History shows that there is a consistent reaction (by some), against new technology. This subset of people often believe that the new technology represents some type of negative shift in thinking and/or production. It is no different with parametric design. Is this an appropriate reaction?
In What contexts is parametric design best employed, and what can it not do? Clearly, when there is a project with complicated, dense and/or fluctuating data, parametric programs can help make live responsive models. As a tool of analysis and quick form generation it works quite well.
However, what types of data are beyond quantification, and therefore beyond parametric design? Can we define qualitatives such as happiness? Ineffable qualities may always be beyond the programs, or our own abilities to record.
Some might also argue that natural systems are too complicated to model accurately. Discrete parts of a system (i.e. wind or precipitation), may be analyzed separately, but the whole system defies modeling. John T. Lyle warns of this danger in his book on regenerative design. Deductive thinking is a great tool of analysis, but we run the risk of missing the wider, connected picture when we try to quantify and categorize everything. I believe this is simultaneously parametric designs greatest potential and fault.