Verge reviewer logs 100 days on a Twin Health digital twin

The Verge's Victoria Song spent 100 days feeding a digital-twin program called Twin Health with data from a Dexcom G7 continuous glucose monitor, a smart scale, a blood pressure cuff and an Apple Watch, letting an AI model color-code her meals green, orange or red. Baseline labs through Quest Diagnostics flagged nonalcoholic fatty liver disease, insulin resistance despite an ideal HbA1c and fasting glucose, and high triglycerides.

Follow-up labs told a flattering story. Writing in her Optimizer column at The Verge, Song reports her metabolic health score rose nine points, triglycerides fell 70mg/dL, and hs-CRP, a marker for inflammation, dropped significantly. Liver enzymes returned to normal with one exception. She remains insulin resistant.

Song pairs the numbers with the work it took to produce them. She notes that "tracking fatigue was frequent feedback from Twin Health users," and the regimen required bimonthly coaching with a Twin Health coach named Amanda on top of continuous glucose readings, smart-scale weigh-ins and daily meal logging across the trial. Weight did not meaningfully move, but she ran an uninterrupted 5K and resumed strength training.

She closes on where the digital-twin idea goes next, naming cancer care as the harder case: privacy, trust and the data burden on patients all climb when the twin is modelling a tumour instead of lunch. Two of the researchers we track had already circulated the column.