Update Your Priors
The smartest people I know never say “this is how it is.”
I’ve been fortunate to learn from my mentors - people I deeply admire, who aren’t famous but they’re the most well read people you’ll ever meet. And after years of observing how they operate, I’ve noticed something fundamental about how they think.
They think in probabilities.
There’s a concept in statistics called conditional probability - the likelihood of something happening given that something else has already happened. It comes from Bayes’ theorem, the idea that our beliefs should mathematically update when new evidence appears.
In Bayesian thinking, a “prior” is your existing belief about something before you see new evidence. It’s the lens you carry into every decision. And the quality of your thinking depends entirely on whether you’re willing to update that lens.
Most people aren’t.
Here’s how it plays out:
In 2007, Steve Ballmer laughed at the iPhone. “There’s no chance that the iPhone is going to get any significant market share,” he said. His prior was simple - phones need physical keyboards, business users won’t switch, Nokia and BlackBerry own this market. It was a reasonable belief at the time.
But then the evidence started pouring in. Consumers loved the touchscreen. App developers flooded in. The entire mobile ecosystem shifted. And yet, Microsoft didn’t update its prior. They kept building Windows Phone as if the old rules still applied. By the time they acknowledged the shift, the game was already over.
Contrast that with Satya Nadella, who inherited the same company years later and did the opposite. He looked at the evidence - cloud was eating software, open source was winning, mobile-first was real - and completely rewrote Microsoft’s priors. He bet everything on Azure. The stock went from $40 to $400.
Same company. Same resources. The difference was one leader who refused to update his priors, and another who made it his entire operating system.
This is how most people operate. A fixed worldview, calcified over time, resistant to new evidence. They form a view once and hold onto it like a religious belief.
The smartest people do something different.
When they see a new signal - someone succeeding where they once failed - they don’t dismiss it. They get curious. They ask: What changed? What do they know that I didn’t? What’s the probability of success given this new information?
They update their priors.
Think of Bezos, who famously changes his mind constantly and considers it a sign of intelligence. “People who are right a lot of the time,” he says, “are people who often change their minds.” Or Munger, who built his entire investment philosophy around identifying when he was wrong. Or Darwin, who kept a special notebook for evidence that contradicted his theories - because he knew his mind would conveniently forget it otherwise.
The pattern is consistent: rigid thinking is the enemy of insight.
And this has never mattered more than right now.
We are living through the fastest technological shift in human history. AI isn’t just another tool - it’s rewriting the fundamental rules of how work gets done. What a team of ten engineers built in six months last year, a single person with the right AI workflow can ship in a weekend today. The entire playbook for how we hire, build, sell, create - all of it is being rewritten in real time.
This means every manager, every leader, every founder needs to be updating their priors not once a year, not once a quarter - but constantly. The half-life of a valid assumption has collapsed. The leader who still plans based on how things worked even twelve months ago is already operating on expired beliefs.
The cost of a fixed worldview has never been higher. Your beliefs should have an expiration date. When credible new evidence appears, you should feel compelled to update, not defensive.
The question isn’t whether your current view is right. The question is: what would it take to change your mind?
If the answer is “nothing” - that’s not conviction.
That’s religion.