Show HN: A tiny C program where an LLM rewires its DAG while running

11 points - today at 1:43 AM

Source

Comments

youknowme123 today at 1:04 PM
This is nice, it reminds me of earlier work where DAG-based optimization was used to discover effective multi-agent architectures under supervised learning.