#167 - Gary Taubes: Bad science and challenging the conventional wisdom of obesity

Episode Summary

Executive Summary: Peter Atiyah interviews Gary Taubes about his path from physics and investigative science journalism into nutrition, using his career to explain how bad science, groupthink, and weak controls can distort fields from particle physics to obesity research. Taubes argues obesity is better understood as a hormonal/fuel-partitioning disorder than simple calorie imbalance, and that the evidence base remains contested.

Main Topics: Taubes’s path from physics to investigative science journalism (Priority: 5/5): Taubes explains how Harvard/Stanford physics training, skepticism, and writing led him from astrophysics and aerospace toward journalism and a lifelong focus on exposing weak science. Pathological science and the importance of controls (Priority: 5/5): Using the Shroud of Turin, CERN, and cold fusion, Taubes argues that scientists often fool themselves by failing to understand background noise, choosing the wrong controls, and committing to conclusions too early. CERN, Nobel Dreams, and the sociology of discovery (Priority: 4/5): He recounts his reporting on Carlo Rubbia and the W/Z boson work, emphasizing how ambition, hierarchy, and selective listening can distort even elite physics collaborations. Cold fusion as a case study in premature certainty (Priority: 5/5): Taubes describes how Pons and Fleischmann’s claims spread rapidly, why replication failed, and how the episode reinforced his view that science must prioritize skepticism and rigorous background analysis. How epidemiology can generate false causal claims (Priority: 5/5): He criticizes observational public-health research for inferring causality from associations without adequate controls, using power lines and cancer as a key example. Nutrition controversies: fat, salt, and obesity (Priority: 5/5): Taubes traces how his reporting on salt and dietary fat revealed weak evidence behind public-health dogma and led him to question the low-fat consensus and the calorie-balance model of obesity. Carbohydrate-insulin model vs calorie balance (Priority: 5/5): He argues that obesity is driven by hormonal regulation, especially insulin, and that low-carb diets can improve weight and metabolic markers despite contradicting conventional energy-balance theory.

Key Arguments: Science advances by aggressively testing hypotheses against alternative explanations, not by collecting confirming evidence after a conclusion is already embraced. Many high-profile scientific errors come from inadequate controls and poor understanding of background noise rather than outright fraud. Physics has a useful theory/experiment split and high standards for replication; biology and medicine often merge theory and experiment in ways that increase bias. Observational epidemiology often mistakes correlation for causation because it cannot adequately control for confounding and selection bias. The obesity epidemic likely reflects a shift in diet composition and hormonal regulation, not simply people eating too many calories. Insulin is central to fat storage and fuel partitioning; obesity should be studied as a disorder of fat accumulation regulation, not just energy imbalance. Low-carb/high-fat diets have repeatedly shown clinical benefits in trials, challenging the idea that they are inherently dangerous or that saturated fat is the main culprit. The persistence of a dominant paradigm can suppress competing hypotheses and slow progress even when the prevailing model is failing in practice.

Data Points: Harvard/Stanford training: Physics major at Harvard; master’s in aerospace at Stanford - Taubes’s early academic background before journalism CERN collaboration size: About 150 physicists - Rubbia’s UA1 collaboration during the W/Z boson work Particle accelerator scale: About 4 miles in circumference - Taubes describing the CERN collider Detector scale: 3–4 stories high; mansion-sized - Size of the particle detectors used in the experiment Discovery threshold: 5 sigma (or more) - Taubes contrasts physics standards with weaker evidence thresholds in other fields Cold fusion replication window: Within 2–3 months it was clear the claim was wrong - Taubes’s account of the rapid collapse of the cold fusion claim Number of people interviewed for Bad Science: Around 300 - Research for the cold fusion book Power-line cancer article interviews: About 85 people - Reporting the Atlantic piece on epidemiology and power lines Dietary fat article interviews: 145 researchers and administrators - Reporting for the fat controversy piece DASH diet article timing: 1998 or 1999 - The first nutrition assignment that led Taubes into the field Atkins weight loss: 25 pounds in 6 weeks - Taubes’s self-experiment with Atkins Obesity epidemic onset: Between 1978 and 1991 - Taubes’s estimate of when obesity rates began rising High-fructose corn syrup adoption: 55% HFCS in soda by 1984 - He links beverage reformulation to the obesity era Nurses’ Health Study size: 110,000 nurses - Example of large observational epidemiology cohort Low-carb trials discussed: Five studies - Early clinical trials comparing Atkins-style diets with low-fat diets Book advance duration: Enough to live for 4 years; book took 5 years - Taubes on the economics of writing Good Calories, Bad Calories Obesity research history: Virtually no obesity research before 1930 - Taubes’s historical framing of the field Newberg experiment size: 6 or 7 people - Early obesity experiment used by Lewis Newberg NUSI-funded research groups: 2 groups - One group aligned with energy-balance thinking, one with carbohydrate-insulin model Clinical trials on ketogenic diets: Over 200 - Taubes notes the breadth of current keto research

Pivotal Quotes: "You must not fool yourself and you're the easiest person to fool." — Gary Taubes: Explaining the scientific mindset he learned from Feynman and applied to bad science "The science of things that aren't so." — Gary Taubes: His definition of pathological science and the core theme of his career "I think the problem here is that when you entered the field, there was a paradigm... that you never questioned it." — Gary Taubes: His explanation to obesity researchers about why the energy-balance model persisted

Implications: The episode argues that obesity science may still be operating under an incomplete paradigm. For listeners, it supports skepticism, better controls, and openness to low-carb approaches while the field works toward more rigorous, testable answers.