1,000 agents write 5,000 rule-based PyBot versions in pure Python, beating a 360M LLM
LoubnaBenAllal1 · x · 2026-11-16
HuggingFace researcher Loubna Ben Allal shared the PyBot experiment: a swarm of 1,000+ DeepSeek Flash agents was tasked with building a chatbot with no neural network — pure Python, if-else branches, regex patterns and lookup tables, echoing ELIZA from 60 years ago. The agents wrote and evaluated over 5,000 versions; the resulting PyBot beats a 360M-parameter LLM on SmolTalkBench (LLM-judged) while using one CPU core and about 1/10 of the memory. Code is open-sourced at huggingface/pybot.
More from coding & agent
- A zero-AI chatbot: 1,000 DeepSeek agents wrote 300,000 if-else rules — victormustar · 2026-11-16
- Open-source Swamp compiles repeated agent skills into workflows: 8x fewer tokens, 2x faster — udmrzn · 2026-11-16
- 67k-star diagram-design skill makes AI coding tools draw editorial-grade diagrams — adnan_hashmi · 2026-11-16
- Dev's take on AI workflows: feed the right data and build verification loops — iamsahaj_xyz · 2026-11-16
- Every tool is becoming an API for AI agents: a field guide to the agent stack — DEVIL_QI · 2026-11-16
- Alibaba open-sources OpenCodeReview, its internal AI code reviewer using ~1/9 the tokens of Claude Code — udmrzn · 2026-11-16