Anthropic's sixth Economic Index report, Cadences, tracks Claude usage by hour of day

Anthropic has published the sixth installment of its Economic Index, subtitled "Cadences," shifting the research series from broad occupational trends to the hour-by-hour rhythm of how people actually use Claude across a day and week.

What's new

The new methodology moves beyond the coarse, weekly-aggregated usage snapshots of earlier reports. Anthropic writes that the update lets researchers "sample data at a higher rate, allowing us to view usage patterns down to the hourly level," combined with conversation-level classification of what people actually walked away with (an explanation, a piece of code, a plan) and a survey of roughly 9,700 Claude users whose self-reported answers were linked to their real usage through a privacy-preserving matching system.

That hourly resolution surfaces concrete daily rhythms: people ask for news at 7 a.m. local time, sleep advice clusters in the hours before dawn, and recipe requests are "2.3 times more frequent at 6 p.m. compared to the average." Work-related queries taper off on weekends, though less sharply among the highest-paid occupations, and tax-related requests spike around filing deadlines. A separate chapter breaks down what kind of output different products actually deliver: chat-style products skew toward explanations, while Claude Code sessions skew toward working code.

On the survey side, Anthropic reports that people who delegate more of their work to Claude are also the most optimistic about their own labor-market prospects: the report states that "people who delegate to Claude the most are the most optimistic about their future labor market outcomes, and feel their skills are growing in value." More than a third of the surveyed workers expected their job responsibilities to change significantly within twelve months, while only around one in ten rated their own job loss as likely.

Context

The Economic Index launched as Anthropic's recurring attempt to measure AI's real economic footprint from usage data rather than survey speculation alone, previously covering topics like which occupations show the heaviest automation-versus-augmentation split. Cadences is a methodological upgrade to that series rather than a new product: the value is in resolution (hourly instead of weekly) and in pairing raw usage logs with a matched survey sample, letting Anthropic connect what people do with Claude to what they say about their own job security and skill growth.

Why it matters

Most public data on AI's labor-market effects comes from either usage telemetry (what people do) or surveys (what people say), rarely both linked to the same individuals. By matching real usage patterns to self-reported career sentiment, Anthropic is building a dataset that speaks more directly to the automation-versus-augmentation debate than either signal alone — and the finding that heavier delegators feel more secure, not less, cuts against a simple narrative that AI use correlates with anxiety about being replaced by it.