The study of a sample of river sediment enables the determination of spatially averaged denudation rates that provide an exceptional perspective on erosion and weathering processes that have taken place within a landscape. These measurements are done with in-situ-produced cosmogenic nuclides (e.g., 10Be, 26Al), mostly in quartz from alluvial sediment. Cosmogenic nuclides are produced when secondary cosmic rays interact with the very uppermost layer of the Earth's surface. They are produced within a characteristic depth scale of about 1 m, which means that the measured concentrations record an integrated denudation history while material passed through this depth interval. Depending on the denudation rate the resulting integration time scales are 103 to 105 years, and one obtains a robust long-term estimate of natural denudation that is relatively insensitive to short-term changes. The last 10 years have seen significant research activity using these methods, and an array of fascinating tectono-geomorphologic and geochemical insights are emerging. Amongst these is the ability to identify the physical and chemical processes with which a landscape responds to tectonic activity or climate change. A compilation of world-wide denudation rates in non-glaciated areas, that however, does not yet include some of the world's most active mountain belts, has resulted in the following findings, some of which have been unexpected: (1) No obvious relationship between precipitation or mean annual temperature and total denudation is apparent. (2) Topographic relief alone does not result in high rates of denudation. (3) Denudation rates are high in areas of landscape rejuvenation; that is triggered and controlled by tectonic activity (faulting, escarpment formation and retreat, rifting, surface uplift). (4) Rates of weathering (using a combination of cosmogenic nuclides and zirconium-normalised cation loss balances) co-vary primarily with physical erosion rates and much less with temperature or precipitation. (5) In some areas of high land use short-term rates (from river load gauging) exceed those from cosmogenic nuclides by several orders of magnitude, which serve to highlight the severity of geomorphic change caused by human action. In the future, the control mechanisms over denudation will be determined on all spatial scales, ranging from the single soil section to entire river basins. The same analysis can be done back through time on well-dated terraces, lake records, and marine sediment cores, which is possible with 10Be for the past 1–2 My. The rates obtained will be used to develop a quantitative understanding of tectonic, geomorphologic, and geochemical landscape processes, which in turn is a prerequisite to design and calibrate models of the response of landscapes to tectonic, climate, and anthropogenic forcing.