Half the world population could be diabetic or insulin resistant by 2050 | Hacker News Reader

Half the world population could be diabetic or insulin resistant by 2050

fortune.com

12 pointsBrajeshwar8 comments

> “The good news is that we have all the tools we need today to help make the world metabolically healthy,” Hashemi added, pointing to AI-enabled technology that allows users to know the amount of protein, fiber, carbs and the like in a food—and their likely blood sugar response to it—just by taking a picture of it.

Knowledge of that generally isn't the problem though. It is that dietary change is fundamentally too difficult for most people to psychologically pull off day after day.

For a lot of people it’s also financially or physically impossible (talking about food deserts, free time in the day, etc) to make the necessary dietary changes day to day

no one is required to eat above bmr? how can less food be impossible?

That’s just not how diabetes works. Eating healthier is just more expensive. If we’re talking about hitting raw calorie count, it’s cheaper to go with the stuff that is packed full of synthetic sugars (sugars that will increase your risk of diabetes).

People love to point at AI solutions when databases are a thing.

The AI tool goes further than a database, so you can in theory just take a photo of food and get the nutrient info. It is definitely more convenient. Accuracy is an issue, but it does have value from a user perspective. However, the main users are athletes who always tracked their macros.

The issue is more fundamental. They are not identifying the right problem that needs solving. For most people, awareness of what is bad for them is not the issue.

Knowledge was definitely a problem for me even after extensive education on diabetes. It wasn't until I started looking at the results of my glucose monitor multiple times per day and associating the spikes with specific meals that I found the motivation to make the necessary changes. It was usually (but not always) the foods that I expected to be a problem, but I didn't really emotionally understand it, or the magnitudes, until I started collecting the data. Since then I've removed so many of the problem foods from my diet that my CGM is mostly a boring flat line and not worth the expense of wearing it all the time.

This article is not about diabetes; it is about AI. Call me a skeptic, but I cannot believe that AI will be an effective defense against an onslaught of unhealthy foods.

Even if this AI could provide a useful service, I am sure there would be major conflicts of interest. If a corporation has a large user-base sending them pictures of every meal, they are definitely going to be mining and selling that data. And the major food companies (who largely got us into this health mess) will hone their targeted advertising.

Reply on news.ycombinator.com