So I play an open source game called Cataclysm: dark days ahead , and I had my save get bricked over some jank. I used llms to unbrick my save (I compiled the code put my save in and I was able to continue). So I did fork request and disclosed that I used llms to generate the code just to be honest and respect the people that care about llms.
So since I used an llm articulated solution for the problem and the community doesn’t allow llm content could that pull request be used an an excuse as a derivative source for the problem that bricked my save?
The rest is just a rant. I did go to the discord and Post that I made a pull request and I wanted to make sure I following procedure. I disclosed that I used LLM in my pull request and the user proceeded to get hostile, like actually triggered. I have to defend myself for my disclosure and they try to say I’m violating a discord rule with my disclosure, even though the rule is that I can’t post ai generated content in the discord. Then they demand the subject get dropped after they escalate.
I probably needed to get disillusioned over a game I like so much. I’ll put the game down and move onto other things.
I just looked at your PR, it’s bad, shouldn’t be merged. The short version is it was clearly built with LLMs, and it would be more of a hassle to review your code than to write the fix by yourself. Which is the main reason small teams opt against LLM generated code, as allowing it would keep them from writing or reviewing better code.
I know from your point of view it seems harsh, but imagine you could do your job, or review what a kid did when trying to do your job. One kid might be something you’re willing to do, but if you have to review thousands of kids you will never get to do your job.
They clearly state they don’t allow LLM code, opening an issue and possibly linking to your fork with the LLM written fix would have been the right way to go.
If their rules say no llms, you’re showing how little you respect the project by intentionally violating those rules. Idk why it’s surprising that someone who says no submitting llm generate code is mad at someone submits llm generate code anyways. That’s a pretty simple cause-and-effect chain right there. I’m not surprised devs who ban the use of llms are mad at someone asking them to incorporate llm code into the project.
Read the rules and abide by the rules or challenge the rules when you think they need to be changed; don’t just ignore them and play victim when called out in it.
Look at this guy 🤣. Rules didn’t say I couldn’t submit llm codes in a pull request just that the project won’t accept it. Plus they can just reject the pull request and move on
I still respected the people that care about llms by disclosing my submission was made with llms
Well, no, you didn’t actually. Because the project rules say explicitly that LLM submissions will not be accepted. Just telling them that you used an LLM to do it does not change the fact that it’s an LLM generated PR. It’s going to get rejected as a waste of everyone’s time including your own.
This is not “respect”, this is “yeah I read your rules but I’m electing to ignore them anyway”.
It doesn’t, but it does imply that my submission will be rejected because it’s sourced from llm.
But I disclose in line 1 that it’s sourced from an llm.
Sharing the solution to a save bricking glitch is still worth it
It does. From CONTRIBUTING.md :
All contributions must be authored by the submitter, with a few exceptions such as porting changes from another fork while maintaining attributions. All contents sourced from LLMs such as ChatGPT (including but not limited to code, configuration, issue and PR text, research, and testing results) are not allowed as they do not have any authorship, and additionally the provenance of this kind of content is legally and morally questionable at best. If you are porting content from another fork, you must arrange your commits such that the original authorship is preserved, i.e. the commits should appear in your branch the same as they do in the fork you pulled them from with the Author field indicating the original author. You must additionally perform due diligence that the changes you are porting are themselves authored by a human and not generated by a LLM.
Sharing the solution to the glitch is probably still worth it. Just don’t expect your pr to get accepted, as it breaks the stated rules of the repo. It’ll be a good starting point for someone else to go fix the problem themselves, however.
I said multiple times my pull request would get rejected.
But someone could look at my solution and just digest it into a human made solution
or you could have just submitted a bug report with useful details and how to reproduce the problem
But it’s better to submit a solution to the bug report
If its an open source game you can likely fork the entire thing with your fix and keep playing it.
True and only complaint is that it takes like an hour to compile the game, if I wanted my fix in newer versions


