‒ AMA : A guide to medications and supplements: determining what to take, what to skip, and how to know if they're working for you

In one sentence

Stop "impulse shopping" for molecules by first defining a tightly constrained problem with a specific metric, threshold, and timeline, then categorizing the intervention’s job (Disease Treatment, Symptom Relief, Risk Reduction, or Optimization) to dictate how much evidence you should demand.

Why read this (and who it's for)

This guide is for anyone currently taking—or considering taking—supplements or medications who feels overwhelmed by conflicting advice or unsure if their regimen is actually working. It solves the "paradox of choice" in the supplement space by providing a rigorous mental model to filter out noise, identify true risks (including opportunity costs), and ensure that any intervention you choose is actually addressing a defined problem rather than a vague aspiration. After reading this, you will be able to audit your current supplement stack using a formal "Problem Definition" framework and decide which items to keep, skip, or research further based on their specific "job."

The big idea

The central thesis of this episode is that most people fail at health optimization because they define their problems at the wrong level of abstraction. When you say "I want more energy" or "I want to live longer," you are creating a vague, unmeasurable goal that allows for "storytelling." Because the goal is fuzzy, your brain will find any minor change—a slightly better night's sleep or a tiny shift in a lab value—and attribute it to the supplement you just started taking. This leads to false positives where people continue taking ineffective (or potentially harmful) substances for years because they can't prove they aren't working.

To fix this, you must shift your focus from the molecule to the problem. You need to define a problem so tightly that you could actually be proven wrong. Once a problem is defined, you must then categorize what "job" the intervention is supposed to do. This classification dictates your evidence threshold: if you are treating a life-threatening disease, you might accept higher risk for stronger evidence; if you are merely trying to "optimize" a healthy baseline, you should maintain high skepticism and low risk tolerance.

Key ideas

Actionable Problem Definition

What it is — A problem definition consists of three specific components: an actionable metric (what you are measuring), a threshold (the specific number that indicates success/failure), and a time horizon (how long you will give the intervention to work). Why it matters — Without these three pillars, any intervention you take becomes an "impulse shop." A vague goal allows your brain to manufacture stories of success, leading you to waste money or ignore side effects because you have no objective way to measure progress. In the book — Instead of saying "my cholesterol is bad," define it as: "My ApoB is 130 mg/dL; I want it below 60 mg/dL in the next six months." Instead of "my sleep is bad," say: "It takes me 60 minutes to fall asleep four nights a week; I want that under 10 minutes within two months."

The Counterfactual

What it is — This is the question: "What happens if I do nothing?" You must evaluate whether the current state of the problem meaningfully increases risk, reduces quality of life, or creates downstream consequences. Why it matters — This separates real problems from things that merely feel actionable. If doing nothing carries low risk, your evidence bar for an intervention should be significantly higher. In the book — The author argues that poor problem definition almost guarantees false positives because you are reacting to a feeling rather than a measurable risk of inaction.

Categorizing the "Job"

What it is — Every medication or supplement falls into one of four buckets: Disease Treatment, Symptom Relief, Risk Reduction, or Optimization. Why it matters — You cannot apply the same evidence threshold to all four. If you misclassify a job (e.g., treating an "optimization" goal as "risk reduction"), you will likely accept a level of risk that is unjustifiable for the expected benefit. In the book — The author notes that many longevity interventions are actually "optimizations masquerading as risk reductions." They use the language of prevention and healthspan, but the actual evidence often looks like speculative optimization.

Evidence Thresholds & Risk Tolerance

What it is — Your willingness to accept downside (side effects, cost, hassle) should be inversely related to the speculativeness of your goal. Why it matters — For serious problems (Disease Treatment), you want strong evidence but can tolerate more risk because the counterfactual (doing nothing) is severe. For speculative goals (Optimization), you must default into a skeptical state and minimize any downside. In the book — The author explicitly states: "The more serious and concrete the problem, the more downside you may be willing to keep... and the more speculative the goal, the less downside you should be willing to tolerate."

Mechanism vs. Outcome

What it is — Just because a supplement has a plausible biological mechanism (how it works in a cell) does not mean it produces a meaningful clinical outcome (how it changes your life or health). Why it matters — People often get seduced by "compelling stories" about how a molecule works, but if there is no validated surrogate marker or hard outcome trial, the intervention may be useless. In the book — The author distinguishes between well-validated surrogates (like ApoB) and vague markers (like "inflammatory markers" or "heavy metals") which are often used to pontificate about detoxifications without real evidence.

Frameworks & tools

The Actionable Problem Definition Framework

Use this before purchasing or starting any new supplement/medication:

1. Identify the Metric: What specific biological marker or lived experience am I measuring? (e.g., Blood pressure, sleep latency, ApoB levels).
2. Set the Threshold: What is the specific number that constitutes a "success"? (e.g., < 120/80 mmHg).
3. Define the Time Horizon: How many weeks/months will I give this intervention to reach that threshold?
4. Analyze the Counterfactual: If I do nothing for X months, what is the actual risk to my health or quality of life?
5. Categorize the Job: Is this Disease Treatment, Symptom Relief, Risk Reduction, or Optimization?

The Intervention Decision Matrix

Determine your evidence requirements based on the "Job" classification:

• Disease Treatment: Demand hard outcome trials or well-validated surrogates; Accept higher risk/side effects.
• Symptom Relief: Prioritize actual lived benefit (how you feel); Watch for placebo and noise.
• Risk Reduction: Demand high evidence bar (hard outcomes or validated surrogates); Minimize unknown safety risks.
• Optimization: Default to extreme skepticism; Minimal risk tolerance; Require highest evidence bar.

Memorable examples & stories

The Storytelling Machine — The author describes how humans are "storytelling machines." If you have a vague goal like "longevity," your brain will find any tiny change—a lab value improving slightly or feeling better one Tuesday—and weave a narrative that the supplement is working. Lesson: Vague goals guarantee false positives.

ApoB vs. Detoxification — The author contrasts ApoB (a well-validated surrogate marker) with vague "inflammatory markers" or "heavy metal" detoxifications. Lesson: Not all biomarkers are equal; some are validated tools for risk reduction, while others are often used to pontificate without evidence.

The Optimization Masquerade — Many longevity interventions use the language of "prevention," "aging," and "resilience" to sell products that are actually speculative optimizations. Lesson: Be wary of any intervention that uses high-stakes language but offers only low-certainty, mechanistic evidence.

Key quotes

"Do not start with the molecule, start with the problem." Why it matters: Shifts your focus from "buying stuff" to solving a specific biological issue.

"Most of the longevity interventions are really optimizations masquerading as risk reductions." Why it matters: Warns you to apply a higher skepticism threshold to most "longevity" products.

"If you can't state the metric, the threshold and the timeline... you're probably not in a good spot to start anything." Why it matters: Provides a litmus test for whether you are making an informed decision or just impulse shopping.

How to apply it

This week

• Audit your "Why": Pick one supplement you currently take. Write down the specific metric, threshold, and timeline you expected it to achieve. If you can't find a specific number/time, acknowledge that you are likely "storytelling."
• Identify the Counterfactual: For that same supplement, ask: "What happens if I stop taking this today?" If the risk of stopping is zero, evaluate if the evidence for continuing is actually high enough to justify the cost and hassle.

Ongoing

• The "Job" Filter: Before buying any new product, categorize it into one of the four buckets (Disease, Symptom, Risk, Optimization).
• Evidence Check: If you categorize a purchase as "Optimization," immediately move your skepticism dial to maximum. Do not accept "mechanism" as proof of efficacy.

Common pitfalls

• Vague Problem Definition: Starting with "I want to be healthier" ensures you will never know if an intervention actually worked.
• Mechanism Bias: Falling for a "compelling story" about how a molecule works in a lab without seeing any data on human outcomes.
• Ignoring Opportunity Costs: Failing to factor in the cost, hassle, and the potential downsides of taking something when the baseline risk is low.
• Relative Risk Confusion: Being misled by relative risk figures rather than absolute risks (though the text notes this as a point covered, it doesn't provide the specific math).

Connects to

• The concept of "Storytelling Machines": Relates to behavioral economics and cognitive biases where we find patterns in noise.
• Longevity/Healthspan: Extends the discussion on how to actually measure these abstract concepts using hard data.

If you remember only three things

• Start with a specific problem (Metric + Threshold + Timeline), not a molecule.
• Categorize every intervention into one of four "jobs" to determine your risk tolerance.
• Mechanism is not outcome; don't let a "compelling story" substitute for validated evidence.

Internalize it

• Reflection questions

  • What is the specific metric I am trying to improve right now?
  • If I do nothing about this problem for six months, what is the actual risk?
  • Is the supplement I'm taking an "optimization" or a "risk reduction"? (If it's optimization, why am I tolerating any risk?)
  • Can I prove that my current supplement is working, or am I just telling myself a story because I feel slightly better today?

• Flashcards Q: What are the three components of an actionable problem definition? A: Metric, Threshold, and Time Horizon. Q: What are the four "jobs" of a medication/supplement? A: Disease Treatment, Symptom Relief, Risk Reduction, and Optimization. Q: How does your evidence threshold change based on the "job"? A: Higher risk tolerance for serious diseases; higher skepticism for optimizations. Q: Why is "mechanism" alone insufficient? A: Because a plausible biological mechanism doesn't guarantee a meaningful clinical outcome. Q: What is the danger of vague problem definitions? A: They allow you to become a "storytelling machine," creating false positives where you think an intervention is working when it isn't.

Fidelity note

I am highly confident in the summary of the mental models provided (Problem Definition, The 4 Buckets, and Mechanism vs. Outcome). However, the provided material mentions that the author provides a "short list of over-the-counter supplements" at the end of the episode, but this specific list was not included in the transcript provided. Readers should verify the specific supplement recommendations against the full AMA episode.