How does "observation" affect physics? I watched this video which very very comprehensively demonstrates concept that sometimes particles behave differently based on whether or not they're being observed.
The idea that observing something (this is not related to the absorption of light involved in physical observation) can change the way that it behaves in any way is hard for me to grasp. It almost sounds like such a concept would have to come from one of those theories where reality just exists within our minds, or something wild like that. 
How does "observation" affect or interact with physics? - 
 A: You speak of the as yet unresolved quantum measurement problem. 
Although it is unresolved, most physicists think that it has nothing to do with solipsism, which is a sensational explanation but there are many reasons to think that this is not the right one:


*

*Conscious observers themselves are quantum systems. Observation always comprises interaction between physical systems and the observer, so it is impossible that observation cannot affect physics, whether quantum or classical. This is the observer effect, not to be confused with the uncertainty principle;

*Since observation is only a special kind of interaction, there is no need for them to be conscious: an observation could be a molecular change induced by a scattering or absorption. In QM we simply postulate nondeterministic mathematical "machines" (called observables) as observers that take a quantum state in and force it into one of its so called "eigenstates" of the particular machine we are using. These "machines" are very simple: they are mathematical models that could equally well model a mind or a molecule.
In connexion with the second paragraph, QM interpretations of the measurement problem break roughly up into two camps: "objective collapse" theories, which postulate the action is a true change in the observed system (our mathematical machines have an objective action) and that the quantum state is a real, objective thing changed by the obseveration, or there are people who think that change in the quantum state is simply an updating of the observer's knowledge akin to the change in a statistical distribution of an event that is wrought by new knowledge: if you learn that a flipped die landed on an even number, you would change the assumed distibution of the outcome to the conditional probability distribution conditioned on the outcome.
