The market pays for certainty, not for being right

Attention rewards a clean answer, and being wrong later carries almost no cost.

Consider the incentives facing anyone whose living depends on being listened to. A confident, memorable, unqualified claim travels. A carefully bounded one does not. If the confident claim turns out to be wrong, the cost is usually small: attention has moved on, few people are keeping score, and the failure can be reframed as early or as right in spirit. Meanwhile the person who was correctly uncertain gets no credit at all, because nobody remembers the accurate forecast that said it could go either way.

This is a selection effect operating on the whole supply of public expertise, and it does not require anyone to behave badly. Confident voices are amplified, hedged voices are not, and the population of people you can easily hear from is therefore skewed towards certainty regardless of accuracy. The individuals may be sincere. The system still produces a distorted sample, and you are standing at the end of it.

Two habits help. First, discount confidence in any setting where the speaker is competing for attention, and weight it more where they are not, such as when someone is talking to peers who would notice an overreach. Second, look for scorekeeping. Anyone who publishes their record, revisits old claims, or maintains a public account of what they got wrong has voluntarily accepted a cost the market does not impose on them. That willingness is rare, it is cheap for you to check, and it is one of the better signals available.