My AI doppelganger tells me what to eat: 100 days with a metabolic digital twin
The Verge senior reporter Victoria Song describes roughly 100 days using Twin Health's digital twin program: each day she logs data from a smart scale, a blood pressure cuff and a Dexcom G7 continuous glucose monitor into the app, which color-codes foods red, orange or green to predict her glucose response, alongside bimonthly check-ins with a human coach. She reports her follow-up labs improved — triglycerides fell by 70 mg/dL, her coach said her metabolic health score rose nine points, and hs-CRP dropped notably — though she remained insulin resistant, saw no meaningful weight loss, and did not manage to taper her medications as hoped. She notes the service is not available for widespread download and is currently offered only through corporate or private health plans, and that it demands a continuous stream of sensitive health data, a burden she found tiring; the figures come from her own and her coach's reports, not a controlled trial.
It shows what an AI digital twin looks like in low-stakes metabolic care: plausible benefits in exchange for relentless personal-data logging and a great deal of trust.
My AI doppelganger tells me what to eat
In return for feeding a ‘digital twin’ my metabolic data, it predicts how I’ll react to certain foods — and one day, maybe medicines.
by Victoria Song
Oct 14, 2026, 5:00 PM UTC


#Twin Health#Dexcom G7#Apple Watch#Quest Diagnostics#The Verge