“It’s certainly a weird feeling to be outdone by several bots at this point.” This single quote from professional forecaster Ben Shindel encapsulates the seismic shift occurring in the world of prediction, after an AI from British startup ManticAI ranked eighth in a global competition, leaving many human experts in its wake.
The Metaculus Cup, a rigorous summer-long contest, became a stark illustration of AI’s rapid progress. Participants were tasked with forecasting 60 complex events, and ManticAI’s system consistently delivered more accurate predictions than scores of people who have dedicated their careers to this work. The experience of being surpassed by an algorithm is a new and unsettling one for many.
Shindel’s comment also points to the astonishing speed of this change. “We’ve really come a long way here compared with a year ago when the best bot was at something like rank 300,” he added. This exponential leap from mediocrity to the top tier in just one year is what makes the AI’s ascent so profound and, for some, so unnerving.
The AI’s success stems from a multi-agent system that can work on dozens of problems at once, relentlessly updating its forecasts with new information. This tireless analytical capability is something human forecasters, with their cognitive limits, simply cannot match. It’s not just about intelligence; it’s about persistence and scale.
While the world’s very best forecasters still maintain a slight edge, the feeling of being outdone is likely to become more common. The industry is now grappling with what this means for the future of human expertise. The emerging consensus is one of adaptation, where professionals will learn to work with AI, leveraging its power to augment their own skills in a new, hybrid model of forecasting.