[kritiko]

The point you're making seems similar to the argument that GMOs have a lot in common with earlier selective breeding techniques.

I think the difference is that while historically grandma or restaurant chefs were biased towards repeatable, appetizing products, they weren't using focus groups and industrial/scientific methods to get there. I.e. you can't send your grandma to General Mills and have her create Spaghetti-Os.

There seems to be medical consensus that "ultraprocessed" food, as squishy as that term is, is obesogenic. Part of that may be influenced by the marketing and marketing science behind the food as well, but there seem to be human and animal studies that support that our food has changed.

[crazygringo]

> they weren't using focus groups and industrial/scientific methods to get there. I.e. you can't send your grandma to General Mills and have her create Spaghetti-Os.

But that's exactly what I'm pushing back against -- you've never needed focus groups or fancy methods to figure out what the tastiest amount of sugar or salt or fat is for a recipe. It's remarkably easy to figure out on your own. Focus groups just serve as objective evidence over the opinion of a single chef, and can reveal that different markets prefer slightly different ratios (e.g. Brazilians prefer sweeter desserts than Americans). And Spaghetti-O's are just spaghetti in tomato sauce in a different shape, for those who prefer their tomato sauce sweeter rather than saltier, like kids do -- and of course grandma made spaghetti.

> There seems to be medical consensus that "ultraprocessed" food, as squishy as that term is, is obesogenic.

There isn't even remotely consensus on what the causality is here -- it's one of the thorniest problems to untangle. But the idea that tastier food is contributing seems extremely far-fetched. Look at sugary soft drinks first of all, look at decreased physical activity, look at changing cultural norms around portion sizes and body weights, look at the reduction of fats that have raised carbohydrate intakes.

Snickers bars have been around since 1930. People in the 1950's were eating every kind of supposed "hyperpalatable food" and obesity was a minor issue. So the idea the flavor is somehow new or responsible for health issues just doesn't hold up to scrutiny.

[kritiko]

Are you familiar with Kevin D. Hall's Ultraprocessed food study?

https://www.cell.com/cell-metabolism/pdf/S1550-4131(19)30248...

They did it in a metabolic ward, so about as controlled as you can get in a diet study. Created 2 diets that were matched - so caloric density and macronutrients of the food were the same. Let people eat as much of the food as they wanted, and the ultraprocessed group consumed 500 calories per day more.

[crazygringo]

Thanks for linking that -- it's a rigorous study for what it attempts.

However, there are two fatal flaws. The first it that it neither controls for nor even measures the glycemic index of the two diets, and the second is that it neither controls for nor even measures the vitamins and minerals in either diet.

The major critique around dividing foods into "unprocessed" vs. "ultraprocessed" is that high-GI foods like rice and raisins are considered unprocessed, while lower-GI foods like pasta (and zero-GI foods like hamburger patties) are considered ultra-processed. And in this study, the authors could assemble whatever GI they wanted for each diet, and never revealed it.

But there's a clue -- in the study, the additional calorie intake and weight gain in the "ultraprocessed" diet is entirely consistent with a vastly higher GI of that diet, which is almost certainly indicated by the nearly doubled non-beverage energy density (2.147 vs 1.151, Table 1).

Which means the study is actually entirely consistent with the idea that higher-GI foods lead to weight gain, and that "ultraprocessing" itself is entirely irrelevant.

[kritiko]

The study’s author, KevinH_PhD is a good follow on Twitter and generally responsive to his critics there:

“Glycemic index was calculated to be ~52 for both diets with respect to oral glucose. There were no significant differences in CGM determined mean glucose or glycemic variability as assessed by glucose CV.”

From this thread - https://twitter.com/KevinH_PhD/status/1536782331628822528