TOWARDS A THEORY OF
ARCHITECTURE MACHINES
TEXT ANALYSIS AND SYNTHESIS

In “Towards a Theory of Architecture Machines”, Negroponte proposes the idea of creating a system or a “machine” that is able to have a learning and evolving mechanism. He proposes that by embedding the learning mechanism, the machine will be free from just doing repetitive tasks, and will instead continually analyse itself and find more effective and efficient methodologies to arrive at the solution. As a result, the evolving relationship between the architect and the machine will allow more time to be spent in the creative process.
The assembly of the learning machine would involve 5 processes/components.
_1// a heuristic mechanism – limitation of the search for the solution – hence reducing the time for alternatives
_2// a rote appartarus –
“storing of the event or a basic part of an event” – creating a definition for an instant predetermined response.
_3// a conditioning device
creating habits from repetitive events – which would lead to a somewhat reflex response
_4// a reward selector
criteria for selection depending on the preferences of the “teacher”
_5// a forgetting convenience
ability to discard redundant functions due to irrelevance.
Eventually, the relationship between architect and machine would be extended with a somewhat control system that would allow the machines
DISCUSSION & IDEAS
The expert group discussions branched in many streams, regarding to the concerns of authenticity, loss of a human element within the design process, the new role of the architect and other ethical topics.
One of the major concerns of the discussion was the recognition of the authenticity of the design. If the creator makes a machine that is able to make a submachine that would create the final product, who receives the credit of the design? If the system is perfectly constructed and established when will it realise that it no longer needs the creator? The growth rate at which the machine is able to compute information and learn from other architectural machines to disseminate experiences and stored information between other machines and architects, hence creating more efficiencies and effectiveness in machine processes.
At the point where this process becomes somewhat autonomous, new layers of complexities and functions can be added to the overall machine. The new layers revolves around its new found ability to sense and evaluate changes in its surroundings. Negroponte describes the process in 3 steps, first involving the recognition of an “event”, second the “manifestation” of the series of functions that it requires to respond to said event, and third would require the “representation” of the process.
would arguiably be greater than the sum of whole entire combination of all future potential, traditionally educated human architects. The creation of such a machine will be the end of the traditional architect as we know it. This effect can already be perceived in the shift from handdrawing to CAD design. The architect population within the worldwide industry dramatically reduced as tasks were now able to be completed at a faster rate, and with less resources than normal, conventional drawing methods.
This shift is already re-occuring in the emergence of parametric design. Endless of iterations of design can be generated at a such a speed that would normally take months and years to complete normally.
Is what Negroponte proposing, the new shift after parametricism?
The discussion soon shifted towards the measure and comparison of the loss of authenticity with the new level of growth and progress, to a level that would exceed that of the current exponential growth that we are experiencing today. If this growth is misguided, it could mean the end of the world as we know it. I believe. however that this level of growth is necessary to reverse all the social and environmental that we have inflicted over the years.
Is this then the new role of the architect? A shepherd, a guide, a driver and a leader that would represent the needs and desires of society and mother nature?
If this was to be the case, what sort of criteria or vision should we impose on the world to ensure the collective is represented accordingly?
Architects would then be purely dedicated to the generation of ideas and implementation of acceptable moral ethics rather than the execution and construction of the idea. The genesis of an idea that is transcending requires a lot of experimentation and exploration into the unknown which may lead to the infinite loop of philosophical discourses.
As architects, we should be concerned with a simple question,
Where is the threshold where growth will result in the eventual destruction of our world?