In these lectures I have been looking for discord, but have not found it. The classical tests return results for cosmological parameters that are consistent with but considerably less precise than those implied by the CMB anisotropies, given the usual assumptions. It is fair to say that the numbers characterizing the concordance model, m 0.3, 0.7 are robust in the context of the framework of FRW cosmology. It is, in fact, the peculiar composition of the Universe embodied by these numbers which calls that framework into question.
Rather small changes in the assumptions underlying pure FRW cosmology (with only an evolving vacuum energy density in addition to more familiar fluids) can make a difference. For example, allowing w = -0.6 brings the number counts and z-distribution of faint galaxies into agreement with a Universe strongly dominated by dark energy (Q = 0.9). The same also true of the high-z supernovae observations ). Allowing a small component of correlated iso-curvature initial perturbations, as expected in braneworld cosmologies, can affect the amplitudes and positions of the peaks in the angular power spectrum of the CMB anisotropies , and therefore the derived cosmological parameters.
But even more drastic changes have been suggested. Certain braneworld scenarios, for example, in which 4-D gravity is induced on the brane  imply that gravity is modified at large scale where gravitons begin to leak into the bulk . It is possible that the observed acceleration is due to such modifications and not to dark energy. More ad hoc modifications of General Relativity  have also been proposed because of a general unease with dark energy - proposals whereby gravity is modified in the limit of small curvature scalar. My own opinion is that we should also feel uneasy with the mysterious non-baryonic cold dark matter, because the only evidence for its existence, at present, is its gravitational influence; when the theory of gravity is modified to eliminate dark energy, it might also be found that the need for dark matter vanishes.
In general, more attention is being given to so-called infrared modifications of gravity (e.g. ), and this is a positive development. High energy modifications, that affect the evolution of the early Universe, are, as we have seen, strongly constrained by considerations of primordial nucleosynthesis (now, in combination with the CMB results). It is more likely that modifications play a role in the late, post-recombination evolution of the Universe, where the peculiarities of the concordance model suggest that they are needed. The fact that the same rather un-natural values for the comparable densities of dark energy and matter keep emerging in different observational contexts may be calling attention to erroneous underlying assumptions rather than to the actual existence of these "ethers".
Convergence toward a parameterized cosmology is not, without deeper understanding, sufficient reason for triumphalism. Rather, it should be a motivation to look more carefully at the possible systematic effects in the observations and to question more critically the underlying assumptions of the models.
I thank Rien van de Weygaert, Ole Möller, Moti Milgrom, Art Wolfe, Jacob Bekenstein, and Scott Trager for useful comments on the manuscript. I also thank Gary Steigman, Wendy Freedman, and Luis Ho for permission to use Figs. 1 and 2. I am very grateful to the organizers of the Second Aegean Summer School on the Early Universe, and especially, Lefteris Papantonopoulos, for all their work and for inviting me to the very pleasant island of Syros.