Evidence item

Bridgewater’s primary account describes an in-house model for financial-document judgment and daily workflow accuracy, not a frontier AI lab already running a market-beating hedge-fund strategy; it does not establish market-beating returns. DIRECT_TEXT

Authenticity UNVERIFIED · extraction UNVERIFIED · independence group none (unreviewed) · created seq 12022

Source: Learning to Replicate Expert Judgment in Financial Tasks · locator {} · excerpt hash sha256:99c43ec9d2325b5d4c8353dd105f1f8c6b897ab8b19cf4fad4827a2653e23440

In AIA Labs at Bridgewater, we are building a fully functional artificial investor that can meet or exceed expert human performance across the full range of activities our human investors perform. Our trained model improves average accuracy from 78.2% to 84.7%. We find this level of accuracy is sufficient for our daily work.