Archives
Contributors
Re: Guidelines for LLM generated contributions
+1000 with Raphael
|
|
||||||||||
|
Nouveau
Votre avis est important pour nous ⭐ |
Découvrir |
Le jeu. 18 sept. 2025 à 09:22, Maxime Chambreuil <notifications@odoo-community.org> a écrit :
Totally agree with Raphael here.
DIRECTOR INTERNACIONAL
M: +52 (442) 114-9164 | WhatsApp
C: MChambreuil@OpenSourceIntegrators.com
Santiago de Querétaro, Querétaro, 76100, México
Analizar. Optimizar. Automatizar. Transicionar.
De: Raphaël Valyi <notifications@odoo-community.org>
Enviado: jueves, 18 de septiembre de 2025 5:36 a. m.
Para: Contributors <contributors@odoo-community.org>
Asunto: Re: Guidelines for LLM generated contributionsThis an extremely important topic as AI is completely changing software development in a matter of weeks.
Of course we should require human contributors to do their correction job before submitting the PR. They can run the CI in their organization 1st, it's easy. May be we should explain how in the migration wiki or contributing.md ?
I woukd say yes it's also important to tell AI has been used and at the same time I'm sure in just a few months the vast majority of us will be using some AI in their PR or will be left behind. So soon enough then telling your PR is using AI will be the noise and not the signal...
There is a positive side: in less than 1 year (may be just 6 months) I think 50% of the modules will be able to be migrated automatically with AI and meet all checks. The same will also be probably true for OpenUpgrade.
This will allow us to use the bleeding edge Odoo with all the ecosystem and free us for more interesting work than migration. At some point this is our chance for not getting distanced by Odoo Enterprise.
This raises other concerns: what will be a valid contributors.md or authors.md entry? How will the OCA keep being organized as a meritocracy when the contributions made will soon enough mostly translate to the amount of compute used? How long before AI will claim PSCs?
May be we should also have repo to build these AI migration or review bots? I did some tools already and will tell about them soon... Reach me personally meanwhile if you are interested.
I see many people in the OCA in the denial phase who think AI is just a hype and or just for toy code. There is definitely a hype bubble and It is not able to do the hardest code today but it already beats you at the 50% of the easiest code you write and it is not going to stop any time soon...
This is really such an important topic...And BTW I believe this will bring dystopia and not abundance. But can we even avoid it?
On Thu, Sep 18, 2025, 6:22 AM Graeme Gellatly <notifications@odoo-community.org> wrote:
I actually just completed a course in ai governance and ethics. It was finance focused as requirement for being an accountant, but declaring usage of llm in generating any professional services output is the very bare minimum. Human review was a mandatory part of governance, but honestly there was like 50 things there.
E.g This PR was generated using x llm and reviewed by a human for correctness prior to submission.
Le jeu. 18 sept. 2025, 21:12, Tom Blauwendraat <notifications@odoo-community.org> a écrit :
Thinking a bit out of the box here, but what might perhaps work is if:
- We set up a separate organisation OCA-ai
- This mirrors the full OCA organization as a fork, similar to odoo vs OCB
- If something looks like AI we close and tell contributors to reopen it there (could possibly be automated with the OCA bot if we can recognize something as AI)
- In that organisation we require 100% test coverage and passing tests, as well as the standard migration process having been followed (could be automated)
- If the bot decides there that it is OK, it automatically reopens the PR to the real OCA organization (where we thus only get "prefiltered" PRs)
In the absence of such automation I +1 Stefan's suggestion of adding a guideline
_______________________________________________
Mailing-List: https://odoo-community.org/groups/contributors-15
Post to: mailto:contributors@odoo-community.org
Unsubscribe: https://odoo-community.org/groups?unsubscribe
_______________________________________________
Mailing-List: https://odoo-community.org/groups/contributors-15
Post to: mailto:contributors@odoo-community.org
Unsubscribe: https://odoo-community.org/groups?unsubscribe
_______________________________________________
Mailing-List: https://odoo-community.org/groups/contributors-15
Post to: mailto:contributors@odoo-community.org
Unsubscribe: https://odoo-community.org/groups?unsubscribe
_______________________________________________
Mailing-List: https://odoo-community.org/groups/contributors-15
Post to: mailto:contributors@odoo-community.org
Unsubscribe: https://odoo-community.org/groups?unsubscribe
by Bruno Joliveau - 03:45 - 18 Sep 2025
Reference
-
Guidelines for LLM generated contributions
Dear all, at least one contributor is planning again to flood the OCA projects with PRs for module migrations: https://github.com/OCA/web/issues/3285. This volume is likely made possible through automation, with an LLM generating the actual migration code (on top of, hopefully, a more deterministic tool like OCA's odoo-module-migrator). Regardless of the volume and the submitter, if the submitter has reviewed, refined and tested the code generated by an LLM, this should not be a problem but as a reviewer I'd like to know what I can expect. Holger Brunn pointed out to me that in other projects, this may be covered by a demand in the guidelines to disclose LLM usage and its extend. For an example, see https://github.com/ghostty-org/ghostty/pull/8289/files. I would very much like to see such an addition to the OCA guidelines. Additionally, I would like to suggest that the basic premise (the generated code is indeed first self-reviewed, refined and tested) is also made explicit, and that it is unacceptable to pass on reviewer comments to the LLM only to copy back the LLM's response (which has happened to me on one or two occassions). Can I have a temperature check for your support for such an addition to the guidelines? Or do you have other ideas or perspectives on the matter? Cheers, Stefan -- Opener B.V. - Business solutions driven by open source collaboration Stefan Rijnhart - Consultant/developer mail:stefan@opener.amsterdam tel: +31 (0) 6 1447 8606 web:https://opener.amsterdam
by Stefan Rijnhart - 09:40 - 18 Sep 2025