+1 for Xiang vision! We need more reviewers to let us spot some bad quality code.
Forbidden AI will not prevent someone from pretending to be the author of a code, it is better to announce he/she is using AI because we will raise the analysis criteria even more to spot bad code. BR, Alan On Thu, Jul 16, 2026 at 6:58 AM Xiang Xiao <[email protected]> wrote: > As one of the most active reviewers/committers in this project, what I > actually care about when reviewing a PR is: > Does the code follow our style and architecture? > Is it tested, and can the author explain and defend it during review? > > These criteria catch bad AI submissions just as well as bad human ones. > A policy that bans the tool rather than the bad outcome is both > unenforceable (we can't reliably detect AI authorship) > and would penalize contributors who use AI responsibly. > > So my suggestion: > Require disclosure of substantial AI assistance in the PR description — > agreed. > Require the author to fully understand the change and respond to review by > themselves — agreed. > Keep "all PRs reviewed and approved by a human before merging" — this is > already our practice. > But do not prohibit AI-assisted or AI-generated code per se; judge the > code, not the tool. > > In short: hold humans accountable for what they submit, instead of banning > how it was produced. > > Thanks > Xiang > > On Thu, Jul 16, 2026 at 4:18 PM Sebastien Lorquet <[email protected]> > wrote: > > > Note: this email is not an attack on anyone. it's just my personal > > analysis. > > > > Hi Alan, > > > > If it was realistic, I would agree, but we all know that "ai" tools are > > designed to trigger the gambling addiction inherent to human nature and > > *absolutely no one* uses these tools "responsibly". The temptation is > > just too strong when you have started using it to generate code. Let > > alone verify it. Once you've generated 3-4 working functions with it, > > the temptation to trust the generation is immense. > > > > This addiction is very strong, and hard to realize once you've switched > > your deep mind into thinking that ""ai" could do something useful". > > > > You will probably argue that I am wrong and that you are reasonable. > > okay, but what about other people you have no control on influence on? > > > > Just remember that this whole conversation started for the EXACT reason > > that "ai" was not used responsibly for nuttx contributions. > > > > So, I dont believe in "reasonable "ai" usage". It's a theoretical > > well-being thought to allow yourself the use of a known bad tool that > > save time when your boss gives you pressure. > > > > > > Also, a calculator gives exact results 100% of the time. I dont need to > > verify that a calculator actually gave me sqrt(2) as a result. > > > > On the contrary "ai" only gives hallucinations that happen to do some of > > the job some of the time, and no one reviews the code most of the time, > > except in rare situations. > > > > But it talks to you like a human would, says sorry when it wipes your > > home dir, has a human name, confirms your opinions, and then we fall for > > the "junior assistant dev" mental image, because it's designed for that. > > > > > > We all know that seriously proof reading code is 10 times harder (and > > more boring) than writing it in the first place. Because we're not in a > > flow mental state when proofreading, there is no dopamine, no > inspiration. > > > > This gives even more hard work to nuttx reviewers. > > > > Sebastien > > > > > > On 7/15/26 21:27, Alan C. Assis wrote: > > > Hi Sebastien, > > > > > > The number of real contributions developed using AI is still not high, > > that > > > is happening at this moment are contributions for fixes using AI. > > > > > > But the number of contributions using AI will increase for sure as the > AI > > > agents become more reliable. > > > > > > Honestly I'm not against the use of AI, when done with responsibility > by > > > people with skill to understand if that generated code has good quality > > or > > > not. > > > > > > AI is just a tool, like a calculator. You can decide not to use a > > > calculator and do all the maths by yourself (in our head or in pencil > and > > > paper). > > > > > > Both approaches are fine, since you already know how to do the math in > > the > > > first place. The only problem in this case is that you will become lazy > > > (why do mental math if I can use a calculator?) > > > > > > And just like a calculator, if used by a kid that doesn't know how to > do > > > math, he will never master it (I think this is the real issue, for the > > > person themself). > > > > > > BR, > > > > > > Alan > > > > > > On Wed, Jul 15, 2026 at 5:21 AM Sebastien Lorquet < > [email protected]> > > > wrote: > > > > > >> Hello, > > >> > > >> I agree with you but unfortunately, we're way past this stage already. > > >> > > >> The usage of AI in NuttX was started "as a normal thing to do" even > > >> before permission was requested and things were discussed on the > mailing > > >> list. > > >> > > >> This is pretty much a tainted project and it has been for months, and > > >> there is no turning back at this point. > > >> > > >> Sebastien > > >> > > >> > > >> On 7/15/26 06:36, Lwazi Dube wrote: > > >>> Was there an option presented for those of us who want to keep our > > >>> open-source projects 100% free of AI-generated code? That discussion > > >>> absolutely should have happened. If there’s a link to it, please > > >>> share—imposing these tools on people who didn't ask for them is not > > >>> acceptable. > > >>> > > >>> On Wed, 1 Jul 2026 at 11:47, Matteo Golin <[email protected]> > > >> wrote: > > >>>> Hello everyone, > > >>>> > > >>>> NuttX still has not adopted an AI policy, and the number of > > >> substantially > > >>>> AI-generated contributions is continuing to grow. Recently, the > Godot > > >>>> project adopted a new AI policy which I think is quite reasonable. > [1] > > >>>> > > >>>> I would like to suggest that NuttX adopt the AI policy from Godot > [1] > > >>>> (slightly modified for more clarity), as follows, and include it in > > our > > >>>> contribution guide: > > >>>> > > >>>> - *No autonomous AI agent use or vibe coding* > > >>>> > > >>>> - A human must be involved in the coding process if patches are > > >>>> submitted > > >>>> > > >>>> - *No use of AI to generate substantial pieces of code* > > >>>> > > >>>> - We require all code to be human authored. AI assistance > should > > be > > >>>> limited to menial things (like code completion, regex, > > formatting, > > >> or > > >>>> find > > >>>> and replace). > > >>>> - If you do use AI in some capacity to author code, you must > > >> disclose it > > >>>> in the PR discussion. > > >>>> > > >>>> - *No AI-generated text in human-to-human communication* > > >>>> > > >>>> - When our maintainers volunteer their time to review your > issue, > > >> PR, or > > >>>> proposal, they do not want to talk to a machine. This is a > basic > > >>>> principle > > >>>> of respect. > > >>>> - Machine translations are still acceptable as long as the > > original > > >>>> content was written by a human. > > >>>> - This includes PR descriptions and comments. > > >>>> > > >>>> - *All PRs must be reviewed and approved by a human before merging* > > >>>> Please let me know your thoughts, I really think it is time to adopt > > >> this > > >>>> change as I am seeing more and more frequently that substantially > > >>>> AI-generated PRs are submitted (what is really most frustrating is > its > > >> use > > >>>> in human-to-human communication). > > >>>> > > >>>> Best, > > >>>> Matteo > > >>>> > > >>>> [1]: https://godotengine.org/article/contribution-policy-2026/ > > >>>> > > >
