The thesis of this book by Braid is courageous but potentially brilliant: many of the attributes that we associate only to high-level organisms (which, to put them into one, are the capacity of creating an internal image of the external world) are actually present also at single cell level - in this liquid computational processing environment called wetware.
If you can excuse the fact that the style of Braid is boring, I guess partly because of the intrinsically hyper-descriptive character of biological trainings and expositions, then this book ranks as a potential gem. The author's opinion is essentially a corollary of recent developments and evidence of the computational power of chemical networks, particular of protein networks, enshrined and compartimentalized within membranes, that is within cells.
2 main reflections are in order, if the conjecture reveals true: 1) the central nervous system, to which we attribute the origin and control of such supposedly high-level behaviors, is not necessary to them, thus not probably its origin, as well, but more probably a refinement and structuralization of the sets of feedback and logical loops already working at molecular level; 2) mimicking molecular biology (liquid, carbon-based, energy-efficient, resource-efficient yet robust) in the perspective of artificial intelligence or, more simply, enhanced artificial computational power has still a lot to progress to reach evolution's achievements - though they can be back-engineered, eventually.
Recommended to newcomers and to those fascinated with networks in search for biological substrates to investigate.