ParentFull threadsaagarjha·It’s not crazy at all, the blog post is clearly about human-guided optimization with LLMs rather than hillclimbing with an autonomous agent.View on HN