Top AI experts badly underestimated how fast the field is moving, study finds
A Forecasting Research Institute study finds experts badly underestimated the speed of AI progress.
The Forecasting Research Institute compared forecasts collected since mid-2022, including a 339-person LEAP expert panel, with later AI outcomes. AI reached International Mathematical Olympiad gold in July 2025, roughly five years before the median expert forecast and ten before superforecasters; experts had given that result an 8.6 percent probability. FRI says a virology troubleshooting benchmark was likely matched by April 2025, while a small trial found language models did not improve biological lab-task completion, 5.2 percent versus 6.6 percent with internet alone. Revenue forecasts were also low: the highest group median was about $25 billion ARR, against roughly $65 billion reported in July 2026 and about $100 billion FRI cites for Anthropic in September 2026.
- IMO gold in July 2025 beat median expert forecasts by about five years.
- Experts assigned only 8.6 percent odds to that Olympiad result.
- FRI says a top virology benchmark was likely matched in April 2025.
- Cited AI-lab revenue far exceeded forecasts of $16–25 billion ARR.
- Organizations
- Forecasting Research InstituteAnthropicOpenAI
From the original article
3 min read at the-decoder.com
Leading AI experts have consistently underestimated how fast AI is advancing, according to the Forecasting Research Institute. AI reached gold-medal level at the International Mathematical Olympiad five years ahead of the median expert forecast, and Anthropic's annualized revenue is about five times what experts predicted. But forecasts for real-world uses like self-driving cars paint a more mixed picture.
By Matthias Bastian · the-decoder.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.