What does it take to make computational/parametric design matter?

...and not just the latest formal fetish? 

a humble model built in David Mah's "fabricating terrains: advanced landscape as digital media" workshop...

Reflections on the recent Computational Design Symposium at the GSD: 

Products of the building boom, none of the invited speakers were able to justify their use of the tool beyond essentially a formal gimmick. Now that these tools have been integrated into the design mainstream (although let's not forget their origins in the cold war military industrial complex), it's up to us to make them relevant. The novelty has worn off, though the power of these tools are too great to ignore. How can these data driven design processes interface new and established ecologies?