Early embryogenesis seems to activate weird genetic pathways, whose function scientists suspect evolved for the sake of complexity alone, to be a genetic integrity stress test.
Mind you, mutation in sexual reproduction isn't at all just accidental "bit flips". The cross-over stage in meiosis always recombines the chromosomal information randomly (to increase environmental adaptability, off-setting the slow rate of sexual reproduction). There is no concept of a "gene" there, the cuts can happen in the middle of anything important, too. This intentional instability needs to be accounted for, biologically.
Before anyone even detects pregnancy, a lot of non-viable genetics have been aborted by the body. As tragic as they are for the becoming parents, miscarriages are a good thing, not misfortune. It's exactly this fail-fast selection process weeding out failed random recombination attempts, which likely would have caused severe functional problems later anyway.
The body isn't a lump of omnipotent cells, but a complex system of functionally primed cells. You cannot do a selective genetic stress test on a developed organism like that. Without radical selection, the rejuvenation process wouldn't be safe, rather a guarantee for cancer and metabolic dysfunction. At any point in time, there are malignant cells somewhere in the body, which are constantly snatched by the immune system. In fact, "rejuvenation"/immortality is a hallmark of cancerous transformation. And let's not forget, embryonic development isn't fully self-governed. The mother's body also checks the embryo's signaling and may abort anything funky. A developed organism doesn't have this external oversight. Auto-immunity and genetic checkpoints is all there is and those mechanisms can become malignant, too.
Did you know, with age, men's somatic cells increasing lack the whole Y-chromosome? It's not at all mission critical and just got lost at some point during mitosis, no biggie, but I think it shows the extent of accumulated genetic disarray in somatic cells.