They tested this with mice, one set would have unlimited fat, another unlimited sugar, and the third set the combination of sugar and fat. Surprisingly, the first two sets of mice would stop eating when they were full. Only the third set of mice would go on eating and eating..
What? What about avocados, coconuts, olives, artichokes, soybeans…
I'd more buy the argument that our body tries to maintain a certain ratio of fats to sugar (which is usually mediated by fiber), rather than that humans historically haven't "experienced" foods with both fat and sugar…
Lustig and other scientists believe that excess fructose consumption may be a key driver of many of the most serious diseases of today. These include obesity, type II diabetes, heart disease and even cancer."
See many, many studies about this on http://nutritionfacts.org
I still don't know how much I trust the claims that fat+sugar is particularly bad, but I suspect it may have to do with some kind of second order insulin response. Like, the presence of triglycerides in the blood cause fat cells to be more likely to respond to insulin, storing fat, instead of releasing it. I dunno.
Also keep in mind GI is falling out of fashion as a key indicator. Glycemic Load seems to be more indicative of poor health indicators. That is to say, that the area under the curve is potentially more important than the max height of the curve. I'd say limiting both is probably important as well. I also highly suspect that the slope of the curve is very important in how we perceive hunger, which is why we tend to feel hungry shortly after eating foods with fast carbs...
FWIW, at the end of the day, in order to keep my wife healthy, we ended up giving up fast carbs, and she had to keep her carb count low in general. Her basic formula for a snack was to add the grams of fat+ grams of protein, and if that was bigger than the grams of carbs it was safe to eat.
Sorry I don't have links to this stuff, this was like 5 years ago, and I'm not sure if I ever saved those bookmarks. I can say it was really hard to find credible data sets. There didn't seem to be anything from America (draw your own conclusions :) ), most of everything I found was published datasets from Australia and New Zealand. Also, don't take this for anything other than anecdata, do your own research. You'll find very quickly a lot of really bad nutrition information. I had to plow through most of it, and might have given up if it wasn't for my pregnant wife, coupled with a friend with a digital signal processing background pointing out engineering problems with most of the current thoughts on controlling diabetic blood sugar levels.