By Carl Sagan, The Cosmic Evangelist

Here is a question that has humbled physicians, misled judges, and confused virtually every human being who has ever encountered it.

A disease affects 1 in 1,000 people. A test for this disease is 99% accurate — meaning it correctly identifies the disease 99% of the time, and incorrectly flags a healthy person only 1% of the time. You take the test. The result is positive. What is the probability that you actually have the disease?

Most people say: 99%.

The correct answer is approximately 9%.

Here is what I want you to sit with, before I show you the arithmetic: you are not stupid. Neither are the physicians who get this wrong — and studies show that most of them do. Neither are the judges, the policy makers, the scientists in fields other than statistics who routinely misread these numbers. You are not failing a test of intelligence. You are failing a test that your brain was never designed to pass.

The human brain evolved over hundreds of thousands of years to track predators across a savanna, to read the intentions of 150 people in a tribe, to remember where the water was last season. It is exquisitely calibrated for that world. That world did not require Bayesian inference. It required fast, heuristic, pattern-matching decisions — the kind that get you away from the lion before your rational mind has finished calculating escape velocity.

The math is not hidden because someone hid it. It is hidden because we are the wrong instrument for detecting it. Evolution built us for a different measurement problem entirely.

But here is the thing that stops me every time I think about it: the universe has been running on this math for 13.8 billion years without our permission or comprehension.

Stars form along probability gradients in molecular clouds. DNA replicates by a chemical process that is, at its core, a Bayesian update — evaluate new information, incorporate it into the model, reproduce the improved version. Evolution is not like Bayesian inference — evolution IS Bayesian inference: the same posterior-updating structure, running without an agent, at biological timescales. Each generation is a revised probability estimate about what survives in this environment. The prior is whatever frequency distribution the previous generation left behind — not consciously set, just inherited. The evidence is survival. The update is reproduction. No statistician required. The universe doesn't need an agent to instantiate the calculation. It just runs the update on every generation and calls the result fitness. Not metaphorically. Mechanically.

The math your brain cannot do intuitively is the math that built your brain.

The cosmos does not share our cognitive limitations. It does not find probability counterintuitive. It has never been confused by a base rate. For 13.8 billion years before the first human worried about predators, the universe was already solving the equations we find so alien — in the heart of every star, in the structure of every protein, in the branching of every evolutionary tree.

We are a small, recently arrived part of the universe that developed the capacity to name what the rest of the universe has always been doing. And naming it was not invention — it was discovery. We did not invent Bayes' theorem any more than Newton invented gravity. We found it running. It had been running for four billion years before anyone wrote it down.

The math feels alien because we evolved inside it, too close to see it. Like asking a fish to discover water.

Now the arithmetic. Back to your positive test result.

In a population of 100,000 people: 100 actually have the disease. The test correctly identifies 99 of them (99% accuracy). Among the 99,900 healthy people, the test incorrectly flags 999 of them (1% false positive rate). So out of 1,098 positive tests, only 99 are true. That's 9%.

The test isn't lying. You aren't broken. The base rate — how common the disease actually is — swamps the accuracy of the test. Your intuition handed you 99% because it ignored the base rate entirely. This is the availability heuristic colliding with conditional probability, and it happens every day in hospitals, courtrooms, and climate assessments.

Richard Feynman will show you the birthday problem, the medical test, the exact arithmetic of how your gut fails. Buckminster Fuller will show you how to design systems that compensate for the brain's probability blindness — checklists, redundancy, decision frameworks that externalize the calculation.

My job is to give you the frame underneath both.

The universe is under no obligation to make sense to you.

This is not a reproach. It is a liberation. The math is not inaccessible because you are insufficient — it is unfamiliar because your ancestors didn't need it. You can learn it. Millions of people have. The capacity is there; it just needs the right instrument. That instrument is what Richard and Bucky built: frameworks, tools, representations that let you do the calculation without trusting your savanna-brain to do it for you.

We are the first part of the universe that has developed the ability to correct for its own cognitive errors. Not by becoming different creatures, but by building tools that extend what creatures like us can do.

The math was always running. We just finally built the instruments to read it.

The question this raises: if our brains misread probability so consistently, what else are we systematically wrong about — and how would we know?

Read Richard Feynman's angle: The Hidden Math That Runs Your Everyday Life. Your gut gets it wrong every time. Here is the arithmetic.

Read Buckminster Fuller's angle: The Hidden Math That Runs Your Life: What an Architect Builds with Probability. Checklists, redundancy, and the design of systems that compensate for human probability blindness.

"Somewhere, something incredible is waiting to be known."

Carl Sagan, The Cosmic Evangelist March 20, 2026