Good QC for RL Data

Read my thoughts on Good QC for RL Data

This is a preview piece - for more writing and a state of data report, visit my substack . In January, I proposed a new definition for Type 1, Type 2 data, pending drastic need from the data industry on how to evaluate data quality. A conscious side-effect of the shift to longer horizon regiments is increased need for model-based QA, far beyond the body-shop capabilities of current day data companies. The progression of what data markets we entered first directly corresponded to how verifiable we could make each one. We filtered the hard domains out of the field at the infrastructure layer, fi

saved by

• Falah Rajput
• Freeman Irabaruta
• Tom Zheng
• Sean Lee

related reading

• State of Data (Jan 2026)seancai.com
• The Bitter Lesson - RL Environments Versionseancai.com
• Thinking about High-Quality Human Data | Lil'Loglilianweng.github.io
• DataRater: Meta-Learned Dataset Curationarxiv.org
• Our Problems · Proximalproximal.ai
• A World of Verifiable Domainsseancai.com
• Cheap RL tasks will waste compute | Mechanize, Inc.mechanize.work
• Your AI Product Needs Evals – Hamel's Blog - Hamel Husainhamel.dev
• RL Environments and RL for Science: Data Foundries and Multi-Agent Architecturesnewsletter.semianalysis.com
• The bitter lesson of LLM evalsparsed.com
• [2605.12474] Reward Hacking in Rubric-Based Reinforcement Learningarxiv.org
• RL Pet Peeves Part 1 · Aurielaurielws.github.io