I've been wondering, Stallman was driven to create free software after an incident trying to get the code for firmware on his office printer. I'm wondering if today, would he have just reverse engineered it with AI?
Edit: I'm also thinking of what he did rewriting all of Symbolics code for LISP machines
I think one of the more prominent issues folks take with mass training on OSS is that the companies doing it are now profiting for having done it.
In his follow-up post he talks about him open sourcing old games as a gift, and he doesn't much care how people receive that gift, just that they do.
He doesn't acknowledge that Anthropic, OpenAI, etc, are profiting while the original authors are not.
The original authors most of the time didn't write the software to profit. But that doesn't mean they don't care if other people profit from their work.
> But that doesn't mean they don't care if other people profit from their work.
He clearly states his opinions. He doesn't care if other people profit from his code.
>> GPL would prevent outright exploitation by our competitors, but those were to allay fears of my partners to allow me to make the gift
He believes other members in OSS community should have this mindset. Of course it might not be fair, especially for members who are as financially fortunate as him. His point is clear nevertheless.
Isn't that permitted by some of the more popular licences? If you care about others profiting from your work you'd choose an appropriate licence. And then you'd temper your expectations and hope for the best because you know there will be less than perfect compliance. It's like lending money to family or friends. You can hope they pay you back, but better to consider it a gift because there's a good chance they won't.
Is it worse because it's AI for some reason? I'm having trouble pinning down exactly what the gripe is. Is it license compliance? Is it AI specific? Is it some notion about uncool behavior in what some people see as a community?
Yeah the main difference seems to be that he open sourced the games after he got very wealthy from them not before. So of course at that point you can easily feel magnanimous about bestowing gifts.
Open sourcing something from the start and essentially giving up any ability to profit from the use of your work when companies are often making huge profits from it seems less easy in comparison.
As I understand it, the anti-AI stance of open source software people in particular has nothing to do with AI learning from code bases, and everything to do with AI slop clogging all unrestricted community feedback channels.
There is code I gift to the world that I license as MIT or similar and there is code I publish as a means for furthering what I perceive as a advanced society which I license as GPL or similar.
I don't ask anyone to share my ideals but conflating these two is dishonest.
> - OSS is valuable for decentralizing power and influence
That was the intention and hope, but I think the past twenty years has shown that it largely had the opposite effect.
Let's say I write some useful library and open source it.
Joe Small Business Owner uses it in his application. It makes his app more useful and he makes an extra $100,000 from his 1,000 users.
Meanwhile Alice Giant Corporate CEO uses it in her application. It makes her app more useful by exactly the same amount, but because she has a million users, now she's a billion dollars richer.
If you assume that open source provides additive value, then giving it to everyone freely will generally have an equalizing effect. Those with the least existing wealth will find that additive value more impactful than someone who is already rich. Giving a poor person $10,000 can change their life. Give it to Jeff Bezos and it won't even change his dinner plans.
But if you consider that open source provides multiplicative value, then giving it to everyone is effectively a force multiplier for their existing power.
In practice, it's probably somewhere between the two. But when you consider how highly iterative systems are, even a slight multiplicative effect means that over time it's mostly enriching the rich.
Seven of the ten richest people in the world got there from tech [1]. If the goal of open source was to lead to less inequality, it's clearly not working, or at least not working well enough to counter other forces trending towards inequality.
Personally for me I don't see it as gift, he licensed out the engine but didn't want to be in the engine business, after selling enough it feels he just put it out there so it's his stamp forever with the GPL infection. I think he already felt the diminishing returns at the time. He knew about the sharing of floppy discs and hacker scene and eventually someone would've done it and I think he felt cornered and said fuck it might as well put it out there to beat them to it.
Surely we can all agree that there is a difference between:
- Sharing/working on something for free with the hopes that others like it and maybe co tribute back.
- Sharing something for free so that a giant corporation can make several trillion dollars and use my passion to train a machine for (including, but not limited to) drone striking a school.
> those were to allay fears of my partners to allow me to make the gift
I respect Carmack so much more now. I always scratched my head why he made Quake GPL. It was such a waste. Now it doesn't matter anymore. I so thankful copyleft is finally losing its teeth. It served its purpose 30 years ago, we don't need it anymore.
In my mind, AI is making a lot of engineers, including Carmack, seem fairly thoughtless. At the other moments in recent history where technology has displaced workers, labor has either had to fight some very bloody battles or had stronger labor organization. Tech workers are highly atomized now, and if you have to work to live, you're negotiating on your own.
It seems like Carmack, like a lot of tech people, have forgotten to ask the question: who stands to benefit if we devalue the US services economy broadly? Who stands to lose? It seems like a lot of these people are assuming AI will be a universal good. It is easy to feel that way when you are independently wealthy and won't feel the fallout.
Even a small % of layoffs of the US white collar work force will crash the economy, as our economy is extremely levered. This is what happened in 2008: like 7% of mortgages failed, and this caused a cascade of failures we are still feeling today.
Software engineers have been automating away workers' jobs from the beginning. "Computer" was once a job title. There were armies of switchboard operators at the phone company. Companies had typing pools, mail clerks, and file clerks. We write shell scripts and development tools to automate our own jobs.
The purist forms of capitalism I’ve seen are places with low prices, a large working class, practical marketing, and high competition - often they’re considered “3rd world” places.
The US economy, if it wants to remain “1st world” must have high prices. It has to contain an element of scarcity (however faux) in order to be sold at a premium, or be able to impart some privileged (institutional) knowledge as a firm - which should be as esoteric as it is scarce.
It can’t be quality alone since all building and manufacturing is effectively outsourced. It has to have a premium brand recognition or monopolistic aspects to it that necessitate a high price.
So the challenge for the first world, during the rise of China (Mexico, etc.), is to find new ways to justify the privileged position using this new technology as a lever to do so.
There's one elephant in the room that's not being addressed:
Training an AI on GPL code and then having it generate equivalent code that is released under a closed source license seems like a good way to destroy the copy-left FOSS ecosystem.
People were violating the terms of GPL without consequence long before AI. It is very difficult to determine if binaries were compiled from fragments of GPL code.
The places I have found AI most useful in coding is stripping away layers of abstraction. It is difficult to say as a long time open source contributor, but libraries often tried to cater to everyone and became slow, monolithic piles of abstraction. All the parts of an open source project that are copyrightable are abstraction. When you take away all the branching and make a script that performs all the side effects that some library would have produced for a specific set of args, you are left with something that is not novel. It’s quite liberating to stop fighting errors deep in some UVC driver, and just pull raw bytes from a USB device without a mountain of indirection from decades of irrelevant edge case handling.
Well, if Carmack wants to give gifts to AI companies then he's free to do it, but it doesn't mean that other people want it too.
I think this debate is mainly about the value of human labor. I guess when you're a millionaire, it's much easier to be excited about human labor losing value.
> and the GPL would prevent outright exploitation by our competitors, but those were to allay fears of my partners to allow me to make the gift.
I can understand his stance on AI given this perspective. I have a harder time empathizing his frustrations. Did he also have a hard time coming to terms with the need for AGPL?
Thinking of open source as a gift is such a strange take. It implies that the relationship is merely a transaction where the giftee is the beneficiary and the gifter is a philanthropist. It has subtle financial undertones, and a sense that gifters are somehow morally superior.
It is far healthier to see it as a collaboration. The author publishes the software with freedoms that allow anyone to not only use the software, but crucially to modify it and, hopefully, to publish their changes as well so that the entire community can benefit, not just the original author or those who modify it. It encourages people to not keep software to themselves, which is in great part the problem with proprietary software. Additionally, copyleft licenses ensure that those freedoms are propagated, so that malicious people don't abuse the system, i.e. avoiding the paradox of tolerance.
Far be it from me to question the wisdom of someone like Carmack, but he's not exactly an authority on open source. While id has released many of their games over the years, this is often a few years after the games are commercially relevant. I guess it makes sense that someone sees open source as a "gift" they give to the world after they've extracted the value they needed from it. I have little interest in what he has to say about "AI", as well.
Hey John, where can I find the open source projects released by your "AI" company?
Ah, there's physical_atari[1]. Somehow I doubt this is the next industry breakthrough, but I won't look a gift horse in the mouth.
115 comments
[ 3.6 ms ] story [ 92.5 ms ] threadEdit: I'm also thinking of what he did rewriting all of Symbolics code for LISP machines
(similar to the person that accidentally hacked all vacuum of a certain manufacturer trying to gain access to his robot vacuum? https://www.theguardian.com/lifeandstyle/2026/feb/24/acciden...)
In his follow-up post he talks about him open sourcing old games as a gift, and he doesn't much care how people receive that gift, just that they do.
He doesn't acknowledge that Anthropic, OpenAI, etc, are profiting while the original authors are not.
The original authors most of the time didn't write the software to profit. But that doesn't mean they don't care if other people profit from their work.
It's odd to me that he doesn't acknowledge this.
He clearly states his opinions. He doesn't care if other people profit from his code.
>> GPL would prevent outright exploitation by our competitors, but those were to allay fears of my partners to allow me to make the gift
He believes other members in OSS community should have this mindset. Of course it might not be fair, especially for members who are as financially fortunate as him. His point is clear nevertheless.
Isn't that permitted by some of the more popular licences? If you care about others profiting from your work you'd choose an appropriate licence. And then you'd temper your expectations and hope for the best because you know there will be less than perfect compliance. It's like lending money to family or friends. You can hope they pay you back, but better to consider it a gift because there's a good chance they won't.
Is it worse because it's AI for some reason? I'm having trouble pinning down exactly what the gripe is. Is it license compliance? Is it AI specific? Is it some notion about uncool behavior in what some people see as a community?
Open sourcing something from the start and essentially giving up any ability to profit from the use of your work when companies are often making huge profits from it seems less easy in comparison.
I don't ask anyone to share my ideals but conflating these two is dishonest.
- OSS is valuable for decentralizing power and influence
- AI as it is being developed is likely to centralize it
That was the intention and hope, but I think the past twenty years has shown that it largely had the opposite effect.
Let's say I write some useful library and open source it.
Joe Small Business Owner uses it in his application. It makes his app more useful and he makes an extra $100,000 from his 1,000 users.
Meanwhile Alice Giant Corporate CEO uses it in her application. It makes her app more useful by exactly the same amount, but because she has a million users, now she's a billion dollars richer.
If you assume that open source provides additive value, then giving it to everyone freely will generally have an equalizing effect. Those with the least existing wealth will find that additive value more impactful than someone who is already rich. Giving a poor person $10,000 can change their life. Give it to Jeff Bezos and it won't even change his dinner plans.
But if you consider that open source provides multiplicative value, then giving it to everyone is effectively a force multiplier for their existing power.
In practice, it's probably somewhere between the two. But when you consider how highly iterative systems are, even a slight multiplicative effect means that over time it's mostly enriching the rich.
Seven of the ten richest people in the world got there from tech [1]. If the goal of open source was to lead to less inequality, it's clearly not working, or at least not working well enough to counter other forces trending towards inequality.
[1]: https://en.wikipedia.org/wiki/The_World%27s_Billionaires
Fine for him, but it's totally reasonable for people to want to use the GPL and not have it sneakily bypassed using AI.
A lot of work was done by the team to release the source code of many of these games.
I know TTimo did a lot of work to help.
https://youtu.be/ucXYWG0vqqk?t=1889
I find him speaking really soothing.
This is demonstrably incorrect given how LLM are built, and he should retire instead of trolling people that still care about workmanship. =3
"A Day in the Life of an Ensh*ttificator"
https://www.youtube.com/watch?v=T4Upf_B9RLQ
- Sharing/working on something for free with the hopes that others like it and maybe co tribute back.
- Sharing something for free so that a giant corporation can make several trillion dollars and use my passion to train a machine for (including, but not limited to) drone striking a school.
I respect Carmack so much more now. I always scratched my head why he made Quake GPL. It was such a waste. Now it doesn't matter anymore. I so thankful copyleft is finally losing its teeth. It served its purpose 30 years ago, we don't need it anymore.
It seems like Carmack, like a lot of tech people, have forgotten to ask the question: who stands to benefit if we devalue the US services economy broadly? Who stands to lose? It seems like a lot of these people are assuming AI will be a universal good. It is easy to feel that way when you are independently wealthy and won't feel the fallout.
Even a small % of layoffs of the US white collar work force will crash the economy, as our economy is extremely levered. This is what happened in 2008: like 7% of mortgages failed, and this caused a cascade of failures we are still feeling today.
The purist forms of capitalism I’ve seen are places with low prices, a large working class, practical marketing, and high competition - often they’re considered “3rd world” places.
The US economy, if it wants to remain “1st world” must have high prices. It has to contain an element of scarcity (however faux) in order to be sold at a premium, or be able to impart some privileged (institutional) knowledge as a firm - which should be as esoteric as it is scarce.
It can’t be quality alone since all building and manufacturing is effectively outsourced. It has to have a premium brand recognition or monopolistic aspects to it that necessitate a high price.
So the challenge for the first world, during the rise of China (Mexico, etc.), is to find new ways to justify the privileged position using this new technology as a lever to do so.
Training an AI on GPL code and then having it generate equivalent code that is released under a closed source license seems like a good way to destroy the copy-left FOSS ecosystem.
The places I have found AI most useful in coding is stripping away layers of abstraction. It is difficult to say as a long time open source contributor, but libraries often tried to cater to everyone and became slow, monolithic piles of abstraction. All the parts of an open source project that are copyrightable are abstraction. When you take away all the branching and make a script that performs all the side effects that some library would have produced for a specific set of args, you are left with something that is not novel. It’s quite liberating to stop fighting errors deep in some UVC driver, and just pull raw bytes from a USB device without a mountain of indirection from decades of irrelevant edge case handling.
This is what it was before AI, and it remains so today.
AI reproducing code without holding rights to it are a failure case that should be eliminated,
I think this debate is mainly about the value of human labor. I guess when you're a millionaire, it's much easier to be excited about human labor losing value.
You can’t put this genie back in the bottle, and this type of event of human labor being deprecated isn’t new.
The forecasts are still not well defined and speculative.
I can understand his stance on AI given this perspective. I have a harder time empathizing his frustrations. Did he also have a hard time coming to terms with the need for AGPL?
MIT asks for credit. GPL asks or credit and GPL'ing of things built atop. Unlicense is a free gift, but it is a minority.
AI reproduces code while removing credit and copyleft from it and this is the problem.
Why? The software is still there and you can still go choose to use it.
It is far healthier to see it as a collaboration. The author publishes the software with freedoms that allow anyone to not only use the software, but crucially to modify it and, hopefully, to publish their changes as well so that the entire community can benefit, not just the original author or those who modify it. It encourages people to not keep software to themselves, which is in great part the problem with proprietary software. Additionally, copyleft licenses ensure that those freedoms are propagated, so that malicious people don't abuse the system, i.e. avoiding the paradox of tolerance.
Far be it from me to question the wisdom of someone like Carmack, but he's not exactly an authority on open source. While id has released many of their games over the years, this is often a few years after the games are commercially relevant. I guess it makes sense that someone sees open source as a "gift" they give to the world after they've extracted the value they needed from it. I have little interest in what he has to say about "AI", as well.
Hey John, where can I find the open source projects released by your "AI" company?
Ah, there's physical_atari[1]. Somehow I doubt this is the next industry breakthrough, but I won't look a gift horse in the mouth.
[1]: https://github.com/Keen-Technologies/physical_atari