On large scales, the agreement between CDM and observations is spectacular. CDM simulations also account remarkably well for smaller scale observations such as galaxy clustering , and even for the radial distribution  and total mass associated with galaxies .
The problems of big bang nucleosynthesis may be a clue to new particle physics or astrophysics. On cusps there has been tremendous progress on observing velocity fields in nearby galaxies, and also real progress in improving simulations. Observed simulations may agree with observed velocities in galaxy centers better than seemed likely a few years ago. But it is something of a scandal that there is still so little theoretical understanding of dark matter halo central behavior, although people are making progress on this problem . It is likely that triaxial halo structure and poorly understood gastrophysics will turn out to be relevant. On dark matter halo substructure, it looked last year as if a challenge might be turning into a success for CDM, if the amount of substructure predicted by CDM is indeed what is required to account for the number of satellites seen and for the flux anomalies observed in radio lensing. The main question is whether the amount of dark matter in subhalos and the predicted radial distribution of such substructure agrees with lensing and observed satellites. Much work remains to be done to test the theory quantitatively. Regarding angular momentum problems, resolving the crucial issues again involves developing better understanding of messy astrophysics. Fortunately, in all these areas wonderful new telescopes are providing crucial data that will help develop and test theory.
At this conference, the updates on direct dark matter detection experiments showed impressive progress, with greatly expanded parameter space now being probed experimentally. Unfortunately, the relevant weakly interacting massive particle (WIMP) elastic scattering cross section is quite uncertain. The annihilation cross section is much better constrained, since this is what determines the WIMP abundance today in most models, but the dark matter density at the center of the Milky Way is uncertain because there are competing processes that can enhance it or diminish it. It is possible that the new H.E.S.S. array of atmospheric Cherenkov telescopes (ACTs) has discovered dark matter annihilation at the galactic center, with a dark matter particle mass of approximately 18 TeV , which is unexpectedly high but not impossible for the lightest supersymmetric partner particle. The high energy gamma rays appear to come from the sort of centrally peaked density profiles predicted as a consequence of scattering of WIMPs by the star cluster surrounding the central supermassive black hole Sag A* , and the necessary dark matter density is consistent with theoretical expectations  if there is baryonic contraction [47, 81]. It is unlikely that the gamma rays come from near the black hole event horizon, since there does not appear to be any time variability in the gamma rays although the X-ray and optical radiation from the black hole is quite variable. The main alternative explanation for the high energy gamma rays is that they come from the supernova remnant known as Sag A* East, which covers a region several pc across surrounding the galactic center. These alternatives can perhaps be distinguished via the different angular distributions expected, when data taken with all four H.E.S.S. ACTs are analyzed. Another important discriminant is the energy spectrum of the gamma rays, which must sharply cut off if the gamma rays come from annihilation.
I thank David Cline for inviting me to present this opening talk at Dark Matter 2004, and I thank my colleagues and students - especially those who participated in the UCSC Workshop on Galaxy Formation in August 2004 - for enlightening me about recent work in cosmology. I am grateful to Ben Metcalf for a careful reading of this manuscript. I acknowledge support from grants NASA NAG5-12326 and NSF AST-0205944.