> This idea that we are so sure of everything is pervasive
I think this is a common misconception made mostly by people who distrust science as a whole, but not really broadly held. I don't know anyone who understands an area of science at any depth who holds such a view.
The really broad problem is that in science there are things we have very high confidence in, and there are things we have very little confidence in but are the best working models currently - and everything in between. However, communication about where the confidence lies is often poorly done. Many observers who mostly hear about things through the lay press have no good tools to differentiate between "some person with a Ph.D said Y once" and "the vast majority of experts in this field are confident that X".
You are quite right that real experts have a healthy feel for what they don't know - and often humility - but those are often very different things from those that worry people outside the field. All of those same experts could happily give you a list of things they and their colleagues have very good confidence in. It's almost always the right thing to do to accept these and move on.
On the flip side any scientist will tell you there is small but real value at the margins for someone plugging away at a "crazy" theory. They are almost always wrong, but occasionally it turns up something really interesting. The trick is not to involve more people or too much resources, as these are long bets that will mostly fail.