In education and research at this time there is arguably no longer a need to build a strong case for the power of CAAD to support designers - the evidence is there. The major challenges no longer centre on hardware, software and graphics potential or on skills acquisition and adoption. The research that we will report here reveals that computational complexity and geometrical complexity are emerging as the sharp issues that demand a major review of how we model the large hybrid spaces that we seek to construct in design. Computational relational modelling and scripting may have opened a trove of creative possibilities. But it may delude us into painting ourselves into a corner: infinite variety within a much reduced palette of opportunities.