👀 … https://sfconservancy.org/blog/2026/apr/15/eternal-november-generative-ai-llm/ …my colleague Denver Gingerich writes: newcomers' extensive reliance on LLM-backed generative AI is comparable to the Eternal September onslaught to USENET in 1993.
-
@mathieui
I agree FOSS projects should make their own policies. Some will (& should!) have a zero-tolerance abstinence policy on any contribution that has even been slightly interacted with any LLM-backed generative AI systems.
Yet, even among SFC projects, some asked us to help them create a more nuanced policy.
Should we just kick those projects out of SFC, or have a nuanced, humans-only conversation?
It's ok if you do not want to join that, but we'd also be glad to have you.
Cc: @tito @ossguy@bkuhn @tito @ossguy I understand the need and do not intend to throw stones at the SFC here at all, I have diverging ethical considerations and am way too tired of it all (particularly writing non-FOSS software at work in the current LLM-crazed atmosphere) to even think about joining an oral conversation about it, in a language I am somewhat fluent but not articulate in.
I'm all for welcoming volunteers who want to do work on FOSS projects, but that means onboarding and doing actual work; if I wanted to run Claude on my code to do stuff, I don't need other people to do that, so what would be the point of recruiting volunteers ?
-
@bkuhn @tito @ossguy I understand the need and do not intend to throw stones at the SFC here at all, I have diverging ethical considerations and am way too tired of it all (particularly writing non-FOSS software at work in the current LLM-crazed atmosphere) to even think about joining an oral conversation about it, in a language I am somewhat fluent but not articulate in.
I'm all for welcoming volunteers who want to do work on FOSS projects, but that means onboarding and doing actual work; if I wanted to run Claude on my code to do stuff, I don't need other people to do that, so what would be the point of recruiting volunteers ?
Actually, I'm absolutely 🤮y re: talking about LLM-backed generative AI too! I've been talking about it for 4 years now
:
https://sfconservancy.org/blog/2022/feb/03/github-copilot-copyleft-gpl/But, I'm senior policy wonk in FOSS, & it's my day job. Everyone has crap they gotta do in their day job that isn't their favorite, & this is mine.
Speaking of bad stuff at day jobs: many people's day jobs MANDATE LLM-backed AI usage. Such a mandate is definitely immoral; it should always be the developers' choice.
-
@bkuhn @ossguy The surprising thing about saying "seriously consider cautiously and carefully incorporating their workflows with ours" is that it doesn't address at all my *biggest* fear: the copyright status of LLM generated contributions seems currently unsettled.
I know there's been assertions to the contrary floating around: the Supreme Court deferred to a lower court in the US. However that is not the same thing as the Supreme Court making a specific decision. And internationally, the copyright situation of output is even murkier... it will take a long time for this to settle.
Does Conservancy not think this is the case? I would be surprised if so, but perhaps you all have an interpretation that I am not currently aware of.
If there *is* concern, then we hit a serious risk: we may be seeing many contributions with legal status which has *yet to be determined* entering seasoned codebases. And this worries me a lot.
-
@richardfontana @bkuhn @ossguy In which of the 5 million ways I could parse that sentence do you mean it?
-
@bkuhn @mathieui @tito @ossguy i think it is good to have a nuanced conversation, but still be stern in that this unethical technology will not be allowed. the ethical issues of it are just too big, it would almost be as bad as allowing proprietary software in, i would say
education is important, and it is important to first educate and give some time before making a decision, but still be stern about it, as this is a deep ethical issue where we should be having a zero-tolerance
zero-tolerance here would mean not allowing the project to endorse or use any genai. if usage of it is snuck in, try and revert it to the best ability possible. if it was used before, do the same. but having some genai commits in is not that important, to me
of course, mistakes may be made. we should not be scrutinizing commits very heavily and going on witch hunts. but genai usage, for code, assets, writing, docs and anything else, must not be allowed
what's important to me is the stance of the project going forward. to be against it completelyThese are good ideas. I hope you can come to one of the chats and share them, but I've bookmarked your post so I am sure the ideas get considered.
-
First, I speak for myself, not SFC on this account. I work for SFC, but my words are not SFC's words by default. I *often* am unable to convince SFC to take policies or positions that I want.
By nuanced, I mean avoiding two sides showing up like it's a protest where one site shouts "NO AI" and the other side shouts "ALL AI ALL THE TIME" won't get us anywhere at all.
I'm very close to the "NO AI" side, but I'm a few steps toward other direction.
-
Normally, when someone is invited to a real-time public forum and would prefer to submit written comments, they would ask how to do so rather than being sarcastic and cruel. How odd.
In seriousness, in parallel @ossguy added info on just what you're asking for at the top of https://sfconservancy.org/blog/2026/apr/15/eternal-november-generative-ai-llm/
-
@richardfontana @bkuhn @ossguy In which of the 5 million ways I could parse that sentence do you mean it?
@cwebber May I please introduce you to the cryptic @richardfontana oracle.
He's often right and predicts the future well, but figuring out what he means is the riddle.
I've had this moment with @richardfontana more than a dozen times at least.
-
@josh @silverwizard @ossguy @bkuhn @karen @wwahammy
I can understand having an absolutist position against LLMs. I find that most arguments are either irrelevant to me or directly map to existing arguments about late-stage capitalism. So for me, there's nothing novel to object to about LLMs.
So with that in mind, I find "all contributions derived from LLMs should be rejected" to be misguided. I look at things like the bug fixes coming out of CodeMender (back in Feb, which is an LLM lifetime ago), and I am a huge fan. Fixing stuff found by a fuzzer:
https://issues.oss-fuzz.com/issues/486561029It's a small example, but it's an area that humans alone have not been able to remotely keep up with. (There are hundreds of open syzkaller bug reports, for example.) Gaining tools that will help with this is a big deal, and I'm glad for the assist.
-
@glitzersachen @josh @silverwizard @ossguy @bkuhn @karen @wwahammy @xgranade
I consider the cognition impairment hazards to overlap with the existing manipulation/critical-thinking hazards that capitalism depends on, with advertising being probably the most dangerous example (both explicit and implicit manipulation of many cognitive systems: confidence, selection, recency, etc etc).
IMHO LLMs are "just" a subset/extension of this existing problem. And I categorize it there because I think the defenses against their negative impacts are very similar.
-
@glitzersachen @josh @silverwizard @ossguy @bkuhn @karen @wwahammy @xgranade
I consider the cognition impairment hazards to overlap with the existing manipulation/critical-thinking hazards that capitalism depends on, with advertising being probably the most dangerous example (both explicit and implicit manipulation of many cognitive systems: confidence, selection, recency, etc etc).
IMHO LLMs are "just" a subset/extension of this existing problem. And I categorize it there because I think the defenses against their negative impacts are very similar.
@kees @josh @silverwizard @ossguy @bkuhn @karen @wwahammy @xgranade
> to overlap with the existing manipulation/critical-thinking hazards that capitalism
I think it's more, not only the manipulation part. LLMs actively corrode skills of the users. Not by by not using them. No, actually worse.
I hope you have heard about this possibility (whether you believe in it or not).
-
@glitzersachen @josh @silverwizard @ossguy @bkuhn @karen @wwahammy @xgranade
I consider the cognition impairment hazards to overlap with the existing manipulation/critical-thinking hazards that capitalism depends on, with advertising being probably the most dangerous example (both explicit and implicit manipulation of many cognitive systems: confidence, selection, recency, etc etc).
IMHO LLMs are "just" a subset/extension of this existing problem. And I categorize it there because I think the defenses against their negative impacts are very similar.
@kees @glitzersachen @josh @silverwizard @ossguy @bkuhn @karen @xgranade I think you are wildly underestimating the cognitive hazards. Like I hesitate to even say "wildly underestimating" because that phrase is not strong enough.
-
@kees @glitzersachen @josh @silverwizard @ossguy @bkuhn @karen @xgranade I think you are wildly underestimating the cognitive hazards. Like I hesitate to even say "wildly underestimating" because that phrase is not strong enough.
@wwahammy @kees IMO you're both right.
LLM-backed gen. AI is a dangerous tool w/ potential to not only atrophy the skillsets of experienced developers *but also* lead newcomers to *never develop those skills*.
Our charge is to create policies that encourage extremely disciplined use of these systems.
I support decriminalization of recreational substances. But, such has to come with major funding for addiction support. IMO the analogy is apt.
@glitzersachen @josh @silverwizard @ossguy @xgranade -
@kees @josh @silverwizard @ossguy @bkuhn @karen @wwahammy @xgranade
> to overlap with the existing manipulation/critical-thinking hazards that capitalism
I think it's more, not only the manipulation part. LLMs actively corrode skills of the users. Not by by not using them. No, actually worse.
I hope you have heard about this possibility (whether you believe in it or not).
@glitzersachen @josh @silverwizard @ossguy @bkuhn @karen @wwahammy @xgranade
> LLMs actively corrode skills of the users
Yup, very aware. It's a specific instance of what I still see as a larger critical thinking erosion happening all around us.
-
-
@glitzersachen @josh @silverwizard @ossguy @bkuhn @karen @wwahammy @xgranade
> LLMs actively corrode skills of the users
Yup, very aware. It's a specific instance of what I still see as a larger critical thinking erosion happening all around us.
@kees @josh @silverwizard @ossguy @bkuhn @karen @wwahammy @xgranade
I suspect (as some scientists do as well) that this is not a cultural, but a neurological phenomenon. So I think this is really on a totally different level.
The skill erosion I am talking about has nothing to do AT ALL with critical thinking.
Best case it's just that if of two synaptic circuits (use the translation tool vs retrieve from memory) the one which wanted to activate, but then got not chosen, is actively deleted or weakened. My understanding is that this is how biological brains work / learn.
The worse alternative is that the output of LLMs has some yet not sufficiently described hidden quality which poisons neuronal networks that process them.
One hint in this direction is, that LLM models, who consume the production of other LLM models in training, collapse. That's a clear indicator that on some level LLM output is observably different from human language production, though we as humans have on average a hard time telling the difference.
And this is what I am talking about: Not a loss of cultural techniques or of learned skill by atrophy or not being taught anymore, but the poisoning of neuronal networks by input they cannot firewall because they have not evolved to recognize it as a hazard.
-
@wwahammy @kees IMO you're both right.
LLM-backed gen. AI is a dangerous tool w/ potential to not only atrophy the skillsets of experienced developers *but also* lead newcomers to *never develop those skills*.
Our charge is to create policies that encourage extremely disciplined use of these systems.
I support decriminalization of recreational substances. But, such has to come with major funding for addiction support. IMO the analogy is apt.
@glitzersachen @josh @silverwizard @ossguy @xgranade@bkuhn @wwahammy @kees @josh @silverwizard @ossguy @xgranade
Let me point you to my reply here => https://hachyderm.io/@glitzersachen/116421481982246037.
I really think the issue at the core *might* not be loosing skills by neglecting to exercise them, but rather poisoning of neural networks. Brainwashing them into (skill) oblivion.
The comparison to hard drugs would be apt, if this is true.
And our employers want us to ruin our skills and our brains. They obviously don't believe in a common future with their knowledge workers anymore...
-
@kees @glitzersachen @josh @silverwizard @ossguy @bkuhn @karen @xgranade I think you are wildly underestimating the cognitive hazards. Like I hesitate to even say "wildly underestimating" because that phrase is not strong enough.
@wwahammy @glitzersachen @josh @silverwizard @ossguy @bkuhn @karen @xgranade
Right; it is an extremely focused risk (differing from the larger varieties and sources of critical thinking erosion). And every piece of research I've seen with regard to "how to safely use LLMs in education" confirms this with bright flashing lights: there is none. LLMs appear to have a universally negative impact in education.
-
@wwahammy @kees IMO you're both right.
LLM-backed gen. AI is a dangerous tool w/ potential to not only atrophy the skillsets of experienced developers *but also* lead newcomers to *never develop those skills*.
Our charge is to create policies that encourage extremely disciplined use of these systems.
I support decriminalization of recreational substances. But, such has to come with major funding for addiction support. IMO the analogy is apt.
@glitzersachen @josh @silverwizard @ossguy @xgranade@bkuhn @kees @glitzersachen @josh @silverwizard @ossguy @xgranade
This is not a remotely accurate analogy. The level of rage in this country over AI is uncontrollable and it's accelerating. Two people tried to kill Sam Altman in the last week. An Indiana planning official's house was shot after they approved a new data center.
In the political realm, the shift is unimaginably swift. Ex: 6 months ago, no Democrat for WI governor had a policy on data centers because building unions wanted them. Now every one of them is fighting over how strict their ban on data centers is.
The best analogy I think of is the opioid crisis. When people were ready to kill the Sacklers and everyone at Purdue Pharma, you can't come in and say anything that people think you are tolerant of the damage. You can't even argue "we can punish these people but we have to protect access to opioids". Everyone KNOWS there are uses but you can't build a policy around that because the public doesn't care. At all.
The only time you can have this discussion was years ago or years in the future after the public has taken their pound of flesh. Right now, it's an immensely dangerous idea for SFC.
-
Talking with them is good. Helping to educate them is good. Making it sound as if what they are doing is okay is *not*.
There is a big difference between offering an olive branch to people who *might* be productive contributors in the *future*, and telling them that what they're doing *now* is okay.
The best AI policy remains "do not contribute any LLM-written content, ever". You have published a post that makes it easier for people who oppose such policies to cite your "olive branch" when arguing against it, and it is not obvious from your post that you do not want that to happen.
I don't want to see people *abused* for using LLMs. I do want them to understand that what they're doing is not okay and not welcome and not a positive contribution.@josh @silverwizard @ossguy @bkuhn @karen @wwahammy
As far as I am concerned people should be "abused" for shilling AI from a position where they really don't have any sufficient insight. Like middle management trying to push AI on reluctant software engineers with all the tricks in the book (for example tying performance review results to AI use). This behavior destroys trust and workplace culture. What do they think? That the engineers don't understand their own work mode? The hubris of management: "I'll tell you how you can work better. I know better how you can work better."
And this behavior needs to be called out.
