News for hackers. Stuff that matters.
Re:Claude 3.5 Sonnet by sdwr on Wednesday June 26, 2024 @11:21PM (#40806867)
Are there any tools out there that expose GPT or Claude to a codebase, and let it write PRs (semi) autonomously?
Re:Claude 3.5 Sonnet by CGamesPlay on Wednesday June 26, 2024 @11:31PM (#40806916)
I know of Sweep <https://docs.sweep.dev>, but honestly the SOTA on SWE-bench, which is basically exactly what you're asking for, is only about a 25% success rate, so expect very mediocre results.
Re:Claude 3.5 Sonnet by nyellin on Thursday June 27, 2024 @12:00AM (#40807063)
Re:Claude 3.5 Sonnet by sdwr on Thursday June 27, 2024 @02:43PM (#40813652)
Thanks, tried it out - it's pretty cheap and pretty good. Being able to commit directly into multiple files simultaneously is great.
Re:Claude 3.5 Sonnet by tananaev on Wednesday June 26, 2024 @11:32PM (#40806918)
It's probably not going to be very good at handling it out of the box. It would require quite a bit of fine tuning.
Re:Claude 3.5 Sonnet by kleneway1 on Thursday June 27, 2024 @12:23AM (#40807196)
I’m working on adding Sonnet 3.5 to JACoB this week. So far it’s been very impressive. https://github.com/jacob-ai-bot/jacob
Re:Claude 3.5 Sonnet by afro88 on Thursday June 27, 2024 @04:37AM (#40808541)
Semi autonomously is the key. Aider and Plandex are both pretty good.
Re:Claude 3.5 Sonnet by campers on Monday July 01, 2024 @04:36AM (#40843753)
Another one is https://www.github.com/trafficguard/nous which provides a software dev agent that can find/clone repos, create a branch, search in a repo, then delegates to Aider for the code editing in the edit/compile/lint/test loop, and raise a merge request.