The primary limitation to ground-based optical/IR interferometry is the turbulent atmosphere, which limits sensitivity by restricting the coherence volume, limits imaging accuracy by corrupting the fringe phase, and limits astrometric accuracy by corrupting the angle of arrival. Various advanced techniques can be used to circumvent these limits to some extent. Sensitivity can be increased with adaptive optics and laser guide stars, which should eventually be able to phase the individual apertures of an interferometer down to some cutoff wavelength, limited by tilt sensing. However, the sky coverage for cophasing the interferometer on an arbitrary object will remain limited at short wavelengths. For imaging, closure-phase techniques, well established in radio interferometry, will be used in next-generation instruments. However, for maximum sensitivity on extended objects, redundant arrays will be needed to cophase the interferometer. For astrometry, the limits to wide-field astrometry set by the atmosphere can be reduced somewhat with two-color techniques, but otherwise do not seem reducible by the techniques now being discussed. However, over narrow fields, the astrometric performance of an interferometer can be quite good. In space, without the corruptions of the atmosphere, the fundamental limitation is photon noise. However, technical issues such as metrology accuracy and practical issues such as maximum affordable baseline length will also limit performance.