This would sound insane to me from two years ago but I have recently started insisting on writing all my own commit messages and pull request descriptions. I do usually have an agent review them for factual accuracy, but not rephrase them.
It slows things down a bit, but in the best possible way. It has helped immensely to improve the depth of my understanding of the agent-generated code. When agents are doing everything its way too easy to “skim” diffs and not really absorb them.
I always prided myself on my technical writing, and commit messages and PRs were a great place to hone that skill. I found that I missed it and my work is better now I’ve reclaimed that part of my old job back.
A big win of LLMs in my book is the absolute reduction of commit messages with just “fixes” or “updates”. Commit messages have become more meaningful and useful, even if far from perfect.
We have one dev who uses LLMs to write the code, but still commits by hand. Most of his messages are of the type above, and none of them are useful.
I'm not sure that Claude's "Realigned the shape of the load-bearing ownership gate to reduce the blast radius of the design contract; confirmed, not assumed" is more meaningful than "fix".
> I'm not sure that Claude's "Realigned the shape of the load-bearing ownership gate to reduce the blast radius of the design contract; confirmed, not assumed" is more meaningful than "fix".
For PR/commit descriptions, I mainly use Claude Sonnet 4.5. It isn’t perfect, but it produces significantly less of this weird gibberish than 5.x models or even Opus 4.x do
I also use an iterative process in which it writes the description, I read it, and then either manually edit it or ask it to make changes
I use Astra at Very High, and shit is still bad. It doesn't actually understand anything, so it often says things which are clearly not needed to be stated. Recently, I've learned that I have very high standards for these things. For example, "fixes" as a commit msg just is NOT acceptable and would never fly where I work.
Astra is such a mixed bag. It makes some amazing reviews and sometimes architecture suggestions that I like. But it’s also lazy and will just make up things.
Have another model (or even another instance of the same model) review the output of the first.
Models will hallucinate. They are also quite good at spotting hallucinations in other models' output (with some more hallucinations thrown in). With a threshold for confirmation, and a few iteration loops, you arrive at a fixed point where every claim is supported.
A few things, but generally more chain of thought before generating a response. So the model is tuned to think more. Given that what it outputs for this task is a summary of its thinking, tuning it to think more will just make a more verbose, less useful commit message.
Tune your model parameters to what is right for the task, not the highest you can afford.
Honestly you probably want a model that has only been trained on language and literature. Nothing from online discourse.
And even then… writing is personal expression. Here people are talking about commit messages. That’s fine but AI doing writing for anyone and I WILL NOT READ IT unless it’s literally basic tech manual.
We read to hear and engage with people’s thoughts. If someone outsources that to AI then they should be shunned.
semiquaver · · focus · HN ↗
It slows things down a bit, but in the best possible way. It has helped immensely to improve the depth of my understanding of the agent-generated code. When agents are doing everything its way too easy to “skim” diffs and not really absorb them.
I always prided myself on my technical writing, and commit messages and PRs were a great place to hone that skill. I found that I missed it and my work is better now I’ve reclaimed that part of my old job back.
LeafItAlone · · focus · HN ↗
We have one dev who uses LLMs to write the code, but still commits by hand. Most of his messages are of the type above, and none of them are useful.
this_user · · focus · HN ↗
skissane · · focus · HN ↗
For PR/commit descriptions, I mainly use Claude Sonnet 4.5. It isn’t perfect, but it produces significantly less of this weird gibberish than 5.x models or even Opus 4.x do
I also use an iterative process in which it writes the description, I read it, and then either manually edit it or ask it to make changes
my-next-account · · focus · HN ↗
manmal · · focus · HN ↗
senderista · · focus · HN ↗
manmal · · focus · HN ↗
adastra22 · · focus · HN ↗
Models will hallucinate. They are also quite good at spotting hallucinations in other models' output (with some more hallucinations thrown in). With a threshold for confirmation, and a few iteration loops, you arrive at a fixed point where every claim is supported.
rrr_oh_man · · focus · HN ↗
my-next-account · · focus · HN ↗
mitxela · · focus · HN ↗
rrr_oh_man · · focus · HN ↗
oblio · · focus · HN ↗
oblio · · focus · HN ↗
adastra22 · · focus · HN ↗
Tune your model parameters to what is right for the task, not the highest you can afford.
jester997 · · focus · HN ↗
And even then… writing is personal expression. Here people are talking about commit messages. That’s fine but AI doing writing for anyone and I WILL NOT READ IT unless it’s literally basic tech manual.
We read to hear and engage with people’s thoughts. If someone outsources that to AI then they should be shunned.