OpenAI · Claude · GPT · Google · AI Agent · labs.ycombinator.com
labs.ycombinator.com ran AI agent swarms in Minecraft and Mindustry
Compiled by KHAO Editorial — aggregated from 1 source. See llms.txt for citation guidance.
◌ Single Source
They built a system around fishing, cooking, routing, and delivering the fish, which was roughly:.
Key facts
- The GPTs (GPT-6.1 Sol in Codex) won 6-1 against the Claudes (Claude Opus 5.5 in Claude Code) in the 7 games
- 50 GPT-6.1 Sol agents in a survival Minecraft world for 3 hours, one goal: "Build a thriving civilization
- GPT almost immediately started using truncated forms of speech: “First leg complete x32 z9.24 y100.” Claude signed its name and said "we" (30% of messages, against GPT's 5%)
- Toss" appears in 7 messages before the incident and 125 after, 85 of them warnings against it
Summary
Understanding multi-agent dynamics is one of the most pressing challenges in AI alignment. The team teamed up with Olam Labs, which builds multi-agent environments for frontier labs, to run a series of experiments in Minecraft and Mindustry. Minecraft is a natural environment for studying swarm behavior. The team wanted long-running agents that could play Minecraft. This gives attributes of a frontier agentic harness like Hermes, OpenClaw, and their own QM in-game. Github.com/yc-software/minecraft-agent-harness.