andy jones

andy jones personal site

me I'm an engineer-researcher at Anthropic, where I've moved between projects as demand dictates. So far that's meant building much of our early RL infrastructure, our sandboxing infrastructure, leading the multimodal effort and working on big-model training. The best way to get in touch with me is via email! writing indie Horses: AI progress is steady. Human equivalence is sudden. Debugging Reinforcement Learning Systems: Debugging reinforcement learning implementations, without the agonizing pain. A Clearer Proof of the Policy Gradient Theorem: A policy-gradient proof with less…

saved by

• Benedict Neo
• Timothy Kostolansky
• Nathan Chen
• nathan karri

related reading

• Andy Matuschakandymatuschak.org
• LYDIAlydia.ml
• Andrew Gordon Wilsoncims.nyu.edu
• Recommendations for Technical AI Safety Research Directionsalignment.anthropic.com
• Tips and Code for Empirical Research Workflows — AI Alignment Forumalignmentforum.org
• Core views on AI safety: When, why, what, and how \ Anthropicanthropic.com
• Debugging Reinforcement Learning Systemsandyljones.com
• Non_Interactive – Software & MLnonint.com
• Andi Pengandipeng.com
• Automated Weak-to-Strong Researcheralignment.anthropic.com
• Vincent Huangvvhuang.com
• Welcome!boydkane.com