the federation concept to the Internet more broadly, TBH. So if you set up a Wireguard subnet on a private IPv6 as well so you can independently invite people in and you want to connect networks, LMK.
@lispi314 @hipsterelectron @mhoye @froztbyte Regarding failure due to obscurity, I think I'd rather wither in obscurity than feed the system I want to starve. I regard "failures" like Oberon and Plan 9 more highly than Linux at this point, and I'd prefer being lumped in with them.
Of course, success beyond my wildest dreams would see this OS or its spiritual successor as the OS of the revolution, so to speak; I'm speaking only to failure modes here.
-
The move to Matrix was partly driven by their open source commitments - hi, @matrix , hey yo, love you all - but they mostly won on the strength of two things. First that their accessibility posture was not just unrivalled in terms of FOSS forums but that their team jumped at the chance to work with us improve it.
But 2nd: the moderation tools are unmatched. Nothing else comes close.
Once you have tools and process in place to ablate harassment, nobody willingly goes back to not having that.
@mhoye The unfortunate thing about the move to Matrix is significantly more discussion started happening on Mozilla-internal Slack at the same time, and that really killed some of the open-by-default vibes, with some people scarcely checking Matrix.
-
@lispi314 @hipsterelectron @mhoye @froztbyte Yes, that's a worry I'm keenly aware of. At the moment, the door is open to anyone who passes the "smell test," so to speak. Ask for a Wireguard config and a username and you'll receive.
I like the idea of extending the federation concept to the Internet more broadly, TBH. So if you set up a Wireguard subnet on a private IPv6 as well so you can independently invite people in and you want to connect networks, LMK.
@lykso@tiny.tilde.website @hipsterelectron@circumstances.run @mhoye@cosocial.ca @froztbyte@mastodon.social The difficulty of anonymizing (most things like Tor don’t support UDP) would have me refrain.
I like the idea of extending the federation concept to the Internet more broadly, TBH.
It has upsides and downsides.
Fidonet and Usenet both had this.
So if you set up a Wireguard subnet on a private IPv6 as well so you can independently invite people in and you want to connect networks, LMK.
I will probably refrain.
At the moment I2P is how I’d be inclined to do so, it also has key/password-limited connectivity & the like for allowlisted access.
An option usable from both the clearnet, I2P and Tor (varyingly based on individual node support) might be something like Retroshare, though I have no idea if it has slopped since I last looked at it and no idea how well-specified its “wire protocol” is.
-
@lykso@tiny.tilde.website @hipsterelectron@circumstances.run @mhoye@cosocial.ca @froztbyte@mastodon.social The difficulty of anonymizing (most things like Tor don’t support UDP) would have me refrain.
I like the idea of extending the federation concept to the Internet more broadly, TBH.
It has upsides and downsides.
Fidonet and Usenet both had this.
So if you set up a Wireguard subnet on a private IPv6 as well so you can independently invite people in and you want to connect networks, LMK.
I will probably refrain.
At the moment I2P is how I’d be inclined to do so, it also has key/password-limited connectivity & the like for allowlisted access.
An option usable from both the clearnet, I2P and Tor (varyingly based on individual node support) might be something like Retroshare, though I have no idea if it has slopped since I last looked at it and no idea how well-specified its “wire protocol” is.
@lispi314 @hipsterelectron @mhoye @froztbyte I should investigate I2P then.
-
@lispi314 @hipsterelectron @mhoye @froztbyte Regarding failure due to obscurity, I think I'd rather wither in obscurity than feed the system I want to starve. I regard "failures" like Oberon and Plan 9 more highly than Linux at this point, and I'd prefer being lumped in with them.
Of course, success beyond my wildest dreams would see this OS or its spiritual successor as the OS of the revolution, so to speak; I'm speaking only to failure modes here.
@lykso @hipsterelectron @mhoye @froztbyte They were eventually published so one can regard them that highly, though.
Perhaps long after they could have had a chance (well, Oberon was a personal university project iirc, it was published in the clear, it was just completely ignored by everyone). -
@lykso @hipsterelectron @mhoye @froztbyte They were eventually published so one can regard them that highly, though.
Perhaps long after they could have had a chance (well, Oberon was a personal university project iirc, it was published in the clear, it was just completely ignored by everyone).@lispi314 @hipsterelectron @mhoye @froztbyte Yeah. I would like to be developing in public, of course, but at this point I am stuck on the point of distrusting the legal system to actually defend small timers like myself, even if there were a well-established license to satisfy my concerns.
-
@mhoye The unfortunate thing about the move to Matrix is significantly more discussion started happening on Mozilla-internal Slack at the same time, and that really killed some of the open-by-default vibes, with some people scarcely checking Matrix.
@gsnedders That didn't happen at the same time; the reality is that IRC usability was, compared to what was available, just so bad for so long that long before a single Mozilla slack instance was blessed, long before I was sent to get us an alternative to IRC, there were already _dozens and dozens_ of private shadow Slack instances all over the place for various teams and projects. Like, on the order of 180 of them. One of the biggest impetuses for moving to Slack was... most people already had.
-
@lispi314 @hipsterelectron @mhoye @froztbyte Yeah. I would like to be developing in public, of course, but at this point I am stuck on the point of distrusting the legal system to actually defend small timers like myself, even if there were a well-established license to satisfy my concerns.
@lykso@tiny.tilde.website @hipsterelectron@circumstances.run @mhoye@cosocial.ca @froztbyte@mastodon.social In my case I have no belief my license works as more than a statement of intent/principles.
So I try instead to compensate in things which I believe the generally applicable contribution to autonomy remains applicable regardless of whatever corposcum might copy it (whatever bastardized copy may not retain that, but the one I will continue to provide will).
I believe finding the source of things matters and that the tooling for current code helps a lot.
Ultimately of course even with protections, corposcum have very well demonstrated in book publishing just how effectively they can drown out one’s signal so that one is never heard of anyway (copyright was a monopolist and censorship tool first & foremost and it is what it does best; the same is applicable to patents in a very similar way). The system has always protected the monopolists and corposcum first, anyone else being a pretense or incidental at best.
-
There are plenty of good alternatives to driving your community off a cliff like a gaggle of demographic lemmings, but if you're still convinced that mailing lists and insularity are the path to durability in this godforsaken year of our lord two thousand and twenty six then I'm not sure what words will convince you that there's nothing at the bottom of that cliff but sharp rocks and seawater.
If your community isn't accessible and welcoming to new, learning participants, the clock's ticking.
@mhoye@cosocial.ca Mailing lists do present a good aspect.
Except that mailing lists rely on mail servers that are part of infrastructure captured by corposcum, trivially and frequently leveraged for abuse by the corposcum or various regimes.
A federation of Free NetNews servers or anything new would do a much better job of it.
The centralized forges of the moment are an active step backward, into the awfulness of server-centrism & infrastructure consolidation in reach of trivial disruption.
There is nothing fundamental about forges that requires them to be server-centric abominations. They could be user-agents for a federated or distributed protocol. Ideally a message-centric one, which is the one part email got right.
In fact, there’s nothing fundamental about email that requires it to use the compromised and taken-over SMTP corporate federation. But email has accumulated some cruft, had some of its limitations come into stark relief when used for various things and there is a good argument that something new would do it better for those things.
-
@mhoye Question is: who is willing to fork pre-LLM version of Linux and maintain that? That would be the most practical solution.
AFAIK most people who care walk away from Linux to less mature NetBSD or RedoxOS, which do have strict anti-AI policy.@tapafon@isn.tapaf.onl @mhoye@cosocial.ca LLMs are the straw that broke the camel’s back.
There were already a lot of problems generating considerable discontent. They just hadn’t be quite enough for those people to decide to leave Linux entirely yet.
-
@lykso@tiny.tilde.website @hipsterelectron@circumstances.run @mhoye@cosocial.ca @froztbyte@mastodon.social In my case I have no belief my license works as more than a statement of intent/principles.
So I try instead to compensate in things which I believe the generally applicable contribution to autonomy remains applicable regardless of whatever corposcum might copy it (whatever bastardized copy may not retain that, but the one I will continue to provide will).
I believe finding the source of things matters and that the tooling for current code helps a lot.
Ultimately of course even with protections, corposcum have very well demonstrated in book publishing just how effectively they can drown out one’s signal so that one is never heard of anyway (copyright was a monopolist and censorship tool first & foremost and it is what it does best; the same is applicable to patents in a very similar way). The system has always protected the monopolists and corposcum first, anyone else being a pretense or incidental at best.
@lispi314 @hipsterelectron @mhoye @froztbyte Yeah, this is the right attitude to have regarding licenses now, I think.
I don't think there is anything which is structurally incompatible with cooption, and if it is coopted then it becomes irrelevant whether I am maintaining a "good" version; it gets drowned out. Worse, my maintenance may be upstreamed and used to benefit the capitalist version.
1/
