You’re absolutely right that most people don’t need to measure this data. That said, most people didn’t need to measure step count or heart rate or sleep and we

You’re absolutely right that most people don’t need to measure this data. That said, most people didn’t need to measure step count or heart rate or sleep and we’ve found utility in that. Realistically, the medical community is extremely gate-keeper oriented. The doctor shortage is entirely fabricated by organizations run by doctors. Some barrier to providing medical care is good, of course. Realistically, we don’t have a lot of data on the blood glucose levels of healthy people. The explosion of Fitbit’s and Oura rings and similar products gave us an explosion in data on basic vitals of millions of people during sleep. This can inform a ton of research, even if the quality of data is not as high as in a clinical setting. The difference between collecting 40 hours a week of data on 5 people vs 40k hours on 5k people (or more!) is huge in a research setting. I used to work doing ML to measure and improve athletic performance, and to measure and detect physiological strain and risks during physical exercise. Our limitation was always access to enough data to not overfit. Compared to data corpus’s using in “generative AI” or image recognition models, or any other number models that are now standard in industry, we’re still in the Stone Age.