Anthropic Economic Index report: Learning curves

The Anthropic Economic Index report examines learning curves associated with Claude adoption. It finds that high-tenure users have developed habits and strategies that allow them to better harness the model's capabilities. More experienced users attempt higher-value tasks and achieve higher success rates in their conversations. Users with six months or more of experience have 10 percent fewer personal conversations. Inputs from this group reflect a 6 percent higher education level. The higher-tenure users show a 10 percent higher success rate, an association not explained by task selection, country of origin, or other factors. Model selection aligns with task value, the report states. Among paying Claude.ai users, the Opus model is used 4 percentage points more than average for coding tasks and 7 percentage points less than average for tutoring-related tasks. The pattern is roughly twice as pronounced among API users. The findings raise the possibility that effective AI use requires complementary skills acquired through experience. If so, the report suggests, benefits from early adoption could prove self-reinforcing. The higher success rates may reflect either the sophistication of early adopters or learning-by-doing.