Belief lives on a spectrum, not a switch
Bayesian reasoning treats belief as a matter of degree, not a binary 'true or false.' Before you see any new evidence, you already have some level of confidence that a claim is true. That starting confidence is your prior. When new evidence arrives, your confidence should move — up, down, or sideways — in response. That movement is the update. The key insight is that rational belief is not about flipping from 'I don't know' to 'I'm certain.' It is about calibrated movement along a spectrum from near-zero confidence to near-certainty, with most beliefs living somewhere in the middle.
This framing is powerful because it matches how careful thinkers actually reason. A doctor does not ignore the patient's history when reading a test result. A detective does not treat every suspect as equally likely before looking at the evidence. An engineer does not discard all previous performance data when a single anomaly appears. Bayesian thinking makes this prior-informed updating explicit and disciplined, transforming intuitive adjustments into a repeatable, auditable process.
The spectrum metaphor has a practical consequence: it forces you to quantify, or at least rank, your uncertainty. Saying 'I think it's probably true' is less disciplined than saying 'I'd put my confidence around 70%.' The act of assigning a rough number — even a qualitative one like 'fairly confident' or 'slightly more likely than not' — commits you to a position that evidence can then move. Without a starting point, there is nothing for the evidence to update.