AI Futures Model pushes full coding automation forecast to 2031

The updated AI Futures Model is one of the few serious attempts to turn AGI timelines and takeoff scenarios into a quantitative, end‑to‑end story. The headline result—a median shift for fully automated coding out to roughly 2031—doesn’t mean AI progress slowed in 2025. Instead, the authors conclude their earlier modeling probably overestimated how quickly partially automated AI R&D loops would compound into full automation.([blog.ai-futures.org](https://blog.ai-futures.org/i/182911449/timelines-and-takeoff-forecasts))

For the race to AGI, that’s a quietly important narrative correction. Frontier labs and many market participants have been operating on extremely aggressive 2027‑ish timelines for “AI can write all the code,” which drove some of 2025’s exuberance in funding, hiring, and infrastructure build‑out. A 3–4 year delay still leaves timelines short by historical standards, but it buys a sliver more time for safety research, policy, and institutional adaptation. Crucially, the model keeps post‑automation takeoff relatively fast and still lands median ASI around 2034, reminding us that a slightly later start doesn’t necessarily mean a slow finish.([blog.ai-futures.org](https://blog.ai-futures.org/i/182911449/timelines-and-takeoff-forecasts))