Clearly disclose Claude usage

[swedneck]

Given the many contentions with LLMs, I think it's good practice to make it obvious when it is used in development, for those who wish to avoid it.

[d4rken]

Sorry but I disagree.
This neither needs to be highlighted, nor explicitly hidden (which it isn't).

These days you can safely assume most software projects touched an LLM at some point. Maybe through an IDE assistant, a CLI tool, or just copying a small snippet. It's basically today's Stack Overflow.

We don’t tag code as "IntelliSense generated." It suggests stuff, the developer checks it and decides what stays. Same with LLMs, just bigger suggestions.

This isn't vibe-coding. It's just another tool, and the person shipping the code owns it.

[swedneck]

Except that is not at all certain, the legal status of LLM-generated code is uncertain since the models are trained on data with no regard to licences or copyright, and I'm not aware of any significant court decisions on who owns the copyright of the output.

I also don't understand why you'd be opposed to disclaiming the usage just because it's standard practice?
It's also normal to use libraries, and yet those are expected to be clearly listed.

[d4rken]

You're not wrong that this may be a legal gray area right now, and once it's clearly settled I'll follow whatever rules apply. But this repo is about the code and implementation, not legal or philosophical debates about LLMs. I understand this matters to you, but it's not something I'm going to cover in this repo. Thanks.