Is this machine playing?
cs.AI, cs.LG
Submitted: 2026-10-05
Updated: 2026-10-05
Comments: 13 pages of main text, 17 figures
Code: https://github.com/nacloos/clawblox2https:
License: http://creativecommons.org/licenses/by/4.0/
The gist: We placed a modern AI coding assistant in an unintended role: as the mind of a body on an unknown digital island.
Abstract
We placed a modern AI coding assistant in an unintended role: as the mind of a body on an unknown digital island. With only a minimal instruction mentioning no specific task, reward, or activity, the machine started animating its virtual body. Across thirty-hour runs, the embodied AI agent climbed hills, stacked blocks into towers, drew mandalas, reinterpreted sports, ran experiments on the physics of its world, and learned techniques that later expanded what it could accomplish. These activities recurred across thirteen agents but diverged into distinct histories. We examine whether this behavior satisfies classical criteria for play and ask whether play can become a mode of machine development.
Sources
- GEPA: Reflective Prompt Evolution Can Outperform Reinforcement Learning
- Project Sid: Many-agent simulations toward AI civilization
- Constitutional AI: Harmlessness from AI Feedback
- International AI Safety Report 2026
- Large-Scale Study of Curiosity-Driven Learning
- Exploration by Random Network Distillation
- OMNI-EPIC: Open-endedness via Models of human Notions of Interestingness with Environments Programmed in Code
- MineDojo: Building Open-Ended Embodied Agents with Internet-Scale Knowledge
- Open-Endedness is Essential for Artificial Superhuman Intelligence
- Eureka: Human-Level Reward Design via Coding Large Language Models
- LLM Agents Beyond Utility: An Open-Ended Perspective
- Open-Ended Learning Leads to Generally Capable Agents
- Training language models to follow instructions with human feedback
- Scaling Instructable Agents Across Many Simulated Worlds
- Voyager: An Open-Ended Embodied Agent with Large Language Models
- Paired Open-Ended Trailblazer (POET): Endlessly Generating Increasingly Complex and Diverse Learning Environments and Their Solutions
- Large Language Models as Optimizers
- TextGrad: Automatic "Differentiation" via Text
- Absolute Zero: Reinforced Self-play Reasoning with Zero Data
Related papers
- MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
- Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
- Agents' Last Exam
- The Clinician's Veto: Navigating Trust, Liability, and Uncertainty in Autonomous AI Prescribing
- MindHelper: Closed-Loop Embodied Mental-State Reasoning for Precision Intervention
- Incumbent Advantage: Brand Bias and Cognitive Manipulation Dynamics in LLM Recommendation Systems