News for hackers. Stuff that matters.
Re:Why the push for Agentic when models can barely follow a ... by alwahi on Tuesday October 14, 2025 @05:12AM (#45577849)
if claude generates the tests, runs those tests, applies the fixes without any oversight, it is a very "who watches the watchmen" situation.
Re:Why the push for Agentic when models can barely follow a ... by ErikBjare on Tuesday October 14, 2025 @05:23AM (#45577914)
If you generate the tests, run those tests, apply fixes without any oversight, it is the very same situation. In reality, we have PR reviews.
Re:Why the push for Agentic when models can barely follow a ... by skydhash on Tuesday October 14, 2025 @07:47AM (#45578873)
Is it? Stuff like ripgrep, msmpt,… are very much one-man project. And most packages on distro are maintained by only one person. Expertise is a thing and getting reliable results is what differentiates expert from amateurs.
Re:Why the push for Agentic when models can barely follow a ... by fragmede on Tuesday October 14, 2025 @05:13AM (#45577856)
Gemini?
Re:Why the push for Agentic when models can barely follow a ... by gmb_uk on Tuesday October 14, 2025 @05:27AM (#45577937)
Good lord, that would be like the blind leading the daft.
Re:Why the push for Agentic when models can barely follow a ... by vidarh on Tuesday October 14, 2025 @08:51AM (#45579422)
That is true, so don't give it entirely free reign with that. I let Claude generate as many additional tests as it'd like, but I either produce high level tests, or review a set generated by Claude first, before I let it fill in the blanks, and it's instructed very firmly to see a specific set of test cases as critical, and then increasingly "boxed in" with more validated test cases as we go along.

E.g. for my compiler, I had it build scaffolding to make it possible to run rubyspecs. Then I've had it systematically attack the crashes and failures mostly by itself once the test suite ran.