Galaxies undergoing formation and evolution can now be directly observed over a time baseline of some 12 Gyr. An inherent difficulty with high-redshift observations is that the objects are very faint and the best resolution (HST) is only similar to 0.5 kpc. Such studies thereby combine in a highly synergistic way with the great detail that can be obtained for nearby galaxies through "archaeological" studies. Remark-able advances are being made in many areas, due to the power of our observatories on the ground and in space, particularly the unique capabilities of the HST ACS. Three new developments are highlighted. First, is the derivation of stellar masses for galaxies from spectral energy distributions (SEDs) using HST and now Spitzer data, and dynamical masses from both sub-mm observations of CO lines and near-IR observations of optical nebular lines like Ha. A major step has been taken with evidence that points to the z similar to 2 - 3 LBGs having masses that are a few x10(10) M-circle dot. Second is the discovery of a new population of red, evolved galaxies, again at redshifts z similar to 2 - 3 which appear to be the progenitors of the more massive early-type galaxies of today, with dynamical masses around a few x10(11) M-circle dot. Third are the remarkable advances that have occurred in characterizing drop-out galaxies (LBGs) to z similar to 6 and beyond, less than 1 Gyr from recombination. The HST ACS has played a key role here, with the dropout technique being applied to i and z images in several deep ACS fields, yielding large samples of these objects. This has allowed a detailed determination of their properties (e.g., size, color), and meaningful comparisons against lower-redshift dropout samples. The use of cloning techniques has overcome many of the strong selection biases that affect the study of high redshift populations. A clear trend of size with redshift has been identified, and its impact on the luminosity density and star formation rate estimated. There is a significant, though modest, decrease in the star formation rate from redshifts z similar to 2.5 out through z similar to 6. The latest data also allow for the first robust determination of the luminosity function at z similar to 6. Last, but not least, the latest UDF ACS (optical) and NICMOS (near-IR) data has resulted in the detection of some galaxies at z similar to 7 - 8.