What leads you to think that the templates are necessarily very complex? In my reading, the models requires the presence of polymers, let's call them A, B, and

What leads you to think that the templates are necessarily very complex? In my reading, the models requires the presence of polymers, let's call them A, B, and T; and a cyclical mechanism that causes these polymers to alternately bind to each other and release each other (through chain to chain ionic or hydrogen bonds). The polymers themselves are formed of chains of smaller monomers covalently bonded to each other. A, B, and T are floating around, and occasionally bind to each other. The nature of our template T is such that both A and B can bind to it at the same time. Importantly, T doesn't need to be much larger, or more complex, than A or B, just able to bind to one of the ends of each. So we have our polymers floating around looking something like (this is my purely made up notation): *A*: AAaAaA *B*: BbbBbBB *T*: TTaAabBBT They move into the night cycle, causing them to bind to each other with ionic or hydrogen bonds: AAaAaA BbbBbBB ||| ||| TTaAa-bBBT Due to the ends of the A and B chains now being so close to each other, they form a covalent bond, resulting in a new polymer chain C: *C*: AAaAaABbbBbBB |||||| TTaAabBBT The day cycle comes around and the template breaks away from the newly created chain. Thus the template is acting like a catalyst, able to create many identical copies of this polymer C. C is a more complex polymer than either A or B and could easily become a template itself. It's not much of a stretch to then consider a polymer which either was a template for itself, or was part of a sequence of templates that produced each other. To your complexity point, it doesn't seem obvious that these templates are necessarily "fairly complex" - indeed if you admit the existence of polymer chains in the first place the ability for one of them to act as a template seems trivial.