It's an unfortunate presentation that makes it more difficult to sympathise with - and ultimately it may not matter; either somebody else will fill the gap or the LLMs he resents will help coders create their products without the libraries he seems to think are valueless without a set financial cost.
My point is, monetizing open-source is really hard. Tailwind was giving bricks away for free while selling you a house. Given enough time, people used those bricks to build houses and gave them away for free.
For me the best outcome is taking over all the market currently held by git and GitHub, and the worst outcome would be that Microsoft somehow pulls an embrace-extend-extinguish on us. But my goal, and I think I've been successful here, is to slam the door in the face of anyone who would be able to easily improve on my work and therefore would be able to undercut my value proposition with theirs. We want to be the first people offering a strongly differentiated service by building a platform that does not in any way sit on top of the MS platform, so that nothing we do is actually a direct benefit to them anyway and we actually get consumers excited about change!
This is the same death grip that Microsoft currently has on OSS. They made their tools so ubiquitous that even their most vigorous competitors make things that just feel like knock-offs because the competitors can't afford not to build on top of the same open core as MS: git and LSP. If you can't beat them you have to join them, and so I know exactly what I need to do to win: tweak the economic incentives until joining me is preferable to competing with me, and then instead of charity we'll simply take GitHub's billions in revenue as our revenue.
Yes, LLMs are all take. Yes, they bastardise the licenses and abuse the model. But they don't change why I want to use open source software. They don't change the fact that people need to keep writing it to keep receiving it.
Long story short, I think a lot fewer people are going to use Marc's software. He might be okay with the commercial realities of this (it might be a runaway success) but I think as non-developers get more acquainted with LLMs, being able to ask a bot to change how your system works is going to be more popular, not less, even if it's not the model RMS wanted.
I never had the time or understanding to setup a business around it, so in the end I just got rid of everything. No more open source unless it’s purely for my own enjoyment.
I’m VERY appreciative of free and open source projects and organizations that don’t try to make money. The FSF, GNU, Linux- I love these! I would love if the world just focused on these! I would use a Linux desktop and phone if they were in that Apple sweetspot (iOS 26 not included) of ease for techs and non-techs.
What I’ve noticed about successful open-source projects is that they have:
- a strong leader, who has dedicated their life to it, not as a secondary initiative, and who one day knows they must hand the reins to another
- a willing team that continues to bring in “new meat”
- a project that sells itself by filling a need that will remain for some time by many
- no need or desire for money
Those that end in failure involve:
- a leader/team that sees it as part-time volunteer work, and they do it because it makes them feel good about themselves; they are giving back
- the assumption that one day maybe they’ll get paid
> I've contributed millions of lines of carefully written OSS code over the past decade
I have never tracked my own output, but I would be surprised if I have written even a single million lines of code, let alone several.
Out of curiosity, is this common? Have people here actually written millions of lines of code themselves? Even assuming two million lines over thirty years, that still averages to around 180 lines per day, every day. It is doable but it is a very high level of sustained output.
I am not doubting the OP. Just noting that this represents an extraordinary amount of coding by any measure.
"I need a caching component in my stack. Oh, I just let AI create it from scratch, so AI can also run it, maintain it etc."
or
"I need a caching component in my stack. Oh, let me just grab an existing, well maintained and supported cache component and let AI write the glue logic to run it, maintain it etc."
Personally I think it will be the second one. There definitely is a case for generating one off tools but I don't think AI will replace established, well written and maintained software (being it open-source or not).
Not all software is equal. I can get 80% of my new SaaS to be generated by AI as it's mostly considered boilerplate and mostly the same as any other SaaS. It's the 20% that makes it unique OR something that AI can't touch like a special relationship with a supplier or a first to market position etc.
Just because AI is "taking advantage" of projects being open source doesn’t mean the solution is direct closed source. Going with the most direct solution to a problem is an indication you might not have spent time thinking deeper about the problem
Open source has always felt explicitly like a benefit for companies.
- They get free code, buy vs build is irrelevant when you can just pip install.
- Systems become largely homogenized, thus contributors are replaceable.
- They get an established pool of workers who know the technology already, no training required.
- They get free labor from contributors outside their organization maintaining their dependencies for them in perpetuity.
It's a great deal for employers! Especially if they forbid their employees from contributing back! If you work out the game theory, there's literally no reason for a company to do anything but sit back and siphon the benefits for themselves.
This doesn't really change with LLMs, it just makes the end game much more explicit. The goal was always to capture the intellectual output of open source contributors for private profit. Always. Now that it's actually happening, who's really shocked?
It's quite ironic that they used an LLM to write or at least entirely re-format their post, when their topic is about the impact these systems have on the ongoing sustainability of the humans behind the work.
I personally don't use LLMs and generative models, I find their output way too untrustworthy and their practice of mining the data of others unsettling. Not that anything on the internet can be inherently trusted anyway.
So… SaaS and how every cloud provider already monetizes open-source code?
Problem with this is demand like that can vanish once you paygate it.
OSS was exploited by hyperscalers to build trillion-dollar industries atop of, but without ever suitably compensating, the creators of much of those tools for their work or sharing in the profit. Before AI, the community was already at a breaking point between private enterprise spouting “supply chain” bullshit at them to demand fixes and attention, or steamrolled small devs to prop up a big corp’s trademark or product, all the while never actually paying enough for the proper development and support of those products - look at NPM (leftpad and kik) as prime examples. Now you have these same big tech ghouls scraping small sites into oblivion with hostile bots, making token predictors that are deliberately engineered to never, ever direct someone to a primary source or site except as an absolute last resort (to keep folks “in app” for engagement metrics), and openly pitching AI coding agents as replacements for human coders forever.
In that context, it’s no fucking wonder that the OSS community is becoming increasingly hostile to the very norms that have left most of them broke and increasingly destitute. Hell, for up-and-coming devs emerging from bootcamps, the mantra of “contribute to OSS” makes zero sense in a Capitalist marketplace that’s pivoting hard towards AI-as-human-replacement; as the OP points out, why bother training your own replacement?
OSS won’t die, but this is a particularly painful chapter that emphasizes it cannot support itself through the (non-existent) generosity of Capital. Alternate funding schemes and organization models are needed to prepare and support it for the future, be they government grants, Academia sponsors, or outright Gov-funded Private or Public Corporations (e.g., BBC, Corporation for Public Broadcasting (RIP), etc). Updates to OSS licensing schemes barring use-cases or with more substantial teeth for commercial use are also needed, toeing the line between empowering users of general computing and extracting reasonable payments from businesses or enterprises.
Marx: Did it occurred to you to destroy the system that make robots doing your job a bad thing?
So what are the actual solutions, now that the cat is out of the bag?
Much like in the post, you could make it source available and have a free license for non-profits and development, but ask money for commercial usage, or different feature sets (like iframe-resizer, which we recently bought for a project, saved a bunch of time, but the AI models still were trained on it and knew how to use it). Or provide support, like many FOSS DBs and OSes should, same for libraries, maybe to have PostgreSQL dethrone Oracle and the likes just a little bit more.
Or maybe some ask-for-funding stuff defined at library level with the expectation that the AIs would present this information to whoever uses them. I don't particularly celebrate https://docs.npmjs.com/cli/v9/commands/npm-fund but I get why it's there - similarly we might get IDE plugins and CLI tools and such to present a summary of what libraries or projects were used/suggested in a task/chat session and how to give them money, much like how you'd get references to website sources when trying to ask AI to research something.
> My prediction: a new marketplace emerges, built for agents. Want your agent to use Tailwind? Prisma? Pay per access.
This probably already exists in the form of MCP solutions for up to date documentation for specific libraries and so on, to mitigate hallucinations.
Either way, we need to start implementing actual solutions so we don't keep ending up with https://staltz.com/software-below-the-poverty-line.html all the time. And not just some "human contract" approach of expecting that someone will go to your docs manually and see the banner that they can give you money for more bells and whistles or something. To me, that feels like wishful thinking.
AI or not, people's work should be rewarded and that needs some sort of standardization - and I hold that view even in regards to answering feature requests on GitHub, like if someone demands something, they should immediately have the option to place money into an escrow for when/if a solution is presented to them (or get the money back if not). If they don't, you have no obligation to help them unless from the goodness of your own heart and the innate desire to do so. Similarly, if I use 10 libraries, I should be able to say "Okay I have 50 bucks, I want to donate to all of these projects by executing a single npm command and confirming a PayPal payment or something."
Incidentally, that's also the first step towards building the Torment Nexus (capitalist incentives will find a way to turn this into a hellscape), but go figure.
* This is how billionaire techbros will finally push most humans out of tech.
All they're doing now by putting up roadblocks is collectively punishing humans from competing with AI.
On the other side of the token, don't give away things and expect payment. Give things away because it's cool, not as a primary survival strategy based on mythological expectations of gratuity. It comes across as bitter giver's remorse. And unprofessional poison spewed onto social media also wrecks their professional reputation and drives customers away.
AI is perfect for writing CSS because you can tell it your intent and it does it.
There just is not the need for frameworks/libs to make it less complex.