Lucas Nussbaum <[email protected]> wrote on 23/07/2026 at 21:12:44+0200:
> On 23/07/26 at 18:00 +0200, Pierre-Elliott Bécue wrote: >> I'll leave aside the fact that this classic behaviour brings to ~5x >> tokens consumption just because people are too lazy to reread and do >> prefer the models to fix their own crap based on CI outcome (or the >> fact that sometimes they modify the tests to make them pass rather >> than fix the code). > > AI put aside, this sounds a lot like "CI is for lazy programmers; real > programmers just reread their code to confirm that it will work > perfectly". I tend to reread my code and run tests locally before pushing to a CI, yeah. That being said, your extrapolation is yours, and yours only. >> We have a right to expect better from people. Just because a ton of >> them do trust enough their agents to upload on their >> github/gitlab/whatever account is no excuse to decide that it should >> be the standard we adhere to. >> >> I'm sorry, but what I see here is what I would bluntly call >> accelerationist hype. I know you (and some others) are totally hyped >> by LLMs, to the point that you might maybe believe they achieve >> greater than what they do. >> >> I'm not. I'm no doomer either, I use these tools on a daily basis. >> But I believe the more automated and fast-paced a tool is, the more >> safety nets are required, and do not allow any LLM to sign or upload >> stuff for me. > > My points in this subthread are mainly that: 1) I don't think that a > GR is the correct place to do detailed technical design [of safety > nets]; 2) I think that it's better to ensure that people feel > responsible, and to let them decide which safety nets are necessary > for them, rather than arbitrarily decide that 'git commit' or 'git > push' should be excluded from the scope of agents. 1) I disagree 2) I noted your opinion. I also disagree with it. > On 23/07/26 at 19:17 +0200, Pierre-Elliott Bécue wrote: >> What I want to do is to demand those using these tools to not give up >> their skills and intellect in favour of letting an LLM do all the >> job. > > Which skill am I giving up by asking my agent to 'git commit && git > push' for me, instead of doing it manually? Do you still review all the atomic changes your LLM does? Do you check that it doesn't produce redundant and overly verbose code? Do you still evaluate if it answers to the algorithmic problem you're trying to solve? Most of the people I know working with a generative AI asking it to commit and push read less than half the code it writes. I already see the difficulties they have when they have to code on their own. Any easy path is a path the brain loves to take, and it has a price, whether you like it or not. I don't want to be there when the next generation will push software, as I fear it might be atrocious. Also, can you point me where in my draft ballot option did I mention forbidding "git commit"? To make it easier for everyone, I did not. > Note that for 'git push', my personal safety net would agree with you > (my agent cannot 'git push' because it cannot access the required > credentials). Therefore you should be fine with the restriction I put. > For 'git commit' however, I routinely ask my agent to 'git commit', > and find it very useful: after several interactions to design a > specific change, the agent is usually in a very good position to write > a detailed commit message capturing the various design choices that > were made during the discussion (while humans tend to be more lazy and > write shorter commit messages). And I don't see what I would gain > (except potential errors) from asking the agent to generate a commit > message, and then commit myself manually. -- PEB
signature.asc
Description: PGP signature
