Future of work

OpenAI's second report on research acceleration

Published
Source
OpenAI blog

Summary

OpenAI published a second report measuring how coding agents are changing its own research and development. By internal tallies, the median researcher now uses about $2,400 a day of agentic tools ($23,000 at the 90th percentile), 'agent workdays' per researcher have risen from 0.5 to 5.7, an elevenfold increase, and more than two-thirds of researchers peak at four or more concurrent agents. Success rates without human intervention vary sharply by domain, from 69% for communication tasks to 43% for building tasks, which the report flags as potential bottlenecks. Time spent on safety and alignment work almost doubled this year, the largest increase of any area.

Why it matters for our work

It is a rare quantitative account of where human-AI collaboration speeds up first and where it stalls, useful for setting adoption priorities in research organizations. All figures come from OpenAI's own measurements and surveys.

Translated from the Korean original. Summaries may be translated and edited. Commentary reflects our perspective; forecasts remain the source’s views.

Read original (opens in a new tab)