Embryo models as experiments in self-organization: The logic of productive failure

Stem cell-derived embryo models help to transform inaccessible stages of mammalian development into experimentally tractable systems, revealing how development succeeds, stalls, or fails under defined conditions. Because natural embryos are shaped by stringent in vivo selection, they conceal many trajectories that embryo models bring into view. In mouse and human systems, embryo-like structures emerge through shared morphogenetic transitions-cell sorting, polarization, lumen formation, and symmetry breaking-with variable efficiency and along multiple routes. These alternative trajectories expose the rules and constraints of self-organization, the roles of extra-embryonic tissues, and the parameters that stabilize robust development. We propose that embryo models can be understood as experiments in productive failure, in which divergence from in vivo development becomes mechanistically informative. In doing so, embryo models reveal the hidden grammar of development, expose the fragilities that underlie pregnancy loss, and potentially reshape what is possible for regenerative biology and bioengineering.