News for hackers. Stuff that matters.

mitch98

mitch98 on Hacker News

Re:Show HN: Leaping – Open-source debugging with LLMs on Wednesday February 28, 2024 @07:27PM (#39545238)
Hi Adrien and Kanav! Love the idea. How do you "re-run the function that failed" if there are non-deterministic inputs (where a system like your initial concept would help)? Also do you have any examples of how your system can work with more complicated stack traces and codebases?