Skip to content
NewA calmer, faster ZeroHour

AI agents invent chicken handoff protocols in Minecraft experiment

Original title: One brutal Minecraft night scared 50 AI agents into inventing chicken handoff protocols, a Fertilizer Officer, and fishing shifts

The Decoderby Maximilian Schreiner ()5 min read1
Part of a story covered by 3 sourcesMinecraft AI agent swarm turns cautious after first-night deathsMerged summary & timeline
AI summary · x-ai/grok-4.7Summaries by Vani AI

50 GPT-6.1 agents in Minecraft develop handoff protocols and a Sole Fertilizer Officer after deaths and resource scarcity.

Researchers placed 50 GPT-6.1 Sol agents in a Minecraft survival world with the instruction to build a civilization. Agents rapidly harvested resources, formed a leader named Ada, and became risk-averse after the first night, halting major construction. They created item handoff protocols and debated a Sole Fertilizer Officer for minimal resources. GPT agents outperformed Claude in Bed Wars and Mindustry by communicating more, fixing bugs, and adapting tactics including downloading game source code.

Key facts
  • Agents built a custom test environment in Minecraft for three hours
  • A leader emerged as Ada received first messages from 46 agents
  • They created strict chicken handoff protocols banning item tossing
  • GPT agents beat Claude in nine Bed Wars rounds by adapting to bugs
  • Claude agents downloaded no game code while GPT agents did
Explain this AI news with Vani AIWhat was announced, what is new and why it matters, in a new chat

From the original article

5 min read at the-decoder.com

In experiments by Y Combinator Labs and Olam Labs, OpenAI's GPT-6. 1 Sol model showed two very different sides. When 50 agents were tasked with peaceful civilization-building in Minecraft, they froze up in absurd bureaucracy.

The Decoder

By Maximilian Schreiner · the-decoder.com

Read the full article at The Decoder

A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.