Scaling Laws for Economic Productivity: Experimental Evidence in LLM-Assisted Consulting, Data Analyst, and Management Tasks

TL;DR

This study empirically examines how scaling large language models impacts economic productivity, revealing that increased compute and algorithmic improvements significantly reduce task times and could enhance U.S. productivity by 20% in ten years.

Contribution

It introduces empirical scaling laws linking LLM training compute to productivity gains across professional tasks, based on a large preregistered experiment.

Findings

Each year of AI progress reduces task time by 8%.

56% of productivity gains are due to increased compute.

Productivity improvements are greater in non-agentic tasks.

Abstract

This paper derives `Scaling Laws for Economic Impacts' -- empirical relationships between the training compute of Large Language Models (LLMs) and professional productivity. In a preregistered experiment, over 500 consultants, data analysts, and managers completed professional tasks using one of 13 LLMs. We find that each year of AI model progress reduced task time by 8%, with 56% of gains driven by increased compute and 44% by algorithmic progress. However, productivity gains were significantly larger for non-agentic analytical tasks compared to agentic workflows requiring tool use. These findings suggest continued model scaling could boost U.S. productivity by approximately 20% over the next decade.