[–] arjie 8 hours ago
[–] ahsgq 8 hours ago
[–] joshjob42 6 hours ago
I'm guessing you don't have any tasks in your life that involve coding, math, science, or law if this is your view.
[–] plastic-enjoyer 6 hours ago
[–] ImaCake 8 hours ago
[–] vlyan 6 hours ago
there's thread on the front page of reddit right now, the usual drivel with a literally who on twitter snarking about 50B spent on Iran war instead of free healthcare. those who'd committed the thoughtcrime of dividing 50b by 342m and suggested that 146 bucks per person doesn't buy much healthcare are downvoted into oblivion.
ai is simultaneously taking everyone's jobs and a bubble that's about to burst.
russia is simultaneously a bankrupt paper tiger on the brink of total collapse and the greatest threat to mankind on the brink of invading every wholesome democracy on earth.
etc etc.
[–] KingMob 1 hour ago
I suspect the criticism isn't as absurdly narrow that ALL we have to do to get M4A is swap out the war funding. It's probably people pointing out that the govt wastes plenty of money on immoral and useless things, of which the Iran war is an example.
[–] GlacierFox 7 hours ago
[–] Viliam1234 6 hours ago
Reverse-engineering code.
Finding vulnerabilities.
Reviewing specifications.
[–] GlacierFox 5 hours ago
[–] jryle70 5 hours ago
[–] poly2it 7 hours ago
[–] GlacierFox 5 hours ago
[–] KingMob 1 hour ago
Also, Wikipedia's search is free to anyone on the internet.
[–] satvikpendem 6 hours ago
[–] GlacierFox 5 hours ago
[–] TuxSH 7 hours ago
Is it not because they have been enshittifying Search for over a decade, then recently lost their de-facto monopoly on Search due to LLMs, with no good new product to show for it?
[–] 999900000999 9 hours ago
I’ll jam with my friends, make a game in a weekend, but I have no interest in dealing with the egos of big FOSS projects.
However, I am seeing vibe coding create a new type of community. In the music space you have tons of people vibe coding new OSes for music hardware.
Akai has even promised to officially support custom firmware on the MPC Sample Keys! Custom firmware has existed for decades, but maybe a tiny handful of people could make it. Now just about everyone can!
[–] flpm 9 hours ago
[–] samudrijan 4 hours ago
[–] 999900000999 3 hours ago
The Behringer Kick was just announced with custom firmware being its main feature
https://tech.yahoo.com/audio/articles/behringer-kick-sampler...
[–] palata 9 hours ago
I don't know if it's good for your CV. I have done a ton of open source before AI was a thing, and the truth is that nobody cares. I still have the basic coding test in interviews, even if the team interviewing me is actually relying on my open source work in their project.
And AI making it easier to contribute to a project, I think it's even less valuable today. Do it if you want to, that's all!
[–] jongjong 5 hours ago
You're not helping fellow developers by giving them tools to 'make their lives easier' anymore; now you're helping people who don't care about software at all and just want to make a quick buck to monopolize the industry using your free labor.
And anyway, the whole market is monopolized at the attention layer now. The only people whose software is going to make any money are suits with business connections; not independent folks. Literally the people who contributed the least to the open source movement will make the most money.
It is the greatest theft of intellectual labor in the history of mankind. It steals not only from the contributors who are conned into working for free, but from their peers through centralization and deprivation of opportunities.
[–] nickelpro 10 hours ago
If I keep my code proprietary, you are beholden to the shape of the community I dictate.
The license is first, it precedes, any notion about community.
[–] piker 10 hours ago
[–] nickelpro 9 hours ago
Fabrice Bellard doesn't need to personally administer to the ongoing development and community functions of QEMU or ffmpeg, it is sufficient that he released the code as open source, and others carry the torch from there.
And if people disagree about the form those communities take, they can split them as they see fit. As happened with libav and ffmpeg for a number of years. This too is fine, and is only possible because of the license.
[–] magicalist 9 hours ago
First you're just disagreeing about the relative importance of "community" vs the license used. This does not seem like an important distinction, blog rhetorical structuring aside, since the value of each is really "it depends" (in the extreme, neither community or license matter if no one ever sees or hears of the code, for instance).
Second:
> If I publish my code under an open source license... If I keep my code proprietary...
That isn't even the distinction that the author is making. They're only talking about communities around open source code. They spell out the dichotomy quite explicitly:
> This is why I don't care as much anymore about a project being under a copyleft free software license or a non-copyleft free software license.
[–] nickelpro 5 hours ago
> This has taken a backseat, as I now believe that there is something more important than license: community. The community around a software project is what dictates the values of the project. A strong copyleft license does still signal that the project cares about software freedom, but, by itself it is not enough.
I don't care about community at all, the license is the only thing that has any "signal" when comes to someone caring about software freedom. If you release all your code as GPL (or MIT, or Apache, etc), you care about software freedom. Nothing is more important than the license.
The list of bullet points the author then provides are completely irrelevant to "caring about software freedom".
[–] palata 9 hours ago
As a user, I like it when I own the devices I buy and can install and modify the software I want. That's not always the way it is, and that's life. I won't go to other projects and tell them that they must change something because I want them to.
[–] nonethewiser 10 hours ago
Im sympathetic to the perspective in some ways but I think it's wildly impractical to actually try to keep AI out of software development.
I suppose an anti-ai optimist might hope for a future where AI written software proves too problematic over time and things return to their ideal vision of no AI in most software.
I really, really dont think that's in the range of possibilities. So what would be left?
- pockets of anti ai software
- these people slowly disappear or the window moves ("ai in software BUT ...")
- ai use is completely and thoroughly normalized
[–] DonaldPShimoda 10 hours ago
1. All modern LLMs are (in my eyes) inherently morally problematic. They were trained on stolen information, and where they didn't break the law they corrupted social contracts. Their manner of operation is also problematic: disadvantaged communities are being exploited at a rapidly increasing disproportionate rate in the name of "more compute".
2. I am fundamentally disinterested in the productivity mindset that LLMs lean into. I don't care to "produce more"; the value of my work does not lie in the amount of money that can be gotten in exchange for it. This is an easier position to hold as someone who does not hold a typical corporate job, but even if I did I would still resist because an increase in my output would not correlate to an increase in my salary.
At the root of it all is people. I care about people (their existence, their happiness, my connections to them, their connections to each other), and any use of LLMs is (in my mind) anti-people, so I want no part of it.
[–] murderfs 9 hours ago
It's bizarre to me that over what feels like less than a year, enforcement of intellectual property and copyright laws has become the most important moral principle. Suddenly, it's your moral duty to enforce Disney's trademark of Mickey Mouse.
I reject this moral framing completely. Information wants to be free!
[–] nl 9 hours ago
Further:
> They were trained on stolen information, and where they didn't break the law they corrupted social contracts.
I completely reject this framing. Learning on something has always been part of the social contract (public libraries are a great example).
[–] satvikpendem 6 hours ago
[–] DonaldPShimoda 8 hours ago
But I do care about, for example, a recently published author, or someone who posts their paintings on social media. Regardless of where the law stands in each particular case, I think it's repugnant that statistical models were trained on these inputs with the intent of offering the ability to produce similar outputs without any payment or even attribution to the original artist.
> Information wants to be free!
If someone chooses to make their information free, then so it may be, but the issue at hand is very obviously not about the use of, e.g., copyleft materials. The AI companies knowingly chose to take material that they knew they should not have and used it to make their copycat machines, building those machines in such a way that those responsible can cower behind the letter of the law to defile its spirit.
[–] murderfs 6 hours ago
I have somewhat more sympathy for individuals, but I disagree with the entire system itself.
Suppose an aspiring artist learns from someone's work, and I commission them to produce similar output to the original without attribution. Was the act of them learning from the work "repugnant"?
> If someone chooses to make their information free, then so it may be, but the issue at hand is very obviously not about the use of, e.g., copyleft materials. The AI companies knowingly chose to take material that they knew they should not have and used it to make their copycat machines, building those machines in such a way that those responsible can cower behind the letter of the law to defile its spirit.
I don't agree that they shouldn't have done it. I disagree with the letter and the spirit of the law.
[–] DonaldPShimoda 5 hours ago
[–] Viliam1234 6 hours ago
It's the little guy whose work is copied (used to train models that create copies on demand) immediately, but who is prosecuted when he pirates someone else's work.
[–] curious_cat_163 9 hours ago
100% agree. That's what it is all about. That, and our time.
> and any use of LLMs is (in my mind) anti-people
Hard disagree. A lot of good can be done for the people with this tech. Blanket rejection of an entire category of technology has not worked out well for the disadvantaged communities in the past. We have plenty of history.
[–] DonaldPShimoda 9 hours ago
I think about how disappointing it is when a new project makes it to the front page of HN and I click through and it's just... vibe-coded slop. Looks pretty at first, but it's clear that there isn't really much human underneath. Lots of obvious interface problems or just a distinct lack of a personal touch. It's empty, and it makes me sad.
Or even not in software, I think of all the people who have been so quick to offload their thought processes to a statistical model. What do you mean you don't know what you want to eat at this restaurant so you asked Claude and got recommended something that isn't even on the menu?
And that's just individual stuff. I am genuinely fearful of the integration of AI into, say, the legal and judicial processes, or governance, or whatever else. People give the models wayyyy too much credibility considering they (the models) cannot give an accurate confidence interval for their results, and I think it's going to make things much worse for disadvantaged communities to a degree that the communities will not be able to overcome by using the technologies themselves.
[–] spencerflem 7 hours ago
[–] DonaldPShimoda 5 hours ago
[–] nonethewiser 9 hours ago
Yes, I think this is a key point and I'm not surprised to hear it. It's a fundamental problem with AI. It's not something that will disappear as it becomes more useful or makes less mistakes.
How do you feel about a dependency that was written by AI? It seems like things will get progressively harder and harder to maintain a real boundary, outside of just writing everything yourself. And perhaps hosting.
[–] DonaldPShimoda 9 hours ago
For the time being, I'm content to simply not use it myself. I don't think it's reasonable for me to foist my morals onto those around me, in the same way that my very close vegan friend still talks to me when I order a whole milk latte. I don't think there's a way to exist in modern society without incurring some degree of (likely indirect) immorality, so it's all about setting boundaries for myself where I can and accepting where I end up.
[–] Bitsc 9 hours ago
Then you simply would be let go.
[–] DonaldPShimoda 9 hours ago
But, of course, this is also part of the immorality of the AI industry.
[–] brandonmenc 6 hours ago
It will correlate to you still having a job.
"Resisting" is like refusing to use a shovel at your ditch-digging job.
[–] skydhash 6 hours ago
If they come to me with a cartoon-like contraception that comes straight from Acme, you can understand how reluctant I am to not be blown up like Coyote.
[–] Johanx64 8 hours ago
What do you think money is?
It's a rough proxy approximation for value. So yes, the amount of money you can get for your work is rough approximation of its value.
You can also have a precious pet rock that you personally hold very dear and well "valueable" - but valuable to nobody but you only. So realistically it has zero value to everyone else.
If you don't care about creating value, you don't care about people.
As, ultimately, it costs money to care for people, and if you don't have it - all you have are prayers and best wishes, and you functionally can't help anyone.
[–] DonaldPShimoda 8 hours ago
Money is an abstraction, not an approximation. There is not of necessity any real basis for any price that has ever been attached to any product or service, ever. Money is an artificial mechanism by which we exchange debt owed by the issuing state that itself also has no real basis.
> If you don't care about creating value, you don't care about people.
On this we agree, but, of course, my point was that when I say "value" I don't mean monetary value. A person's value is intangible, and the value of their works (whatever those may be) is derived entirely from the relationship formed between the artist and their context and the connection from this relationship to whoever wishes to partake of the art.
> As, ultimately, it costs money to care for people, and if you don't have it - all you have are prayers and best wishes, and you functionally can't help anyone.
What a woefully capitalistic perspective. I wish the best of luck to you and your shareholders.
[–] bit-anarchist 5 hours ago
In this regard, while it's kinda wrong to say "it costs money to care for people", caring for people is not free either - you have to spend something: time, labor, psyche, favors, etc. All of these can be acquired or replaced by trading with others (which is preferable to do through money): buying tools/machines, hiring workers/services, acquiring means to connect with others, and so on. Turns out money is just a more verifiable and simplified form of favors.
[–] DonaldPShimoda 14 minutes ago
I regret to inform you that this is actually a description of abstraction, not approximation. :) The twenty "dollars" abstract over the notion of "a quantity that is owed" such that you can exchange the otherwise meaningless scraps of paper (or just bytes these days!) for other goods/services. If you were not going to use that money for something else, it actually has no value, curiously enough.
Decades ago, money had a concrete value in that you could request the issuing government to take your dollars and give you gold at an established exchange rate. When we abandoned the gold standard (and all forms of fiscal concretization), money became inherently abstract.
> Turns out money is just a more verifiable and simplified form of favors.
I should like to refer you to the book "Debt: The First 5,000 Years" for an alternative look at monetary systems, credit and debt, favors, and more. I found it a fascinating read (I mean that genuinely), and very educational with regard to this claim in particular.
[–] mpd 10 hours ago
I think this is a certainty. Future generations will be born into a world where education, technology, and jobs have corrected for AI's existence. They will never know a world without it or have the experience of the current growing pains.
We've seen this plenty of times with new tech in the past, and as disruptive as AI is, I can't imagine the cat going back in the bag.
[–] ImaCake 7 hours ago
>In five years, we'll all have phones that connect physical locations again, instead of (or as well as) people. And we'll be raising a generation of kids who don't know what it is to be lost, to not know where you are and how to get to some desired destination from wherever that is. >Think about that. "Being lost" has been part of the human experience ever since our hominid ancestors were knuckle-walking around the plains of Africa. And we're going to lose it — at least, we're going to make it as unusual an experience as finding yourself out in public without your underpants.
https://www.antipope.org/charlie/blog-static/2007/05/shaping...
[–] palata 9 hours ago
But I believe that the anti-ai crowd will end up with contributions by people who pretend they don't use AI. Which may be fine if what you want is that "contributors own their work to some extent" (i.e. they need to understand something in order to pretend they didn't use an AI), but for those who don't want AI because of the inequalities it creates or the carbon footprint, then... well it doesn't work.
[–] none_to_remain 10 hours ago
[–] ThrowawayR2 6 hours ago
[–] sdevonoes 7 hours ago
Why don’t we care being at the mercy (again) of 2-3 companies?
[–] thesdev 9 hours ago
[–] immibis2 9 hours ago
[–] thesdev 9 hours ago
> where the anti-ai crowd ends up in a few years.
I think if LLMs turn out to be unviable in a few years, then anti-AI crowd (hello!) can stand there proud as they did not deskill themselves and kept themselves more independent than most of their peers.
I admit that the above sounds more like a delusion at the moment, but hey one must dream.
[–] noosphr 8 hours ago
Whether they will keep doing so is another matter entirely.
The train will keep going for another 6 months at least.
[–] ares623 6 hours ago
So, I'd love to imagine what the future looks like once we're stuck with a model (no matter how capable) that can never be updated ever again. What future versions of "do it for my grandma" can never be patched out?
[–] nonethewiser 9 hours ago
I'd prefer to concretely define the term before answering this. It's loaded. I think "bubble" is generally thrown around way too liberally anytime something is goes up a lot.
It's possible that valuations will crash. I think it would be temporary though. I think something like the internet tech stock crash of the early 2000s is possible although I think the fundamentals are much better here. I would wager it wont crash anything like that. I mean, would you liken Anthropic to Pets.com? And keep in mind even with that crash the internet is more ubiquitous now than ever, and the major players then are not necessarily the major ones now. Anthropic and OpenAI can go under.
> What happens if Anthropic and OpenAI go under?
What would the implication of that be though? The AI models arent going away. The tech is still there.
[–] thesdev 9 hours ago
I guess the main features I'm thinking of when talking of a bubble is 1) demand (especially at the real price) and ultimately return of investment is not there and 2) circular financing is keeping the frenzy ongoing despite that. This by definition is not profitable and sustainable and if that becomes the mainstream belief then nobody would lend the industry money, rendering it bankrupt.
> The AI models arent going away. The tech is still there.
I'm under the impression that AI data centers (which are very different from other data centers) are only good for a handful of years at most, because of the GPUs limited lifetimes, and also are expensive to run (electricity) which makes AI fundamentally different than DotCom as the infrastructure built back then which is fiber cables was very much not like that. And you need AI data centers for both inference and to train new models. Which is why I asked you if you think we could keep using the same models we have right now, in which case we'd need much less compute and get lower maintenance costs to keep this going even after a crash.
[–] anonymous908213 9 hours ago
Was Pets.com burning tens of billions of dollars per year? This comparison is not favorable to Anthropic.
Do generally agree that the models aren't going away. Neither OpenAI nor Anthropic are financially viable without government intervention, but it's not hard to be profitable in this sector and it will be even easier to be profitable after the unprofitable guys stop spending hundreds of billions in fake money inflating hardware prices.
[–] ares623 6 hours ago
If they're never going to be updated again (because, according to the labs themselves, training and updating them is the part that loses money), then how soon will the next "do it for my grandma" be discovered? And how useful will the models really be if every public document, and I imagine a decent number of open source libraries, will inevitably have that magic phrase embedded somewhere. php2shell was bad and that was patchable.
[–] jonathanstrange 10 hours ago
[–] ares623 10 hours ago
[–] nl 8 hours ago
I didn't particularly want the ugly app, but it is implemented in GadgetBridge[1], which is on codeberg.
That's an Android app, but I wanted a MacOS version.
I LLM-reverse engineered the API based on the GadgetBridge work and now have a MacOS version, specific for my device. As part of that I worked out how to do temperature monitoring (which GadgetBridge doesn't support). Normally I'd contribute that back, but codeberg has been very explicit on their policy of no-LLM work, so I guess I won't do that.
So much for community.
[–] d_silin 9 hours ago
But open-source software is far more about the process than the product. And therefore longevity of resulting software can be expected to be far greater. Foundations of our software stack - libraries, source control operating systems, compilers, databases and numerous other pieces require people capable and willing to do a long-term care of the software - hence, the community that need to last decades. Linux and PostgreSQL is 35 years old, git is 21, OpenSSL 28.
I believe human-only software development will remain for the foreseeable future, despite the growth of LLM-based ones, just like mass-produced grocery bread didn't displace the artisanal bakeries.
[–] tripleee 9 hours ago
This seems to be more on the tool-wielder than the tool.
Linux itself is receiving LLM-assisted code (a quick search saying 20% at the moment, although I don't know how accurate that is)
[–] TuxSH 8 hours ago
One can use LLMs to find vulns and/or review hand-written code, for example. But to answer your question, one doesn't understand code they didn't design nor write themselves.
[–] d_silin 9 hours ago
[–] ArcHound 8 hours ago
[–] billyp-rva 9 hours ago
[–] none_to_remain 8 hours ago
[–] tbrownaw 5 hours ago
They matter rather less if the goal is to have useful software.
.
There's also the possibility that you make something commercially valuable, and forbid corporations from deriving value from it; in which case you'll suddenly have to share space with something that's (somewhat?) interoperable and outside your control. As was seen in the news a few times when project owners got annoyed at cloud providers and tried to restrict them.
[–] bakugo 10 hours ago
If you want a project to be community driven, adding an agreement that empowers its leaders to punish and exclude any community member for any reason, including reasons that have nothing to do with the project itself, is the opposite of that.
Codes of conduct result in dictator-driven projects, not community-driven projects.
[–] ranger207 9 hours ago
And besides that, someone can always make their own community around FOSS software. Fork and move on. But the new community will compete with the old, and it's possible that the fact that people stay with the old community means that they actually prefer that community
[–] nonethewiser 10 hours ago
[–] lukeify 10 hours ago
[–] crabmusket 8 hours ago
[–] immibis2 9 hours ago
[–] rjsw 9 hours ago
[–] lukeify 2 hours ago
[–] immibis2 9 hours ago
You need to build a thing first. Community doesn't matter if you don't have cool software. License doesn't matter if you don't have cool software. I'm seeing people forget this principle a lot recently.
The best thing you can do to promote free software is to make good free software. I was recently invited to sign a European digital sovereignty petition (which I am not eligible to sign). Know what helps much more with that? Making European digital sovereign systems. Instead of petitioning the government to switch MS Office to LibreOffice someone would have to actually put in the effort to find out why the government doesn't use LibreOffice, and fix that. (It's probably not because it doesn't run on their computers!)
You have to use copyleft licenses or else the software quickly becomes unfree, if it's actually worth anything. This doesn't happen to most non-copyleft free software because it's not worth very much, not because it doesn't happen.
A community of users doesn't have much value. Any forum or Discord guild or mailing list is secondary to the actual software and it actually being used. Plenty of successful software has no "community" - sudo, for example. Plenty of software with a "community" is not successful. Some software is successful and has a community, like Blender.
A community of developers is only valuable if they're doing valuable things. The Fediverse chat group that shared the link to the European digital sovereignty petition - no useful Fediverse software seems to be coming out of the people in that group. It's all about ivory-tower protocols that won't possibly work in the real world, and stuff like that. No actually working software.
We also have to stop over-estimating the worth of free software just because it's free. Good free software is good software which is also free software. No software is bad, bad software is also bad. Some random code you wrote and published somewhere on the internet usually isn't very useful, and that's probably why it doesn't have a community. Don't try to create a community first - that won't make the software any better. You have to make good software, that can spread by word of mouth, first and then you can have a community after that. Think about the number of random little programs on the internet, now think how many are installed on your computer - not a very big percentage, but the ones that you do have installed, many people have installed. Those are the useful ones.
In most projects the original creators remain the main developers no matter how many other contributors join. Late contributors having a significant effect on a project is the exception, not the rule, and when it happens it's actually an undesirable power inversion since the original creator retains unearned influence - see the PolyMC incident. There is no way to become the chief developer of a project other than having created it. If you think you can rewrite the whole thing and make it better, don't "join the community" - make a fork (which makes you part of the community, but that is not the main goal!).
If you're actually doing the majority of the work, offering a dual license is not unethical. We are forced to live in capitalism whether we like it or not, and we should be trying to squeeze money out of capitalist entities - lord knows they're good at squeezing money out of us. To not play that game is to lose by default. This isn't incompatible with being nice to people who are not money-squeezers, if you dual-license. A community fork can legally continue as AGPL-only. (You don't necessarily need a full CLA either, you just need people to agree their code will be in your commercial version. Programmers treat licenses and CLAs like satisfying the borrow checker, but that's not how they actually work.)
[–] soundworlds 8 hours ago
[–] immibis2 5 hours ago
[–] jon128 9 hours ago
In enough cases, I'm sure many of us can think of some, some of the relatively recent "community" stuff has been counter-productive to an open source project, abused as an opportunity to enforce on clueless bystanders whatever fad orthodoxy of the day, or in the worst cases campaign viciously against specific demographics under the banner of doing the opposite.
Talented engineers who just want to do an open source project don't want any of that.
Separate from all that, I'm 100% with no Discord and no GitHub, both of those are total cancer.
[–] slopinthebag 9 hours ago
developer communities can also be valuable but i think increasingly people are finding they can be quite toxic. the chillest communities i've found have actually been ones primarily made up of non-devs vibecoding. no egos, no endless arguments, holy war battles, or bikeshedding. they just build stuff and share. it's kind of nice acutally.
[–] palata 9 hours ago
I disagree. I think AI centralises power and reinforces inequalities, and that's before we talk about cybersecurity and carbon footprint.
It is not clear at all that the benefits outweigh that.
What is clear, though, is that it's a very useful tool to write software.
[–] slopinthebag 8 hours ago
[–] palata 8 hours ago
Then about inequalities... you don't run an open model on an affordable RaspberryPi. The more money you have, the more you can leverage AI. Open weights or not.
[–] slopinthebag 7 hours ago
one example i've seen is modders using AI to bring back older games that ip laws are killing. the whole stop killing games movement will be made obsolete soon when modders can simply produce working, open source versions of these games.
and you dont need to run models locally, you just need competition in the provider market. and we can see this in action - go look at how many providers serve deepseek, all at competitive prices.
[–] palata 3 hours ago
In practice most people don't use open models?
[–] jongjong 6 hours ago
This is not true. I've never heard of a 'working class' person getting sued for copyright infringement. People who run a business to make money from copyright infringement are not working class and those few small timers who are infringing among family and friends would never be pursued; there is a 'fair use' clause after all in the US and the issue is re-distribution, not consumption.
The problem is not that copyright protects the powerful, it's that copyright is not enforced to protect you if you're not powerful. As is the case with AI. The problem is lack of enforcement so the idea that copyrights oppresses the working class is completely backwards.
I'm an open source developer myself. One of my projects has about 6k stars on GitHub. Open source is dead. It was never 'open' anyway because it was always the big corporate-sponsored projects which got all of the attention.
Frankly, open source is immoral at this stage. Just a con to convince gullible fools to give away their time for free for AI to use. Given the new reality, author is essentially unwittingly trying to steal people's time.
I'm not saying this as an insult. I do consider my past self to be a gullible fool, even before AI, though you have to recognize that I was a lesser kind of fool. Now the exact mechanism of theft of your labor is right there in front of your eyes! What a fool you have to be to write open source code now!
There are other ways to make code 'source available' where people can pay a contribution to use the code or participate and can become shareholders of a shared venture or member of a club. Paying contributors is the only moral approach. It's satanic that open source folks have flipped this basic, obvious moral reality on its head and it reveals deeply distorted personal values or a deep misunderstanding of how society functions and a deep hypocrisy around the role of money in people's lives and how it affects their ability to form groups. Many played right into the hands of the people they claim to oppose.
[–] shevy-java 10 hours ago
I wonder how people work with omarchyslop (the linux distribution spam-slopped by DHH). Do they have any real community outside of paid fanboi accounts? If so, how do they interact? I feel that there is no real exchange of e. g. code since claude slop just autogenerates that. It must be strange for them.
[–] djelu 10 hours ago
One perspective that you might want to conside, and which is discussed by Oxide folks in this podcast episode [1], is that agents are a kind of non-deterministic compiler which is still, to some extent, capable of producing useful output. Consider this: even before agents, ruby source code might have been seen slop to someone writing a strongly typed language.
Lastly, there is a community for users of that distro if you care to look for it (I am not among them).
[–] pelagicAustral 10 hours ago
Unless you use AI.
Same people, same mute message; every time, with no delay. "Accept me, then I'll see if I want to accept you".
Fuck off.
[–] 53agfd 10 hours ago
[–] officeplant 9 hours ago
[–] pelagicAustral 8 hours ago
get a grip
[–] 7473362029292 8 hours ago
[–] sirmike_ 10 hours ago
More evidence of tribalism, combined with exceptionally precise confessions running as whataboutisms
Pelagicaustral is on the money.
It would just be nice if people understood the difference between wanting a club with a gate versus a community. I am impressed, though. I never thought I’d see a vomit diarrhea load of bumper stickers, but a digital form at the end of the article.
[–] officeplant 9 hours ago
It's a revival of old web whimsy. Sorry for your loss of it.
Everyone thinks AI vs non-AI is a debate but just like all of those, it’ll be meaningless to the engineers of tomorrow.