There is now a substantial body of evidence that a deconfined quark-gluon plasma is created in ultrarelativistic collisions of heavy nuclei. Some key observables which are used to gauge the production of the quark-gluon plasma are the hadronic spectra and multiplicities across many species, azimuthal anisotropies in the production of various hadrons, jet quenching, photon/dilepton production, and heavy quarkonium suppression. A key question in the study of the quarkgluon plasma is what happens to these observables as one changes the system size. In particular, it is very interesting to study collisions of "small systems" such as pp, dA, pA, and e + e − since naively one does not expect to produce a QGP in these cases. One of the surprises from such studies was the existence of sizable azimuthal anisotropies in the hadron spectra in the highest multiplicity classes, which led to speculations that one can generate a QGP in such relatively rare events. This interpretation, however, is complicated by the fact that there can be multiple sources of azimuthal anisotropy. I will discuss our current understanding of the different sources of azimuthal anisotropies and the multiplicity windows in which each mechanism is expected to dominate. I will also provide a critical discussion of the reliability of hydrodynamic models applied to small systems. Finally, I will briefly discuss jet quenching and heavy quarkonium suppression in small systems and discuss whether or not these observables show any indication of the production of a small QGP droplet.