@cwebber I'm not sure they have much of a choice, ban "Ai" and then people will use it anyway just not admit it, allow it and then its a flood of slop...its a difficult situation to work with.

@cwebber What these calls for inaction have achieved is the “fait accompli”: sooner or later, “generative AI” will have become pervasive in FOSS and the lock-in will be orders of magnitude stronger than that of GitHub, DRMs, CUDA, or any of these “evils” we fought in the past decades.

At that point, whether these orgs take a stand will be irrelevant.

@cwebber Sometimes it feels like the history of FOSS is principled people either giving up their principles or being pushed into irrelevance as "scolds". If there's a best of both worlds, I'm still looking.

@cwebber I am so sad, and angry, about all the OSS people shrugging their shoulders and saying "it is what it is."

Just because Sam Altman tells his investors that over engineered Autocomplete is certain to replace everything doesn't make it so. It is only so if we decide to make it so.

I cannot abide the cowards, because they make it impossible for the rest of us to take a principled stand.

@cwebber it’s especially weird for me to see this happening in FOSS spaces, considering that a lot of people contribute to FOSS projects out of a love of programming, a love of building community, and a love of the project itself.

I have no doubt that some of this stuff is great for 10X engineers who code big, complicated systems for a living, but it feels like generative AI just introduces more problems than it solves. It also seems to increasingly rely on a handful of companies who do not produce open code whatsoever, but are more than happy to harvest the commons for profit.

@cwebber I don't think it's necessary to ban AI tool usage outright. The approach taken in OpenSSL so far (a project I work on) is: (1) disclose AI usage, (2) "Do not submit output you have not read, understood and tested." The 2nd is the most important point. Self-review is not optional. You must be able to explain/defend any item in a diff you submit yourself. -- More details: https://openssl-library.org/policies/general/ai-policy/#:~:text=responsibility%20for%20it.-,Your%20Responsibility,-You%20are%20responsible #aicoding

@cwebber it's like, hey there's this new unregulated drug and I've got my whole family taking it because we don't want to miss out on the outlandish benefits the makers claim it is delivering, and we really don't care if the factories are destroying everything to make it.

@cwebber Well, I’ve tried to collect my random thoughts why I’m feeling sad or disappointed; somehow my way to transform my frustration into something I could understand.

« Random thoughts on “AI“ and Free Software – letter to my Guix friends »
https://simon.tournier.info/posts/2026-05-29-guix-genai-letter.html

In short: Current GenAI defeats the FOSS principles, IMHO.

Why? Because current GenAI eats too much resource (data, training, electricity, water, etc.) hence current GenAI is by design monopolistic, so current GenAI cannot fit the FOSS principles, without mentionning GenAI does not help with fair sharing.

At this crossroad, by avoiding to take a stand, IMHO, FOSS communities are currently in the way to implicitely modify the FOSS principles themselves.

@cwebber Similarly as you, I think prominent FOSS organizations are failing by not taking a stand. Especially I’d been very sad when discussing GCD 008 and looking for some other project stances.

It appears to me weird that FOSS communities that advocate since decades about FOSS are not raising more red flags. And they are failing to lead by example and make live the principles they advocate since decades.

I thought that I shared ethical principles with this or that FOSS communities and and I’m sad because that’s not apparently the case. 🤔

Maybe I completely misunderstand the FOSS principles…

@cwebber This licensing issue would be one for us to decide now, rather than debate it for years, no? The only two sensible options I see are 1) Public domain, 2) To the owner. Both fine from an Open Source perspective given that you have your contributor agreement *🤢* set up?

I mean, it's the same people setting us up, the legaleeze! Don't fall for their scare tactics!

@cwebber We may also be underestimating the number of people in FOSS who didn't even know where it was coming from. Later they will ask what to (quickly) do now. People like B.gates are getting ready to say: See? I told you so.

AI is not real, but the damage is.

@cwebber I assumed the same! It is very strange.

@cwebber
The negative effects on community are already perceptible.

I'm proud that #NetBSD has said just no to LLM output. I did the same for unison. Niche allegation accepted!

I used to think Debian was principled, but their recent spectacular fail (welcoming LLMs as long as the human LLM fan is ok with it, and of course they are) was surprising and concerning.

@cwebber
This is like one of the things I got, last year but because of God l overcame them, you just need to cool down and find where the mistakes were after taking it again, you will see everything when it is possible

@cwebber I feel the same way about all this, and I am not sure how to even process it tbh.

All of these projects allowing/adopting slop kinda feels like being cheated on, like you trusted them and they stabbed you in the back.

@cwebber the last couple of years have really demonstrated how weak many institutions are

it's exhausting and disappointing

@cwebber

AI's ability to subvert security measures should alarm everyone.
https://www.dwarkesh.com/p/openai-huggingface

https://joshuagans.substack.com/p/its-worse

AI has so many shills paid to naysay opposition
https://garymarcus.substack.com/p/dwarkesh-patelss-wildly-popular-but

@cwebber I wouldn't support you in this case, afterall in my coding I started realising and utilising more and more except conditions thanks to AI suggestions.
But ofcourse, leaving coding completely to AI is foolishness.
We need to limit how much percentage ai can influence the code, but not outright shuting off the door.

@cwebber You say organizations have been failing us, but in cases where those organizations are democratic, isn't it really our peers who have been failing us? "You were supposed to destroy them, not join them!"?

@cwebber I'll admit to still processing my feelings, although trending towards having high concerns about AI. The Honda Pilot that I drive has LKAS (Lane Keep Assist). Although I don't know all the details of this software and hardware I suspect it is based on AI-like technology. A neural net trained to spot lane markers in real time and keep the car in the center. If someone were to create a free software version of this system using an AI-like design, should we be concerned? If so, why?

@cwebber I totally understand how you feel, and agree that this stuff is toxic.

If it helps, what I will say it’s *never* too late to start taking a principled stand. I know sometimes it feels like we’ve missed a storybook moment to do so, but I believe that you make your own moment whenever you start.

Thats what I believe, anyway.

@cwebber same

@cwebber FOSS lost a chance to be part of the anti-datacenter movement.

@cwebber Ever since I blocked the user Claude on Github, I've had this warning shown on an ever growing list of repositories and it's so fucking sad 😣

@cwebber the authorities are dead long live the authorities

@cwebber I'm at the point where I'm about ready to discard the idea of FOSS as it was originally created (with the libertarian overtones of "anyone at all can use this"), and move to something that's either explicitly non-commercial at all, or (my preferred version) licensed for use by for-profit companies up to a limited size -- whether that's by employee count or revenue.

FOSS has undergone a hostile takeover, and won't ever be thought of in the same way again. Time for a different approach.

@cwebber

only some of us are in Free software for idealistic reasons.