The papers reviewed here have shown strong evidence that almost all types of high-z sources contain dust. The only case where the evidence is not yet convincing is the quasar absorption line systems, particularly at the column densities of the Lyman forest. In that case, it is not even clear that we are dealing with galaxies. All other cases involve galaxies with at least some inferred star formation. Hence at high redshift, as in the local universe, star formation and dust are correlated. The dominant location of high redshift star formation remains debatable. While it remains plausible that the majority of star formation at z > 2 remains completely invisible shortwards of the rest-frame FIR, there is a strong body of evidence that moderately dust obscured but rest-frame UV-bright galaxies should dominate the star formation in the early universe. It is likely that this issue will not be definitively settled until the sub-mm background is fully resolved.
Acknowledgments. I thank Adolf Witt for cajoling me to become Daniela Calzetti's replacement at the conference. My talk and this review benefited from discussions and correspondence with John Blakeslee, Daniela Calzetti, Mark Dickinson, Tim Heckman, Jason Prochaska, and Chuck Steidel. I thank the anonymous referee and Andrew Blain for suggestions that have improved the paper, allowing me to go deeper into the issues than I did at Estes Park.