Sunday, July 17, 2022

"Model" = Facts

 Apparently the Dutch government mandates that will put farmers out of business are based on computer models, so were:

  • Climate change policies;
  • Covaids policies.

Outside experts who examined the covid disease model found that the code was trash and amateurish. They could not reproduce the results the governments were promoting.

I've worked on a handful of computer models of physical systems over the years. In some fields, like say nuclear science, you can get a model that's as accurate as your inputs. If you perfectly describe some system, the model and experimental data will align quite well. In other fields, the model is a vague qualitative approximation because simplifications and assumptions are necessary and some characteristics of a system are difficult to precisely quantify. For example, in an optical system, it's time consuming to really measure the precise characteristics of a reflective surface at some specific wavelength of light, and if that much experimental characterization of something is required, building the model is a waste of time, anyway. Some things are impossible to measure, for example, what's the reflectivity of a surface when viewed from the inside of a crystal?

Climate models include simplifications of atmospheric behavior, because there's insufficient computer power to include them. Some things that massively affect climate can't even be included, like volcanic eruptions, because they're random events. There are most likely aspects of climate that are entirely missing from the models because they're not known.

Anyway, it's pretty interesting that model outputs are now regarded as gospel truth.

No comments:

Post a Comment