Nina Teicholz’s The Big Fat Surprise opens with a tantalizing premise: What if everything we’ve been told about fat in our diets is wrong? With bold assertions and a hefty dose of investigative bravado, Teicholz promises to upend decades of nutritional orthodoxy. And for a moment, she almost pulls it off. Almost. But upon closer inspection, The Big Fat Surprise reveals itself as a buffet of cherry-picked studies, logical leaps, and sweeping generalizations, leaving the reader less with a paradigm shift and more with a lingering sense of indigestion.

Gary Taubes’s Good Calories, Bad Calories is a highly influential and contentious book that challenges decades of conventional dietary advice. Taubes argues that the widespread belief that dietary fat, particularly saturated fat, causes heart disease and obesity is based on flawed science, and he proposes that carbohydrates are the true culprits behind chronic diseases. While the book’s depth and scope are impressive, its criticisms of mainstream dietary guidelines and its promotion of low-carbohydrate diets have sparked significant debate. Below, I provide an in-depth critique of the book, emphasizing its methodological weaknesses, selective use of evidence, and overstated claims — because some of this stuff is just too rich to pass up without commentary.

In previous posts, we’ve established the role of saturated fats in elevating LDL cholesterol as well as the causal role of elevated LDL cholesterol in heart disease, contradicting false claims made by a number of journalists, bloggers, YouTubers, and crank doctors that have tried to confuse the public regarding the relationship between saturated fats, LDL cholesterol, and heart disease. Using GPT-4o, we now explore the precise mechanisms by which saturated fats in the diet elevate LDL cholesterol.

A number of journalists, bloggers, YouTubers, and crank doctors have tried to confuse the public regarding the relationship between LDL cholesterol and heart disease. Mendelian randomization is another method to investigate the relationship between LDL cholesterol and heart disease. Here we use GPT-4o to explain Mendelian randomization and to look at the evidence from Mendelian randomization for LDL cholesterol playing a causal role in heart disease.

In my last post, “What Causes Heart Disease?,” we saw that the most rigorous scientific evidence still supports the fact that elevated LDL cholesterol is the primary driver of heart disease — despite the fact that some profit-motivated journalists, bloggers, and YouTubers claim otherwise.

With so much confusion, controversy, and deliberate deception surrounding the primary causes of heart disease, I decided to put the question to four AI models.