One night of sleep can now forecast a hundred diseases, the headline says. Stanford's SleepFM is real: fed a full night of lab polysomnography, it ranks future risk across 130 conditions. Then read the fine print. The score is a concordance index, a ranking, not your probability of anything; the paper reports no calibration, no false-alarm rate. It was trained on people sick enough to be wired up overnight, not on you. Mignot said the most information came from "contrasting the different channels," the brain-wave channels your ring does not have. Your tracker infers sleep from pulse and movement, and SleepFM was never shown to work on its data. What the ring would hand you is a risk score you cannot act on, and a market ready to sell the fix. Keep the diseases, drop the caveats: that is the version headed your way.