CONTRIBUTING.md: Add LLM contribution policy#264
CONTRIBUTING.md: Add LLM contribution policy#264ahelwer wants to merge 1 commit intotlaplus:mainfrom
Conversation
Signed-off-by: Andrew Helwer <ahelwer@pm.me>
ahelwer
requested review from
damiendoligez,
kape1395,
lemmy and
muenchnerkindl
|
I suggest that the TLA+ project first establishes consensus on its approach to handling contributions that are partially or entirely generated by LLMs, before we develop policies for individual tools. |
| LLM usage also must be disclosed and the precise model used identified. | ||
| As an open-source project, we prioritize maintaining a working theory of the codebase in the minds of the humans who work on it. | ||
| Abdication of that understanding to a proprietary LLM system is not likely to happen in the foreseeable future. | ||
| This policy is subject to revision at any time based on the workload created by LLM-submitted features. |
There was a problem hiding this comment.
My proposal (aimed at transparency, not judgment): "Initial contact with the project must be made by a human. For each commit, any use of an LLM must be disclosed, including identification of the specific model(s) used."
|
|
||
| We welcome your contributions to this open source project! | ||
| TLAPS is used in safety-critical systems, so we have a contribution process in place to ensure quality is maintained; read [CONTRIBUTING.md](CONTRIBUTING.md) before beginning work. | ||
| Especially read this document if you are using a LLM agent. |
There was a problem hiding this comment.
This approach is unlikely to encourage the people who actually need to read CONTRIBUTING.md to do so. For everyone else, it adds unnecessary distraction. It also comes across as somewhat hostile in tone.
|
2 participants
For reference, here is the Linux kernel policy: https://github.com/torvalds/linux/blob/254f49634ee16a731174d2ae34bc50bd5f45e731/Documentation/process/coding-assistants.rst
Opened after #263