Vote on AI contributions

[mcollina]

In order to resolve the blocks of @indutny

• doc: create ai-guidelines and include to CONTRIBUTING node#62105
• Virtual File System for Node.js node#61478

We should schedule a vote, ideally starting after April 1st, which I think it's the PT-friendly meeting where most people in the US can join.

In parallel, I'm working with the Board to provide a legal/high-ground answer to the question of the DCO's compatibility with AI-assisted engineering.

@aduh95, can you take care to set up the vote?

[jasnell]

Agreed, let's push it to a vote.

[3n-mb]

If someone uses an advanced tool, like LLM, should everyone have a higher baseline for expectations?

Take a look at initial review in github.com/nodejs/node/pull/61478 , looks like begging to use node's style for ... core node.
Picture an analogous PR into Linux kernel, and Linus' reaction. Seriously.

This tool, LLM, is supposed to do things in style. Instead there was a re-implementation of path.normalize() ?
Where would a hacker looks for SQL-inline style place for an attack?

User of an advanced tool has responsibility to provide more advanced result. Of course, in coding, LoC count is not a metric for being better.

[jasnell]

@3n-mb ... this issue is not the place to discuss the merits of AI contributions. This issue is just to call for the TSC vote.

[3n-mb]

@3n-mb ... this issue is not the place to discuss the merits of AI contributions. This issue is just to call for the TSC vote.

Please direct me to place, where you collect material to ponder before making a vote.
Uh. Placed it in nodejs/node#62105 (comment)

[mcollina]

Here is the AI policy of OpenJS, answering the question on the DCO: https://openjsf.cdn.prismic.io/openjsf/aca4d5GXnQHGZDiZ_OpenJS_AI_Coding_Assistants_Policy.pdf.