| ▲ | perrygeo 4 hours ago |
| For an outright ban on "LLM, AI or agent assisted code contributions", they do very little to define what that means. This is a terribly written policy - I sincerely urge the QGIS maintainers to put enforceable guardrails on the definition. This reminds me of the legal ambiguity behind the GPL and its definition of "derivative work". A careful analysis of the GPL shows that using GPL software doesn't trigger any other clauses. But distribution does. What is the legal definition of distribution? Who knows!?! That's why most software companies have a partial (edit) ban on GPL software. Even companies that are willing to comply with the GPL, they cannot feasibly come into compliance with such a vague definition. So it's safer not to engage. Same will happen with anti-AI policies. Even contributors who are fully willing to do things "by hand" will start avoiding these projects. No way to know if you'll accidentally run afoul of someone's AI radar. There's nothing to protect enforcement from becoming a witch hunt. |
|
| ▲ | giancarlostoro 4 hours ago | parent | next [-] |
| Some people blindly hate AI, what everyone actually wants to ban is fully automated coding, or at least “vibe coded” submissions, if you cannot support your change without an LLM you shouldnt be pushing it into a major project. People can 100% still use LLMs without “vibe coding” in a meaningful capacity and understand the output, and still be drastically more productive. |
| |
| ▲ | perrygeo 4 hours ago | parent [-] | | My point is, if projects are actually serious about banning AI, write that down in enforceable language. As of today, your personal interpretation of what the policy might mean likely conflicts with other's interpretations. There's no clarity on what exactly has been banned. Put simply, if you're so convinced that AI contributions will ruin your project, you owe the world a full and complete definition of what "AI contribution" means. | | |
| ▲ | giancarlostoro 4 hours ago | parent | next [-] | | I agree with you, I'm just laying out what I believe the real ban that makes the most sense would entail. It does sound like they dont even want to hear that you used Stack Overflow if it came up from Google's Instant Answers AI thingimabob. | |
| ▲ | ragall an hour ago | parent | prev [-] | | > My point is, if projects are actually serious about banning AI, write that down in enforceable language. You're under the false impression that the maintainers of a project are obliged to follow the same standards as a court of law. > if you're so convinced that AI contributions will ruin your project, you owe the world a full and complete definition of what "AI contribution" means False. |
|
|
|
| ▲ | stdbrouw 4 hours ago | parent | prev | next [-] |
| If you are the enforcer, which the maintainers are, then surely any rule you come up with is enforceable in exactly the kind of way you'd like it to be enforced? |
| |
| ▲ | pixl97 3 hours ago | parent [-] | | This is a bit naive. People write emotional rules and laws all the time. They may even work for a little while. Then one of two things happens, maybe even both at the same time. First other people poke at the gray areas of the rules which leads to the "you accepted x, but not y, this must be favoritism when allowing AI posts". Things degrade into insane rule lawyering. Or, you end up with insane sets of rules and enforcement that kill the project. We've seen a lot of project leaders recently just act kind of mental. Quite often it will lead to the project forking and falling apart. Well written rules for a project matter even more than well written code. | | |
| ▲ | Pannoniae 3 hours ago | parent [-] | | This is why I prefer very loosely written rules in a pseudo-common-law system :) i.e. have few, generic rules, and have their boundaries be constrained by implementation. The important part is consistency. There is a boundary, it's just unstated. If you're enforcing stuff arbitrarily, this won't work, but if you're mostly consistent this can make things smooth by eliminating the space for rules lawyering, drama and schemes by getting everyone in the community to "sign up" to the status quo. When you're new you don't know where you're at so you join, observe and listen. After some time, you're familiar with the norms so you can engage with them, advocate for changes, etc. Of course, the alternative is codified rules for everything, but the moral hazard there is that the more rules you add, the more they're likely to be contradictory, leading to the possibility of arbitrary enforcement nonetheless despite the pretence that the rules are specific and clear so the enforcement isn't arbitrary at all. |
|
|
|
| ▲ | jefftk 4 hours ago | parent | prev | next [-] |
| > most software companies have an outright ban on GPL software That doesn't sound right: most software companies, for example, make extensive use of Linux. The standard I'm used to is: * Internal use of GPL software is fine. * SAAS products with GPL software are fine. * GPL software leaving company infra is fraught. Either just don't do it, or bring in legal if there's a strong business need. This is different from, say, AGPL, where many companies do just ban it outright. |
| |
| ▲ | perrygeo 3 hours ago | parent [-] | | > Baby companies do just ban it outright This might be the key point. There's legal ambiguity surrounding these licenses. Large companies with legal resources can draw their own line in the sand to protect themselves. But small companies don't have the resources or the bandwidth to address the ambiguity. So a blanket ban (Zero AGPL and GPL for Internal or SAAS, Linux generally gets a pass) is rational. Likewise, I see small dev shops without the resources to address anti-AI policy ambiguity are going to defer contributions until further notice. | | |
|
|
| ▲ | cr3ative 4 hours ago | parent | prev | next [-] |
| I don't feel like any further definition is required. You would surely know if you had used an LLM, AI or agent while creating the code. Keeping things broad allows for simpler enforcement, otherwise you end up in the weeds arguing if seeing the Google AI search summary counts as LLM influence. |
| |
| ▲ | perrygeo 3 hours ago | parent [-] | | Lol. You say no further definition is required - then proceed to outline your interpretation which is quite distinct from others. Clearly further definition is required. If you're serious about an anti-AI policy, you owe the world a definition. It's not hard. You are making a strong statement - an outright ban. Make an equally strong case for what is and is not in scope. You must get in the weeds, since it is you yourself that is banning the weeds. To quote Gustave Flaubert: “The art of writing is the art of discovering what you believe.” Failing to write the definition is a strong signal that no one thought through the problem. | | |
| ▲ | cr3ative 3 hours ago | parent [-] | | No, my example is one of where you end up if you start doing granular definition (and demanding to know where the line is, to the point of absurdity). A broad stroke based on community trust and your own self-evaluation of a very simple rule is acceptable here. If a contributor wishes to poke the edges of the policy to try and find the line of acceptable AI/LLM usage in the project, they are probably not the kind of contributor they want. | | |
| ▲ | tonyarkles 2 hours ago | parent [-] | | > If a contributor wishes to poke the edges of the policy to try and find the line of acceptable AI/LLM usage in the project, they are probably not the kind of contributor they want I think this is where it becomes a question of whether it’s an ideological rejection of LLMs or a technical/practical rejection of LLMs. A project that is rejecting contributions where LLMs are used purely because of the “we can’t handle the amount of newbie generated inbound slop” would, in theory at least, have minimal issue where the contribution is, in fact, indistinguishable from human-only contribution. “This person has put together a solid clean patch that solves a real problem” sounds like a great kind of contributor to have. On the other hand, a project that rejects LLM-assisted patches because of ideology or social or environmental concerns might actually hate a contributor like this even more than a random drive-by slop contributor. |
|
|
|
|
| ▲ | fn-mote 3 hours ago | parent | prev | next [-] |
| > put enforceable guardrails on the definition This is the issue! Without objective, verifiable criteria, the rules cannot be enforced. Rules without enforcement rely on the strength of community norms, which we don’t have in this case. All you’re going to do is end up banning obviously vibe coded crap. If that’s what they mean, say it. In spite of the positive reaction, this IS a bad policy because it isn’t enforceable. Keep thinking - we need guidelines that focus on our needs and goals, not the basic mechanics of code creation. |
|
| ▲ | ToucanLoucan 3 hours ago | parent | prev [-] |
| > For an outright ban on "LLM, AI or agent assisted code contributions", they do very little to define what that means. It means: did you use AI? If so, we don't want it. Like come on. They don't like your new shiny toy and you're mad about it, but there's no need to be purposely obtuse. You fire up an IDE, you write code, you submit it. That is not an AI contribution. You fire up Claude Code and prompt it, you submit it. That's an AI contribution. Contributing to projects is a privilege not a right, and if you fail to contribute, be it because your PR was written like shit, be it your documentation wasn't complete, be it you didn't abide the style guides, or, yes, be it you used AI and they don't want AI, your submission doesn't go in. That's that. 2 + 2 = 4. To be blunt it's wild that otherwise very intelligent people keep getting bluescreened by the fact that they occasionally must reckon with other intelligent people who do not necessarily believe the same things as them. Have y'all never gone outside or what? |
