news.volyx.in

Godot will no longer accept AI-authored code contributions (pcgamer.com)

561 points by pjmlp · 11 days ago · 403 comments on HN

Article summary

The Godot Foundation, maintainers of the open-source game engine Godot, have announced that they will no longer accept code contributions authored by AI. This decision comes after a significant increase in AI-generated pull requests, which have become 'increasingly draining and demoralizing' for the project's code reviewers. The foundation cites the need for contributors to be accountable for their code and able to fix it if it fails. This change in policy aims to reduce the burden on maintainers and ensure that contributions come from humans who understand their code.

Main themes

  • AI-generated code
  • Open-source software
  • Code review
  • Accountability
  • Software maintenance

What commenters say

  • The use of AI-generated code can lead to subtle hidden bugs that humans may not be able to explain or fix.
  • There are valid non-functional reasons for rejecting AI-generated code, including provenance, accountability, and the importance of human learning and ownership.
  • Allowing AI-generated code can lead to a lack of understanding and responsibility among contributors, making it difficult for maintainers to review and fix code.
  • Some argue that the origin of the code, whether human or AI-generated, is irrelevant as long as it runs and meets specifications, but others counter that this view is naive and irresponsible.
  • The inability to verify the correctness and reliability of AI-generated code is a significant concern, particularly in large and complex codebases.
  • Rejecting AI-generated code is seen as a way to encourage contributors to take ownership of their work and to ensure that they have a deep understanding of the code they are submitting.
  • The use of AI-generated code can also raise concerns about copyright and licensing, particularly if the AI has been trained on proprietary or copyrighted material.
  • Some commentators argue that the focus should be on developing processes and tools to verify and validate AI-generated code, rather than outright rejecting it.