

Ah I misunderstood the purpose. I thought it was to provide the community with a guide they could use to decide which games to support rather than a tool to condemn developers for their choices. Sorry for the misunderstanding.


Ah I misunderstood the purpose. I thought it was to provide the community with a guide they could use to decide which games to support rather than a tool to condemn developers for their choices. Sorry for the misunderstanding.


This is a good idea but I don’t think the current design adds much over existing categorization. There are already ways to find no-ai/handcrafted games. And complete slop is (at least for now) fairly obvious to spot. Imo the opportunity to refine the scale is all in the middle which you’ve collapsed into a single category.
I think most people who are willing to accept some ai use would benefit from a more nuanced approach. For instance, many people might be ok with some minor AI assistance in the programming process, but draw a hard line at ai assets. Your scale doesn’t serve those users.
I’m not arguing with your very valid points on the ethics and result of AI, just pointing out that to add utility past the current options you might need more granularity.
Also a major gray area is the use of libraries that may include ai code. For example, most games are now made with engines like unity, unreal, or godot. It’s almost certain that unity and unreal now include ai code. So can a game dev never use an engine again without falling into the tainted category? Do they have to write their own quaternary code for 3d translations, sprite blitter, texture mapper etc from scratch? What about using the windows API for file handling? That’s likely got ai code in it now. I guess to be considered handcrafted games need to be freebsd only now? Or ship with their own operating system?
I think we’d be better served by a scale like this:
0 - fully human made
0.5 - used a game engine / libraries / os SDK etc which likely includes AI generated code but is itself not considered total slop
1 - used some minor AI assisted coding techniques (e.g. code completions / merged an ai big fix after thorough review, etc)
2 - used extensive but human directed programming, assets all human made. Shows evidence of care put into design, graphics, story, etc
3 - vibe coded or close to it, but all assets still human made
4 - some ai assets, human directed work
5 - obviously totally vibe coded with generated assets
I would actually argue that there is some validity to scrutinizing code supply chain. For example a dev could choose to a human-crafted library in c++ or use a vibe-coded “port” to rust or whatever because that’s the language they’re more familiar with. Those choices are worth investigating.