This AMA episode focuses on developing a structured framework for evaluating medications and supplements. The core insight is that effective decision-making begins with defining a specific, measurable problem rather than starting with a molecule. The episode categorizes interventions by their 'job'—disease treatment, symptom relief, risk reduction, or optimization—and explains how the evidence bar and risk tolerance should shift accordingly.
Summarized by Podsumo
People often define problems at the wrong level of abstraction (e.g., 'I want more energy') instead of making them actionable with a metric, threshold, and timeline.
Interventions should be classified into four jobs—disease treatment, symptom relief, risk reduction, and optimization—each demanding a different evidence standard and risk tolerance.
A key warning is that many longevity supplements are 'optimizations masquerading as risk reductions,' borrowing prevention language but lacking hard outcome evidence.
The host emphasizes starting with the problem, not the molecule: 'If you can't state the metric, the threshold, and the timeline, you're probably impulse shopping on your favorite website.'
For symptom relief, placebo effects can be acceptable if the downside is low; for risk reduction, validated surrogate markers like ApoB are necessary.
"Do not start with the molecule, start with the problem. Define tightly enough that you could actually be proven wrong. If you can't state the metric, the threshold and the timeline and even the consequences of doing nothing, you're not really making an intervention decision, you're probably impulse shopping on your favorite website. — Peter Attia"
"The right question is whether a specific intervention makes sense for a specific person with a specific problem. — Peter Attia"
"I think the challenge... is in part because most of the longevity interventions are really optimizations, masquerading as risk reductions. They kind of borrow the language of prevention, aging, health span, resilience, longevity. But the actual evidence... looks much more like speculative optimization interventions rather than true risk reduction. — Peter Attia"