Repository navigation
DOC: adopt scikit-learn LLM policy - #155
Conversation
* Related to the first item in lanl#108. This patch updates our `README` to switch from banning LLM contributions outright, to adopting the LLM usage policy of `scikit-learn`, and linking to their policy. * The discussion there seems in favor of doing this.
|
I have no issues with this policy. |
|
I guess my only concern with this LLM policy is that we are attaching ourselves to the scikit-learn hyperlink.
I am good with either adopting this commit as is, or, copy and paste those last 5 paragraphs in that section and cite that this came from scikit-learn. |
* Instead of linking to the `sklearn` LLM contribution policy, we now reproduce it in our docs (substituting in our project name), so that we avoid silent policy drift. * A typo in the previous reference to the `sklearn` project was also fixed.
|
Interesting feedback -- I've revised the PR to reproduce the current version of the There was also a typo in my original diff that I've fixed up in the latest commit. Let me know if there's anything else you want changed. The part about banning the offending account is perhaps a bit aggressive, although if someone doesn't stop I've seen it happen on occasion. |
|
I'm satisfied with the status of the most recent commit. You can resolve my conversations after you've read them and when you're ready to merge. |
|
Alright, I resolved the review comments as requested. Let me know if there's anything else. |
|
Nothing else from me. @aganguly-lab? It sounds like you were happy with the LLM policy already. |
|
Yes. That was a good point that it is better to have the policy written out in our own repo. Otherwise, I still have no issues with the policy. |
|
Alright, two approvals and CI is green so I'll self-merge. Feel free to merge these when folks are happy. |
Related to the first item in DOC, MAINT: minor process/docs improvements #108. This patch updates our
READMEto switch from banning LLM contributions outright, to adopting the LLM usage policy ofscikit-learn, and linking to their policy.The discussion there seems in favor of doing this. Team assistance in chipping away at matters like this would be appreciated--I think just helping out with a few things like this each day is probably more productive than being oppositional and complaining about the number of open issues in meetings, especially with the availability of modern automation tools.
To quote from the policy on violating contributions:
I fully agree with that--reviewer bandwidth is very limited, especially if the reviewers are also driving many other aspects of the project, which is just as true for us as it is for them.
AI usage disclosure: No AI used here.