Will we run out of data? Limits of LLM scaling based on human-generated data

TL;DR

This paper examines the future limits of large language model scaling due to the finite availability of public human-generated text data, forecasting potential data constraints and exploring alternative strategies for continued progress.

Contribution

It provides a forecast of data availability constraints for LLMs and discusses possible solutions like synthetic data and transfer learning to sustain growth.

Findings

Models will reach data limits around 2026-2032 if trends continue.

Current data stock is finite and may constrain future LLM development.

Alternative methods could enable continued progress despite data limits.

Abstract

We investigate the potential constraints on LLM scaling posed by the availability of public human-generated text data. We forecast the growing demand for training data based on current trends and estimate the total stock of public human text data. Our findings indicate that if current LLM development trends continue, models will be trained on datasets roughly equal in size to the available stock of public human text data between 2026 and 2032, or slightly earlier if models are overtrained. We explore how progress in language modeling can continue when human-generated text datasets cannot be scaled any further. We argue that synthetic data generation, transfer learning from data-rich domains, and data efficiency improvements might support further progress.