Have your own agent do first pass code reviews, they catch that stuff every time.

goldenCeasar a day ago | parent | next [-]

And then the PR owner's agent can fix it and then after some number of iterations you get a PR for a new and mysterious system.

▲

CuriouslyC a day ago | parent | next [-]

It's only new and mysterious if you don't have a tight spec and guardrails.

▲

angxiaobi a day ago | parent [-]

There seems to be a gap between the theory you're advocating...which I actually agree with and the practical execution in your own public projects which you talk about heavily.

I haven't been able to get any of your recently published projects (Scribe, volknut, a few others) running successfully on linux, whether following your documentation, using the tagged releases, or reverse engineering the true functionality and CLI arguments working from your provided source code I found to have wasted my time.

It's difficult to believe you when your own documentation is entirely wrong or 404'd.

I'd genuinely love to see your projects in action, since the concepts sound promising, but you need to back it up with shipping properly.

	▲	CuriouslyC 17 hours ago \| parent [-]
		There's a reason I haven't dropped release pages yet. These projects are under heavy development and I push to github to checkpoint mostly rather than release. I'm sorry if you feel like I've misled you, I've tried to be clear that what I'm sharing now is to understand what I'm doing, to give visibility, and it's not ready yet. I'm committed to delivering great software, and when I tell you it's ready you can rest assured that it will work. Understand that I'm one man working on 20 projects simultaneously, with 5+ under active development at any one moment, so release stabilization and cadence will take a little bit to lock in.

▲

theshrike79 a day ago | parent | prev [-]

AI PR tarpit, anyone? :)

▲

hamdingers a day ago | parent | prev | next [-]

> they catch that stuff every time

This is why AI code review continues to be mostly useless in my experience. If you instruct an LLM to make PR comments it will come up with some, every time, even when none are warranted.

▲

CuriouslyC a day ago | parent [-]

So give them a schema and get them to give a relevance score, then filter below a threshold. Works better than trying to get them not to reply.

	▲	hamdingers a day ago \| parent [-]
		Seems like sound advice, unfortunate that the developers of the review tools I've had brushes with did not take it.

▲

jitl a day ago | parent | prev [-]

I try to avoid participating in the bullshit economy as much as possible