People have predicted the pervasive use of smart matter - nano-technology with computational ability - in the future. This will be painted on the surfaces of everyday environments, and the suggestion is that these surfaces will be used as video displays, user interfaces, and sensor arrays. The implication is that a coat of `paint' will be enough to add this functionality to a surface. The field of amorphous computing provides one possible realisation of this vision; an amorphous computer is a multitude of identical tiny computers with local communication capability, and could be `painted' onto a surface to form an ad hoc network. An iSurface is a specialised version of an amorphous computer, using a simulator with a physical grounding. In this paper we argue that such a vision of Intelligent Surfaces is fundamentally flawed. The load, both in terms of processing and communication, on the individual elements will always be a hindrance to the responsiveness of iSurface applications, and we present experimental results on several of the expected key features of this approach to demonstrate this. (10 pages)