Um, where’s that?
No, that’s purely a (forced) guide to let people know how a title was made.
Now, one could argue that the form is not needed; one could just post in that category without having to go through those details. But, as I’ve experienced in other communities (as well as here), people tend to be more interested in the process when AI is involved; and also, it hopefully forces/inspires a level of transparency.
Well, it’s not like I spent an inordinate amount of time to come up with the segmentation/categorization, and it can be changed easily and quickly. However, looking at ai-declaration.md, I don’t think that there’s any advantage to do so?
I’m open to tweaking the schema, no biggie, but if we do so it’ll be geared even more towards our own particular needs and workflows, not closer to the generic declaration.
You have to look down the road a little. I’ve seen a lot of communities facing this issue at some point: AI (aka SI according to Trump
) begins to be used, then the community is literally flooded with AI releases, then everyone begins to demand ways to filter said AI releases. I’ve seen it happen to communities ranging from games to mobile apps and from watchfaces to music. And, seeing such releases coming in hot and fast, I am ready to bet my right arm it will pose the same problem for our community as well.
No, because:
A. People tend to misuse/ignore tags altogether. If nothing else, your kinda-regex solution would flood the tags repo. And
B. Because of the previous point; my guess is people will want/demand a better, clearer demarcation line and separate AI titles completely.
The form is flexible enough; we’re not doing forensics here, we just want some transparency and some idea of how a title was made. No need to get bogged down with the many nuances.
That is an absurdum point, but I do get it. Again, though, we’re not putting stuff under the microscope. Whatever a dev feels like sharing, it’ll be enough.
To reiterate, tags is an impossible solution. Plus, it faces the exact same problems you yourself outlined with regards to the current implementation anyhow.
Not really; we already have a WIP tag and @XeNoMoRPH wisely puts [WIP] in his titles whenever such a project appears. That’s a solved problem.
Personal note: I am not against the use of AI. In fact in my personal/work life I use a variety of LLMs and methods to help me produce deliverables: from simple research to local LLM training and knowledge graphs building, and from market research for the best strainer for my kitchen to aid in building complex taxonomies. But I don’t do art, which I consider games to be, and I never pass on the LLM’s work as a final product at work.
By the way, thanks for taking the time to comment so thoroughly
I do enjoy the back-and-forth.
Say, how about an agent to interrogate the knowledge found on the wiki and the forum?