Forecasting is Not Overrated and It's Probably Funded Appropriately

Ben S. pushes back against Marcus Abramov's claim that forecasting is overrated and overfunded, arguing that initial investments in platforms like Metaculus and Manifold have yielded extraordinary returns. These sites, now used by tens of thousands daily, provide free crowdsourced predictions on everything from mundane topics to consequential geopolitical events, improving decision-making at scale. The ROI is evident in the growth of billion-dollar companies like Kalshi and Polymarket, which built on early forecasting research, and in the expansion of the EA/rationality community by roughly 5%, with dozens of individuals taking the Giving Pledge as a direct result. Additionally, forecasting has been a key tool for AI safety advocacy, influencing policymakers and amplifying the impact of figures like Hinton and Bengio. Ben notes that while further funding at tens of millions might not be optimal, the initial spend was clearly justified.

Beyond direct financial returns, forecasting has created lasting epistemic infrastructure akin to Wikipedia or OurWorldInData, helping people think more concretely and quantitatively. It has also served as a recruiting pipeline for talent in AI safety and other fields, and is now being used to allocate resources for the 2028 U.S. election, particularly for Democrats aiming to hold the House. Ben acknowledges that measuring the value of such infrastructure is difficult, but the cumulative impact—from better individual decisions to systemic political influence—makes the investment a clear win. The post highlights that even with modest ongoing funding, the benefits of forecasting continue to compound, reinforcing its role as a critical tool for evidence-based decision-making in tech and policy circles.