Bayesian Priors
Prior beliefs do more than compete with new evidence: they shape how that evidence is interpreted. Updating can fail when the prior distorts the very observation meant to correct it.
The unsettling claim is not merely that you see evidence and then force it to fit your beliefs. What you already believe can change what the evidence appears to be in the first place.
E1The prior sits upstream of judgment
A prior is the expectation already in place when evidence arrives. It influences interpretation before you reach a conclusion, so two people can encounter the same signal yet experience it as supporting different stories. This creates a feedback loop: the prior shapes the perceived evidence, and that perception then seems to confirm the prior. Bayesian updating works only when the observation retains enough independence to push back.
E1Influence is not inevitability
The claim is that priors change how evidence is interpreted, not that every observation is wholly manufactured by belief. This material does not establish when a prior will dominate, how strong the distortion will be, or which kinds of evidence can reliably overcome it.
E1Separate the signal from its interpretation
Before evaluating consequential evidence, write down two things separately: what you directly observed and what you think it means. Then ask what someone holding the opposite prior could honestly infer from the same observation. The gap exposes where your starting belief may be doing the interpreting.
Episodes that teach this
-
Believing is Seeing - How Bayesian Priors Trick Your Senses - Future IQ
· explained at 3:45
6,488 views
"our priors what we believe already changes how we interpret evidence"