A simpler AI timelines model predicts 99% AI R&D automation in ~2032

The headline date is an output of assumptions about research, not a measured deadline. METR hosts a forecasting note that links computing growth, improvements in software efficiency and the share of AI research work that can be automated. As more work is automated, humans concentrate on the remaining tasks, accelerating research within the model.

The author identifies two particularly consequential uncertainties: how fast software efficiency improves, and how those gains translate into automating real tasks. Several parameter estimates are explicitly provisional, and the model omits feedback from the wider economy and a separate treatment of research judgment. Its useful contribution is an inspectable account of what would produce rapid automation; the predicted year should be read with those dependencies attached.