Despite impressive advances, adult stem cells remain less promising than embryonic stem (ES) cells for treatment of numerous human conditions, including age-related ones. The pace of ES cell research has, however, been slowed by three critical difficulties. First, the embryo from which the ES cells are derived will not be totally immunocompatible with the recipient (even if HLA alleles are matched). Second, isolation of ES cells involves destroying a blastocyst about five days post-fertilisation; even though that is well before embryonic structures emerge and twinning or chimerism become impossible, many believe that this is destruction of a human life. Third, human oocytes are in short supply, since adult volunteers cannot donate many oocytes each. Somatic cell nuclear transfer (SCNT) potentially confers immunocompatibility, but is no help regarding the other two problems. SCNT using non-human oocytes theoretically solves problem three, but not in practice, because nearly all the oocyte's mitochondrial DNA (mtDNA) is then non-human and cannot function in concert with the newly-introduced human genome. However, various subtleties of mitochondrial biology combine to side-step this problem, and also to avoid the ethical issues surrounding SCNT (problem 2). The technical problem is two-fold: first, deriving "pseudo-ES cells" from the "pseudo-zygote" product of an inter-species SCNT procedure, and second, replacing the mtDNA of these otherwise differentiation-incompetent pseudo-ES cells with patient-derived mtDNA, so turning them into potentially therapeutic, bona fide ES cells. I will explain why both those problems may be much easier than existing work might imply, if established mitochondriological principles are applied. Regarding the ethical issues, the pseudo-zygote is unambiguously not a human embryo much more unambiguously than, for example, one generated using the "altered nuclear transfer" procedure recently proposed by Hurlbut because it is not only unable to proceed through the earliest stages of embryonic development (even in an ideal environment, such as hypothetical uterine implantation) but also unable to give rise to cells that can perform any natural function in vivo. It is thus far less ethically problematic than a normal zygote, So is the pseudo-blastocyst that is destroyed in isolating the pseudo-ES cells; the eventual bona fide ES cells are derived only by a subsequent procedure. Finally, the "yuck factor" associated with inter-species chimeras does not strongly apply here, because the enucleated non-human oocyte is merely an unusual type of container in which the human nuclear genome is reprogrammed to a totipotent state.