Description du poste
Response to Forecasting is Way Overrated, and We Should Stop Funding It
disclaimer
I'm a part-time RA at the Forecasting Research Institute and these are my personal views.
tl;dr
Forecasting platforms backed by EA funders already inform and influence policy-makers on issues of concern to the EA community, and are being used by businesses responsible for managing risk at scale.
However, information consumers can obtain most of the value generated by forecasters without directly interacting with (or paying) the forecasters and forecasting platforms that generate that value.
This missing feedback loop makes it difficult for forecasting organizations to quantify where and how they add value. In the absence of easily quantifiable impact or a clear thematic mandate, the path of least resistance is improving accuracy. Accuracy provides status games for participants, but I view it as a necessary(ish) but not sufficient property for impact.
Beyond accuracy, businesses, governments, and the general public need coverage, latency, explainability, and extensibility with proprietary data sources.
Forecasting organizations, with the exception of organizations focused on forecasting questions of interest to EA funders, should probably dedicate significantly more effort towards solving for these other components of the forecasting supply chain, rather than continuing to optimize for accuracy.
asset managers
Asset managers with roughly $200 billion in combined AUM have cited prediction market and forecasting venue probabilities in their allocation decisions.
A few examples:
• National Bank Investments ($115 billion AUM, subsidiary of National Bank of Canada) cited Polymarket's 33% probability of U.S.-Iran ceasefire by end of April 2026 and 26% probability of normal Hormuz traffic by same date as reasons to maintain a cautious portfolio posture (cutting equities to neutral, adding bonds).• Westwood Group ($17.1 billion AUM) incorporated prediction market probabilities, including Ma