The data black hole at the center of AI

Intelligence is fundamentally defined by sample efficiency, yet current AI progress relies on massive data scaling rather than improvements in how models learn from limited information. While humans develop complex skills with relatively minimal exposure, frontier models require trillions of tokens and bespoke expert trajectories to achieve competence. This million-fold discrepancy persists because current scaling laws cannot bridge the gap between human learning and the "Frankenstein's monster" approach of stitching together vast datasets. Although training AI is significantly less efficient than human learning, the ability to amortize these costs across billions of sessions makes AI economically transformative for white-collar automation. Ultimately, the industry’s reliance on massive data black holes highlights a fundamental difference in scaling curves, suggesting that solving the sample efficiency problem remains a critical, unresolved hurdle for achieving human-like intelligence.