Peter Attia MD167 - Gary Taubes: Bad science and challenging the conventional wisdom of obesity
At a glance
WHAT IT’S REALLY ABOUT
Gary Taubes traces pathological science from particle physics to obesity research
- Taubes recounts his path from physics and aerospace to journalism, and shares stories of boxing, risk perception and learning to write under harsh editorial criticism.
- He describes how Carlo Rubbia's CERN experiment and the cold fusion episode illustrate pathological science, in which researchers commit publicly to premature signals and neglect their background and controls.
- He criticizes epidemiology and public health science for treating associations as causal, and discusses the conflict between honest uncertainty and the need to change behavior.
- He explains how a one-page assignment on the DASH diet led him to salt and dietary fat, then to the carbohydrate-insulin hypothesis, which he argues was sidelined by the history of obesity research, including lesion and ob/ob mouse experiments.
- Taubes and Peter Attia discuss the dispute over recent energy balance experiments and measurement methods, Taubes's frustration with how research is funded, and the wider acceptance of low-carb diets.
IDEAS WORTH REMEMBERING
5 ideasSignal-to-noise problems underlie bad science
Taubes argues that understanding your background and equipment flaws is the hard, thankless work of science, and skipping it can produce false discoveries without any fraud.
Public commitment to premature results locks in bias
He says scientists should first try to prove themselves wrong, but once they go public they start collecting supporting evidence and listening only to those who agree.
Choosing the right control matters
The Georgia Tech neutron detectors that responded to humidity, and pair-fed lesioned animals that still got fat, were his examples of how a missing control can produce a misleading conclusion.
Epidemiology can't easily establish causation
Taubes argues that observational associations, such as power lines and cancer or nutrition cohort data, have endless possible confounders yet are accepted at a 95% confidence threshold, far weaker than physics' five sigma.
Public health faces a conflict between honesty and action
If scientists list every uncertainty, as Feynman's ideal would require, nobody changes their behavior, which Taubes describes as the central tension of public health science.
WORDS WORTH SAVING
5 quotesThe message from Nobel Dreams and Bad Science was that bad scientists never get the right answer. It's just too hard to get the right answer for you to go in being sloppy and slipshod and lazy and ambitious and get it right. Nature isn't that kind.
— Gary Taubes
You've got a field where you can't do the experiments, so you develop a field of science around basically a hypothesis, and as more and more people see the same phenomena without using rigorous experimental techniques, they start to believe that this hypothesis is true.
— Gary Taubes
Once you decide you know what the truth is, you tend to stop listening to the people who disagree with you and then you surround yourself with the people who agree.
— Gary Taubes
Doing a background analysis is the hard, relentless, rigorous grunt work of science. It's endless and thankless, because if you do it right, all you'll do is prove that you were wrong all along.
— Gary Taubes
You decide you're gonna tilt at windmills for life. You gotta get used to the fact that the windmills are gonna kick you in the ass when you ride by.
— Gary Taubes
High quality AI-generated summary created from speaker-labeled transcript.
