Comment by bobbylarrybobby
23 days ago
You are basically hitting on what has been referred to as high modernism, which promotes a level of confidence in science and technology that can only be maintained by eschewing all the inherent complexity of the world. The scientific method can really only study systems by modifying a handful of variables at a time and keeping the rest fixed, and isn't really capable of handling hundreds of interacting variables. Rather than acknowledge this limitation, high modernism embraces simplification even to the detriment of its products.
Further reading: https://en.wikipedia.org/wiki/Seeing_Like_a_State
> The scientific method can really only study systems by modifying a handful of variables at a time and keeping the rest fixed
Not true. Statistical measures of large systems are a routine thing in the natural sciences. However that's higher effort and tends to make it more difficult to communicate the results to others so it's avoided whenever possible.
Also high dimensional models carry a distinct risk of overfitting.