I think science is primarily attempting to externalize the expertise (the way the experts do inference), to formalize it, so it can be independently scrutinize

I think science is primarily attempting to externalize the expertise (the way the experts do inference), to formalize it, so it can be independently scrutinized, and reproduced. Of course, this is based on belief that experts can be fallible (as stated in the article), but also on belief that opaque expert's intuition is not as valuable as clear understanding of the inference process. Of course, this doesn't make the experts useless. The expert intuition is still mostly right, even if it lacks explainability. It is also a useful shortcut.