News for hackers. Stuff that matters.
Re:AI Blindspots – Blindspots in LLMs I've noticed while AI ... by ezyang on Wednesday March 19, 2025 @08:38PM (#43418792)
One problem with keeping the changes separate is the LLM usually wants to test the code with the incremental new changes. So you need a working tree that has all the new changes. But then... why not use the real one?
Re:AI Blindspots – Blindspots in LLMs I've noticed while AI ... by danenania on Wednesday March 19, 2025 @09:09PM (#43418981)
Plandex can tentatively apply the changes in order to execute commands (tests, builds, or whatever), then commit if they succeed or roll back if they fail.