Consensus is broken, I am reverting it and start a discussion. On Thu, Sep 24, 2026 at 4:24 PM Jarek Potiuk <[email protected]> wrote:
> Sorr. I merged it because I **thought** it was non-controversial (I have > not seen any complaints about it before) - and yeah - I wrote that if there > will be any complaints I will revert it. > > I am going to do now, and start a discussion as apparently it is > apparently controversial. I do not agree with your arguments - I think > it's exact opposite of what you imply - and it's "centaur and not reverse > centaur [1] (or meat proxy)" - where human is in control. > > BTW. I highly recommend the book "Reverse - author of the > "Entshittification" term. The book "The Reverse Centaur's Guide to Life > After AI" is a fantastic book, I literally swallowed it and it opened my > eyes to many things. > > [1] > https://doctorow.medium.com/https-pluralistic-net-2025-09-11-vulgar-thatcherism-there-is-an-alternative-f1428b42a8fd > > > > J > > > On Thu, Sep 24, 2026 at 3:33 PM Ash Berlin-Taylor <[email protected]> wrote: > >> No. >> >> -1 as it stands. >> >> The PR description should already be for humans only. The PR description >> is the way the person opening the PR engages with the community, and being >> a meat proxy shows they don’t care about the consumers of it. Your proposal >> is reversing the priority to my mind >> >> I.e. I think we should remove the `Human Summary ` and `AI Summary` >> headers, and use `<details><summary>AI Summary</summary>` instead of “click >> here”. >> >> Additionally: why did you merge this PR just 30mins after sending this >> email when the date below says “in 4 days”? >> >> -ash >> >> >> > On 24 Sep 2026, at 13:27, Jarek Potiuk <[email protected]> wrote: >> > >> > Hello everyone, >> > >> > Following Shahar's idea - and a number of PRs he has already opened this >> > way - I would like to call for lazy consensus on making it a rule for >> > agent-drafted PR descriptions: >> > >> > https://github.com/apache/airflow/pull/73660 >> > >> > The idea is simple: >> > >> > * A PR description drafted by an AI agent starts with a short "Human >> > Summary" - a few sentences typed by the human author, in their own >> > words: why the change is needed, key decisions, caveats. >> > * The agent-written description is still there, but below it, in an >> > "AI Summary" section that is folded by default. >> > * Agents are instructed to ask the user for the Human Summary and paste >> > it verbatim - they must never write, suggest or paraphrase it >> > themselves. >> > >> > This should make it much easier (not with 100% certainty, but quite >> > reliably) to tell PRs where a human understood and owns the change >> > apart from fully AI-generated ones. It also gives reviewers a quick, >> > human-level "why" before they dive into the details. >> > >> > The PR only changes the guidance our agents follow (AGENTS.md and the >> > PR drafting/publishing skills) plus one bullet in the Gen-AI section of >> > contributing-docs. The regular PR template is unchanged - PRs created by >> > humans without agents are not affected. The PR itself is "dog-fed" - it >> > uses the new format. >> > >> > As this is straightforward and uncontroversial, I plan to merge the PR >> > before the lazy consensus ends. If there are objections, we can simply >> > revert it. >> > >> > The lazy consensus will end on Monday, 28 September 2026, 15:00 CEST, >> > unless there are objections. >> > >> > J. >> > >> > --------------------------------------------------------------------- >> > To unsubscribe, e-mail: [email protected] >> > For additional commands, e-mail: [email protected] >> > >> >> >> --------------------------------------------------------------------- >> To unsubscribe, e-mail: [email protected] >> For additional commands, e-mail: [email protected] >> >>
