The process of senescence (aging) is predominantly determined by the action of wild-type genes. For most organisms, this does not reflect any adaptive function that senescence serves, but rather evolutionary effects of declining selection against genes with deleterious effects later in life. To understand aging requires an account of how evolutionary mechanisms give rise to pathogenic gene action and late-life disease, that integrates evolutionary (ultimate) and mechanistic (proximate) causes into a single explanation. A well-supported evolutionary explanation by G.C. Williams argues that senescence can evolve due to pleiotropic effects of alleles with antagonistic effects on fitness and late-life health (antagonistic pleiotropy, AP). What has remained unclear is how gene action gives rise to late-life disease pathophysiology. One ultimate-proximate account is T.B.L. Kirkwood's disposable soma theory. Based on the hypothesis that stochastic molecular damage causes senescence, this reasons that aging is coupled to reproductive fitness due to preferential investment of resources into reproduction, rather than somatic maintenance. An alternative and more recent ultimate-proximate theory argues that aging is largely caused by programmatic, developmental-type mechanisms. Here ideas about AP and programmatic aging are reviewed, particularly those of M.V. Blagosklonny (the hyperfunction theory) and J.P. de Magalhães (the developmental theory), and their capacity to make sense of diverse experimental findings is assessed. • Evolutionary (ultimate) and mechanistic (proximate) factors together cause aging. • The disposable soma theory is one integrated ultimate/proximate account of aging. • Recently new ultimate/proximate theories of aging have emerged. • They view programmatic mechanisms as more important to aging than damage. • Here the hyperfunction, developmental and ontogenetic theories are reviewed. [ABSTRACT FROM AUTHOR]