In this text Negroponte talks to us about the idea of making an architectural machines, not only machines that can solve specific problems with step by step instructions where the result is unquestionably attributed to the designers creativity, but a machine that is cable of learning how to learn. This machine should be able to learn about architecture and have a dialogue between itself, man and other machines in order to produce an evolutionary system.
The evolutionary machine should be self-improving in order to have better chances of making its computational and informational abilities relevant, to achieve these the machine needs to learn to be adaptable. According to Negroponte the 5 particular subassemblies: would be part of an architecture machine:
1) A heuristic mechanism – heuristic is gained from analogue situations where the machine drastically limit the search in order to reduce time and search for alternatives of solutions.
2) A rote apparatus – a rote mechanism can retain the circumstances of usage when similar situations are next encountered. This helps solving repetition of sub problems in architecture.
3) A conditioning device – repetitious responses will become habits, the machine will respond with a combination of rote apparatus and a heuristic mechanism to develop its own conditioned reflex.
4) A reward selector – the designer overviewing the machine must exhibit a response of his satisfaction with the design in order for the machine to know when the operation has been successful.
5) A forgetting convenience – The machine should learn to forget. Obsolescence occurs over time with irrelevance.
These five items are just pieces, the entire body will be a group of mechanisms interconnected with the designer and a parent machine. The parent machine will have a powerful processor and extensive memory, the reasons to have a parent machine are to acquire large burst of computational power, acquire stored information and communication between designer and other architecture machines. Negroponte states that in order for the machine to have a design solution 3 properties are necessary: an event, a manifestation and a representation, these 3 properties form the interface. The event is how we perceive the situation with our senses, the manifestation measures the event with appropriate parameters and the representation is the act of mapping the information to be compatible with receiving organisms.
Negroponte speaks of an idea really ahead of his time, he proposes the idea of artificial intelligence and a profound connection and relationship between man and machine. The idea of a machine that can perceive what humans perceive with their senses it’s a possible topic that interests me for my personal research, finding a way or a link how humans can transmit or machines can perceive human emotions because that’s part that machines lack in the design process. The designer sometimes makes decisions based on emotion or sensations that may not be understood or rational to the machine, in order for this solution to not be obsolete for the machine, the machine needs first too able to learn and measure emotions in order to approve these solutions.