They make the point that rotary joints are for conceptual illustration, and it could equally use hinge or flex joints. This would be relevant at the nano scale,

They make the point that rotary joints are for conceptual illustration, and it could equally use hinge or flex joints. This would be relevant at the nano scale, > In particular, a molecular version of this architecture could use stiff covalently-bonded nanotubes for the links and single bonds for the joints. In that case joint friction would not be a consideration, or at least I would imagine things like electrostatic or molecular bonding forces would be larger.