There's no obvious reason you couldn't use a carbodiimide for this coupling. What method of amide coupling is used in your 4-step procedure you mentioned?
Similar, we reduced acid to dialdehyde and then just mixture of dialdehyde, 2,2´-(ethylenedioxy)bis(ethylamine) and MnCl2.
My prof told me that it works from ester without carbodiimide just by mixing the 2 reagents to get dipeptide and then reduce the oxygen which is ,according to him the most difficult part, because, reduction by LiAlH4 doesnt give good yields (Im not sure why not, it is used in literature for doing realy similar reductions). Same goes for kirshner-wolf reduction. I guess reduction with zinc amalgame in HCl would be good. Do you know any other metods of reducing carbonyl to CH2?
You have a bifunctional electrophilic intermediate and a bifunctional nucleophile. Would the macrocycle be favored over polymer formation? I don't know.
MnCl2 should help with that as it gets coordinated by 2,2´-(ethylenedioxy)bis(ethylamine) and "bends" the shape of the molecule to favor macrocyclization.