Measuring Time Horizon using Claude Code and Codex
Source
metr.org
Date
Why it matters
Evidence that polished coding-AI agentAn AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.Full definition → harnesses like Claude Code and Codex did not raise measured time horizon over simple scaffolds, which tempers assumptions about scaffold-driven gains.
Terms in this piece · Glossary
AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
