A more structured summary of what I found interesting at the Crystalpunk workshop for soft architecture in Utrecht last week-end:

Pablo Miranda's talk targeted two aspects of architecture: being critical and projective. As for the critique part, a lot of things in architecture comes from the assumption that one can do stuff by drawing (before, people were building directly). The use of computer application has complicated this situation: new tools (3ds max, maya, catia...) gives a very important flavor to what's been designed: so who is going to be the author of a project built using such a tool? the architect or the designers of the tool? Concerning the projects, he presented those of his group which are amazingly interesting, especially those playing with genetic algorithm in architecture.
Adam Somlai-Fischer's point was about 'bionic spaces", not smarthomes. Sdam is against smarthomes, makes you dumb and incapable to interact with reality (which I definitely agree with, I hate the crappy intelligent fridge idea that may be refilled when empty). For him, architecture is not the building itself, not the behavior, it's not a dress code BUT the idea that space that surrounds us is a responsibility AND it's a social process. He thinks that now that social software are paving this way (web2.0, flickr...), we have familiar ground to replicate this in architecture. He then presented his incredible projects.
Jelle Feringa elaborated on Pablo's critique about interactive architecture. For him, each architect style is bound to the software they used (marcus novak - mathematica /greg lynn - maya kas oosterhuis - virtools / frank gehry - catia). His critique was also that specific software designed for architects are regressive. For example piranesi is very limitative: the software creates romantic-sketchy-cheesy models that make architects bound to old-fashioned model to create architecture. For him, archiCAD is also regressive in the sense that it helps quickly resolve design dilemmas and site-constraints unique tu strict municipal-building codes and client requirements through its interactive abilities. One of the most interesting project he presented was a chair design produced with genetic algorithm. They started by asking a question: "what is the minimum volume defining a chair" which they try to answer using genetic algorithm techniques. For Jelle, this was a new way of designing by addressing a specific question before creating an artifact.
Jonas Hielscher (Z-25) presented his project Dat-a: an installation based on RFID technology. The dat-a project allows to track people in an exhibit using rfid (you can see people's name in different rooms). He said: "we are in effect creating and Internet of Things". He and Pablo also mentioned this interesting fact after my presentation about blogject: we will get enormeous piles of data, how would we do deal with that (data mining gets hot as Jonas said).

Thanks Wil for all of this!