The decision of the market seems pretty clear. We've been able to co-operate and build a software commons for decades, iterating on and improving shared infrastructure and solutions to problems common and niche. The work done for these commons, though, benefits everyone, and that's a hard sell for a profit-driven organization. So the commons are enriched with
a) volunteers
b) brief windows in which corporate decision makers are driven by ideology and good intentions, where those decisions carry momentum or license obligations (see Android, and how Google tries to claw it back)
c) corporations attempting to shape the larger landscape or commoditize their complement, see Facebook's work on React, or contributions to the Linux kernel
Of the above, only (a) or rarely and temporarily (b) are interested in collective wellbeing. Most of the labor and resources go into making moats and doing the bare minimum to keep the shared infrastructure alive.
Now companies selling LLM coding agents enter the scene, promising to eliminate their customers' dependence on the commons, and whatever minimal obligations they had to support it. Why use a standard solution when what used to be a library can now be generated on the fly and be part of your moat? Spot a security bug? Have an agent diagnose and fix it. No need to contribute to any upstream. Hell, no upstream would even accept whatever the LLM made without a bunch of cleanup and massaging to get it to conform with their style guides and standards.
Open source, free software, they're fundamentally about code. The intended audience for such code is machine and human. They're not compatible with a development cycle where craft is not a consideration and code is not meant to be read and understood. That is all to say: yes, it is unrealistic to expect companies to donate anything to the commons if they can find any other avenue. They prefer a future where computer programs are purchased by the token from model providers to one where they might have to unintentionally help out a competitor.
> Now companies selling LLM coding agents enter the scene, promising to eliminate their customers' dependence on the commons, and whatever minimal obligations they had to support it.
This is misguided. Maintenance of LLM code has a far greater cost than generating it.
> They prefer a future where computer programs are purchased by the token from model providers to one where they might have to unintentionally help out a competitor.
I don't think that's even a thought. The thought is that "no one can tell me no".
The longevity of code depends at least on whether it's a product or a service.
Services are what the majority of devs already work on and maintain. There's almost no incentive for anyone to use LLMs for that outside of startups. They do indeed last a long time because the code is as fundamental to the recurring revenue of the business as their legal or accounting or marketing. Devs make changes according to the evolving needs of the business, and "productivity" isn't as much of a priority as accuracy and reliability. The implementation details are very relevant to the business, especially for B2B services that need to meet compliance requirements.
Products, however, have always been disposable code written by people being thrown into a meat grinder. I don't think LLM-generated code is better, but it's probably not that much worse either.
> This is misguided. Maintenance of LLM code has a far greater cost than generating it.
I agree. I'm just observing what they're doing.
> I don't think that's even a thought. The thought is that "no one can tell me no".
I doubt there's any one thought driving things. I didn't mean to imply the existence of some grand strategy or scheme. The preference I speak of isn't of any person, it's the direction pointed at by incentives and circumstance. Companies will make decisions to steer clear of helping competitors. Separately, they signal great interest in replacing costs spent on labor with costs spent on services. See the transition to cloud. The result is the preference of a world where code is like gasoline, purchased from a handful of suppliers for metered cost.
a) volunteers
b) brief windows in which corporate decision makers are driven by ideology and good intentions, where those decisions carry momentum or license obligations (see Android, and how Google tries to claw it back)
c) corporations attempting to shape the larger landscape or commoditize their complement, see Facebook's work on React, or contributions to the Linux kernel
Of the above, only (a) or rarely and temporarily (b) are interested in collective wellbeing. Most of the labor and resources go into making moats and doing the bare minimum to keep the shared infrastructure alive.
Now companies selling LLM coding agents enter the scene, promising to eliminate their customers' dependence on the commons, and whatever minimal obligations they had to support it. Why use a standard solution when what used to be a library can now be generated on the fly and be part of your moat? Spot a security bug? Have an agent diagnose and fix it. No need to contribute to any upstream. Hell, no upstream would even accept whatever the LLM made without a bunch of cleanup and massaging to get it to conform with their style guides and standards.
Open source, free software, they're fundamentally about code. The intended audience for such code is machine and human. They're not compatible with a development cycle where craft is not a consideration and code is not meant to be read and understood. That is all to say: yes, it is unrealistic to expect companies to donate anything to the commons if they can find any other avenue. They prefer a future where computer programs are purchased by the token from model providers to one where they might have to unintentionally help out a competitor.