API

Where the book uses it

• Chapter 2 — The Discovery That Changed Everything

  How a formula about compute and capability became a bill: the scaling laws, the first supercomputers built to test them, and why the money came before the paper.

Sources that use it

• Anthropic Economic Index, “Economic primitives,” 15 January 2026 (computer and mathematical work about 40 percent of Claude.ai conversations at its March 2025 peak and 34 percent by November; 44 to 46 percent of first-party API traffic; its definition, “effective AI coverage as the percent of a worker’s day that can be performed successfully by Claude”; of data entry workers, “This is because although only two of their nine tasks are covered, their largest task—reading and entering data from source documents—has high success rates with Claude. AI excels at what they spend most of their time doing”; “Adjusting productivity estimates for task reliability roughly halves the implied gains, from 1.8 to about 1.0 percentage points of annual labor productivity growth over the next decade” and “all signs suggest that reliability over increasingly long-running tasks will improve”). In the Sources
• Anthropic Economic Index, “Learning curves,” 24 March 2026 (February 2026 sample: coding 35 percent of Claude.ai conversations; top ten tasks down from 24 to 19 percent of traffic; personal queries up from 35 to 42 percent; the share of Computer and Mathematical tasks since August 2025 “has increased by 14% in the API and decreased by 18% in Claude.ai”; “the top 20 countries account for 48% of all per-capita usage, up from 45%”, August 2025 to February 2026; “Low-adoption countries fell slightly further behind.”). In the Sources
• Anthropic Help Center, “What is the Enterprise plan?” (accessed 30 August 2026: a seat fee plus usage billed at API rates; seat-based Enterprise agreements begun before 1 January 2025 transition at renewal); first reported by The Information, “Anthropic Changes Pricing to Bill Firms Based on AI Use as Demand Jumps,” 14 April 2026; OpenAI publishes a ChatGPT Enterprise token rate card. In the Sources
• DeepSeek API pricing, as of 29 August 2026 (V4 Pro $1.32/$3.96 per million input/output tokens at peak; half price off-peak). In the Sources
• GitHub (Mario Rodriguez), “GitHub Copilot is moving to usage-based billing,” company blog, 27 April 2026, effective 1 June 2026 (“GitHub has absorbed much of the escalating inference cost behind that usage, but the current premium request model is no longer sustainable”; credits “consumed based on token usage, including input, output, and cached tokens, according to the published API rates for each model”; “Base plan pricing is not changing”). In the Sources

And 32 more: search the Sources.

Related terms

• Docket
• Floating-point operation
• Loss
• Power law
• Pre-training
• Reasoning model
• Reinforcement learning
• Scaling laws
• Transformer