Look at this molecular network. It is so nice. And it is quite large too. The pores are more than 2nm wide. Each hexagon is build from 6 dimers of Angiotensin II. Please look in our latest publication for the precise structure (http://www.nature.com/ncomms/2016/160112/ncomms10335/abs/ncomms10335.html)
The idea to build these networks and tailor their structure and ultimately their properties is around for a while. However, when you only have the molecule, it is next to impossible to predict the structure or the properties you will get. This gets even worse, of the molecules are large and functional. And then, in addition, the synthesis get more and more complicated, if for instance one would want to include a small modification.
Not so with peptides. Here you can order a sequence online and because the synthesis is universal and done by robots, they are quite cheap. In our paper we saw that we can do a similar trick as nature: We change the sequence and get a different structure. Not too surprising, true, but we can now work a lot of different sequences and maybe finally learn how to design molecular networks.