If we have seen further, it’s largely by standing on layers of previous errors.
I had a lecture last week about cognitive biases, their possible adaptiveness and also impacts on science. It also led me to think about the old “hyper-competent scientist” trope so typical for SF. Science-fictional scientists can often recite complex information verbatim and know the answer to every question, even if it’s unrelated to their subfield of research – but then again, there are few molecular biologists focused on studying only one class of receptors in SF. Science-fictional scientists are usually either “generalists”, or very well-informed about basically every single subject of their field. A biologist can easily identify any plant or animal, run various analyses, create a model of a protein’s active site as well as an ecosystem simulation. And if they by chance don’t know something, they are able to quickly look the relevant information up or find out.