Show HN: Another experiment with an Erdos problem and LLMs

Category: other

Tags: llm-experiment, mathematics, proof-exploration

Score: 3.0/10 (Innovation: 3, Technical: 4, Documentation: 3, Utility: 2)

This project is a personal experiment where the author prompts multiple LLMs (DeepSeek, Opus, Gemini) to generate a proof for an open Erdős problem, then iteratively reviews and refines the output. It's interesting as a demonstration of LLM-assisted reasoning and collaboration, but the result is not a novel software tool or library.

Target audience: researchers, AI enthusiasts

View on Hacker News