?? so for certain class of filteredly cocomplete category, we apparently have some reasonably nice sort of somewhat systematic embedding of it into the topos of filteredly cocontinuous set-valued functors on it ... ??? .... ?? let's see, covariant or contravariant, or does it matter ?? ... hmmmm .... ???? .... ?? seems like ... ??? covariant ?? ... ?? which would be opposite of yoneda embedding, i think ... though situation seems confusing for various reasons ... anyway we have reason to think "yoneda embedding" doesn't really exist here ... ?? ....

??? so .... does this alleged embedding curry into something covariant in both of 2 variables ??? ..... ???? .... "global sections of tensor product" ??? .... ???? .... ???? ..... ... ??? .... ?? non-toric analog ??? ...... ????? .....