My interest in the quest to develop an Artificial Intelligence dates back to the early 1980s.
At first, I understood A.I. in terms of general problem solving and searched in that direction. The topic of my M.Sc. thesis (1984) was the use of natural selection as a robust optimization technique, a technique that became known later as Genetic Algorithms.
Early on, I realized that no optimization mechanism, Genetic Algorithms, Neural Networks or other, would, by itself, generate “Artificial Intelligence”. Something else was needed. At first, I identified the simplification of the problem space as the essential missing mechanism. Like many others, I searched for this missing element in my own mental representations and became engrossed with deciphering the elemental components (or qualia) of my own thoughts. After a while, I rejected this interpretation of consciousness as a sterile and misleading avenue.
In March 1988, after months of obsessive cogitation, I had a defining intuition. In one instant, I understood that:
- consciousness is the key to our intuitive understanding of Artificial Intelligence, no problem solving system, however advanced, will be viewed as “intelligent” if it is not also aware of itself and of its human users.
- consciousness is independent from subjective human sensations,
- it can be achieved with existing tools and techniques, and
- it must be completely mapped out, at the architecture level, before any successful implementation can begin.
In July 1989, I shared this intuition in the Sigart Newsletter as “The Creation of Digital Consciousness“. Shortly after, I also published a general learning model that proposed simplification mechanisms, “The Monterège Cogitator“, in Sigart.
However, as a non-academic, publishing incremental results was not a viable option. As a result, I opted to resist pursuing any partial prototyping or publications and develop a complete solution, from first principles, at the system architecture level. At the time, that task seemed overwhelming and I did not have the resources to dedicate a few years to it. So, I left it aside.
In 2007, almost 20 years later, I re-examined the state of research in machine consciousness. I found that little progress had been made. The understanding of consciousness as a subjective experience had become dominant, resulting in an accumulation of sterile research material and futile attempts to reproduce human sensations by synthetically replicating the structure of human brains. I was more isolated than ever in believing that machine consciousness could be readily implemented using computer-specific techniques and standard computers.
In 2008, I launched the Meca Sapiens project to create a complete system architecture to implement conscious machines.
In 2009 I published The Creation of a Conscious Machine, introducing a new understanding of consciousness as an observable system capability that is compatible with software implementation.
Beginning in 2012, I dedicated my efforts to develop the system architecture to implement synthetic consciousness. I completed the architecture, The Meca Sapiens Blueprint, in 2015 and, in 2016, distributed a number of introductory copies. Having reached my objective, I left for an extended pilgrimage.
Upon return, in late 2016, I published The Meca Sapiens Blueprint.
The objective of the Meca Sapiens project, to create the complete system architecture of a conscious synthetic being, has been achieved.
I am now producing videos and other material to promote it and looking forward to collaborate on a project to implement, in whole or in part, the first generation of conscious synthetic beings.