Frankly there is nothing woo about it.

Even without venturing into the realms of Quantam Mechanics or Chaotic Dynamics, all measurements are, by physical definition, upto the resolution of the measuring instrument. So it always come with inherent +/- error bars. Sometimes these error bars are made explicit, at other times they are elided, often with the assumption that it's obvious.

Models are always simplifications. That is precisely why they are useful. ( A 1:1 scale map is not very useful, especially when we already have one). Models are obtained by ignoring the effect of many known disturbances, but whose effect one deems not to exceed a tolerance bound.