Meek Models Shall Inherit the Earth

TL;DR

This paper suggests that due to diminishing returns on compute scaling, smaller 'meek' AI models will eventually match the performance of larger models, impacting AI strategy and policy.

Contribution

It introduces a model showing diminishing returns to compute and argues that meek models will inherit AI capabilities, challenging the focus on scaling for performance.

Findings

Diminishing returns to compute reduce advantage of larger models.

Proxies like training loss correlate with capability measures.

Empirical data supports convergence of model capabilities.

Abstract

The past decade has seen incredible scaling of AI systems by a few companies, leading to inequality in AI model performance. This paper argues that, contrary to prevailing intuition, the diminishing returns to compute scaling will lead to a convergence of AI model capabilities. In other words, meek models (those with limited computation budget) shall inherit the earth, approaching the performance level of the best models overall. We develop a model illustrating that under a fixed-distribution next-token objective, the marginal capability returns to raw compute shrink substantially. Given current scaling practices, we argue that these diminishing returns are strong enough that even companies that can scale their models exponentially faster than other organizations will eventually have little advantage in capabilities. As part of our argument, we give several reasons that proxies like training loss differences capture important capability measures using evidence from benchmark data and theoretical performance models. In addition, we analyze empirical data on the capability difference of AI models over time. Finally, in light of the increasing ability of meek models, we argue that AI strategy and policy require reexamination, and we outline the areas this shift will affect.