Yup, I just put butter on cheese and then I eat it. I think the butter will actually increase the amount of insulin secreted per gram of protein, just like it would with carbohydrate--but I end up eating less grams of protein, so it ends up being less insulinogenic in the end.

I've wondered if things would work as well if I put the cheese and butter through a food processor or something. If you take a mouse, and feed it carbohydrate, fat and protein separately, it's not fattening, and usually the mouse will choose a high-fat diet, depending on the strain. But if you make a high fat chow, with a little sugar, the animal gets fat. So what happens when an animal prefers fat--but isn't offered a pure fat source?