Recent advances in diffraction-limited techniques on 8-10m telescopes using adaptive optics (AO) and integral field spectrographs (IFS) have led to significant scientific achievements and are stimulating the design of future instrumentation. My talk will focus on development and use of current near-infrared AO instruments to study galaxies in the early universe, as well as the design and capabilities of AO instrumentation for the Thirty Meter Telescope (TMT). I will present results of a spatially resolved survey of intermediate redshift (z~1) star forming galaxies that we use to explore the scaling relationships of star forming regions locally and at high-redshift. I will also present a powerful new survey that utilizes IFS and AO observations and multi-wavelength data sets to reveal high-redshift (z~2) radio-loud quasar host galaxies. There are numerous instrument design and observational challenges that need to be overcome in order to exploit the diffraction-limit of an extremely large telescope. I will discuss instrument design, diverse science cases, and our current efforts in the laboratory to maximize near-infrared integral field spectrograph and imager sensitivities for the first light TMT instrument IRIS.