How the non-validity of science at different scales can be quantified
I want to add to the quite well-made point above that there even is a quantitative method describing the “approximatedness” of science, known in physics as the renormalization group. It shows how the relevant laws of physics change as you change the (length-) scales you are looking at. This means that at very large distances (our every day world) very different laws of physics dominate than at very tiny scales such as in particle accelerators. One example is the strength of the fundamental forces we know (weak & strong nuclear force, electromagnetic, gravity). When treated with the renormalization group one can show that the weak and strong nuclear forces, which are dominating the physics in nuclei, is completely irrelevant to length scales that we experience in our every day lives. For us, only the electromagnetic force and gravity dominate.
More is different
This also shows that we can’t know everything yet — at the tiniest scales there could always be new relevant forces appearing of which we don’t know anything about yet.
Theories are always taylor-made to a given scale.
Particle physics describes the smallest scales accessible to us, around the size of the constituents of nuclei, organic chemistry reactions happens at the size of carbon molecules, biology explains what happens on the scale of a cell. And it is always immediately clear that as we go to a different scale, very different laws of nature appear. This fact was beautifully put into words by Nobel prize winner P. W. Anderson in his famous article “More is different”  which I recommend everyone interested in the fundamentals of science to read.