Eric Jang: Building an Automated Researcher with Claude Code to Rewrite AlphaGo
michaelrzhang · x · 2026-08-13
Eric Jang shares his article As Rocks May Think, exploring how to reinvent workflows using modern coding agents in the AI era.
- Context: Since 2022, AI-generated videos are indistinguishable from reality, the global economy is re-organizing around AI scale-up, and chief among all changes is that machines can code and think quite well now.
- Automated Research: The author spent the last two months on a Claude Code bender, no longer needing to write code by hand. He has been implementing AlphaGo from scratch to catch up on foundational deep learning techniques and re-learn how to program with the full power of coding agents.
- Agent Orchestration: He set up Claude not only to write infra and research ideas, but also to propose hypotheses, draw conclusions, and suggest what experiments to try next. He created a /experiment command to standardize an "action" in the research environment into an automated loop.
Related event: Eric Jang Explores Using LLMs and Coding Agents to Reshape Workflows(2 posts)→
More from coding & agent
- Palantir Launches AIP Evolve for Autonomous LLM Optimization and Cost Reduction — downingARK · 2026-08-13
- Claude Generates Walkable 3D City Street in Browser Using Pure Code — prasenx · 2026-08-13
- Auto-Generating Cross Stitch Patterns Using MCP and OpenAI Codex — burhop · 2026-08-13
- Dev Integrates MCP with OpenAI Codex to Auto-Generate Cross-Stitch Patterns — burhop · 2026-08-13
- AI Coding Isn't Fully Autonomous: Complex Projects Require Rigorous Human Review — georgemillo · 2026-08-13
- Hands-on with Grok Bot: Booking Reservations via Mixed Voice Commands — elonmusk · 2026-08-13