[hinkley]

We now know that the intestines have 'taste buds' that can detect sweetness.

There's a theory, and I don't know how tested it is at present, that the body uses these detectors to control the digestion process. Within this theory, the presence of the artificial sweeteners may encourage the body to work harder to scavenge calories from your digestive tract.

The upshot of this model is that if you eat a burger with a glass of water, you absorb X calories. If you eat the burger with a diet Sprite, you might be absorbing X + Y calories instead.

Remember, calorie counts for food are based on measures of caloric value of a unit of food minus the caloric value of what the average person excretes. If we were furnaces instead of meat we would get more calories from our food. And if you have odd digestive microbes you may be absorbing more or less calories from your food.

[jwally]

So I could be totally off, but I think the calories listed on food are derived from the Atwater factors of Carbs:4kc/g, Protein:4kc/g, and Fat:9kc/g.

[hinkley]

The caveats section covers a lot of variations, between food samples and individual digestive tracts.

I've you've ever been diagnosed with anything that isn't mainstream, you're familiar with how much the medical community likes to stuff pegs of any shape into their favorite receptacles. A patient with ideopathic symptoms could be lying to you and sneaking food in the parking lot (I've known a couple of those), or they could be an anomaly. Singular or a few percent of the population. The world of genetics is vast and a handful of rare conditions can net you several examples within your Dunbar number.

We have a lot of people these days who don't seem to correspond directly to calories in/calories out. And unless they're absorbing energy from the universe, then something about our assumptions is broken.

At work I often find myself having to remind people that if your assumptions tell you that an event must be impossible, then it's not your eyes that are wrong but your assumptions. Half of debugging is being able to efficiently name your assumptions and sort them by probability x difficulty of verification.

If people aren't losing weight on diet soda, we need to be dismissing rules of thumb and directly testing patients instead of shrugging and saying, "exercise more, scrub." Which is a polite characterization of how medical people treat my obese friends.