Conversation
.github/PULL_REQUEST_TEMPLATE.md
Outdated
|
|
||
| ## Checklist | ||
|
|
||
| - [ ] This is my own work, I did not use AI, LLM's or similar technology |
There was a problem hiding this comment.
This wording is too restrictive. Most developers "use AI" in some sense, and asserting "this is my own work" should cover it.
There was a problem hiding this comment.
what usage do you think is acceptable?
There was a problem hiding this comment.
I think either we need to follow what Apache says (easy mode!), or discuss what we consider "using AI". If I get an error message while working and ask claude / gemini / google with the automated AI box at the top what it means, is that "using" AI? If I have never really noticed that VSCode autocomplete is copilot now, is that using AI?
My view on this right now is that if we are not confident in licencing, we should constrain this to clearly be about the PR as output, as opposed to the PR as process. "I am a human, and this code, documentation and other associated works in the PR are written by me, not an LLM or similar technology". Something along those lines.
I'd also suggest in the comments on the template making it clear we don't support bot submissions (eg from openclaw or whatever) or LLM work in general due to licencing issues.
There was a problem hiding this comment.
@SCdF that’s a useful view that I am happy with. I’ll make adjustments accordingly.
@garethbowen I take your +1 as support :)
| Thus, it is expressly forbidden to contribute material generated by AI, LLMs, | ||
| and similar technologies, to the PouchDB project. This includes, but is not | ||
| limited to, source code, documentation, commit messages, or any other areas of | ||
| the project. |
There was a problem hiding this comment.
I like @pjfanning 's suggestion to lean on Apache's docs here: https://www.apache.org/legal/generative-tooling.html
There was a problem hiding this comment.
I don’t agree with the ASF’s take tho :)
garethbowen
left a comment
There was a problem hiding this comment.
I'm ok with this compromise. Approving.
This feels significant enough that it would be good to get more than one review from the community before merging.
As discussed in https://lists.apache.org/thread/93vk0hm2z0pvth88qc7kt928r6h08f7b