Computers That Work Like the Brain

Computers That Work Like the Brain 

Kwabena Boahen

Kwabena Boahen in his TED talk explains the problem scientists are now running into with developing new computers.  The chips of today are experiencing data bottlenecks in their CPUs which is preventing the flow of data to and from the RAM.  To mitigate this problem Boahen and other scientists are striving to make computers behave more like the human brain.  In the brain, redundant connections between neurons are made in a diffuse network pattern, allowing data traffic to flow freely.  He explains that the fastest computer in the world processes data at 10^16 bits per second, but consumes 1.5 mega watts of power (about enough to power 1,200 homes).  By contrast the human brain contains the same processing power as the world's fastest computer, but only consumers about the same amount of electrical power as a standard laptop.

Boahen makes two interesting sociological points in his talk about improving the efficiency of computers.  First he says (quoting Brian Eno) that the computer should work more like Africa.  What does that mean?  I can only guess that society in Africa functions in a networked sort of way.  There are some corporations and governments working on the continent that try for a linear organizational model of control.  But I imagine the balance of organization in Africa is modeled on free network connections with nodes sprinkled in.  My symbolic vision of this are people talking on cell phones with connections criss-crossing the landscape, doing business and socializing.  The operators are mobile, flexible, can turn on and off (there are very few service contracts in Africa, mostly pre-paid minutes), and most of all efficient.  Western companies are in fact recognizing this form of efficiency over fixed vertical and horizontal cubical farms, and turing people loose as independent contractors (can turn them on and off), working from home (mobile).  The problem there is that imposing this model on Western firms is like imposing the linear model on Africa, its simply too much change all at once, at the whim of some arbitrary organization.

The second interesting sociological point that Kwabena makes is that as a boy growing up in Tanzania he immediately recognized the inefficiency of computers' inner workings.  He saw a disconnect between how data is processed and how his society operates.  Not only does this say that the society he grew up in either was efficient enough to work, despite the difference of opinion of outside NGOs like the World Bank who try to build up institutions in places like Africa as they see fit, but that those society's contain the resources needed to attack problems from different angles which could be beneficial.  It was refreshing to have the opportunity to hear Africa school the "developed" world for a change.