Home » Blog » Africa » “The problem with computers is that there is not enough Africa in them.”

“The problem with computers is that there is not enough Africa in them.”

Long-time TEDster Jay Walker uses a three minute talk to issue a challenge to the TED community: help figure out how to create at least ten million new jobs. He argues that “if you’ve got capital tools, you’ve got jobs”, suggesting that a person with a sewing machine rapidly becomes a sewing business. Mobile phones are a capital tool with a billing platform built in, which have already sold 1 billion units. What sort of services can people with mobile phones and no training sell to the world? Storytelling? Jokes? Interpretation? Language practice? Chris Anderson notes that Jay has given this talk two years previously and is evidently still looking for the right solution.

Zip Zap, a circus troupe from South Africa, have a presentation that wouldn’t be well-delivered by mobile phone. They’re introduced by a board member who explains that “social change will come from brave, non-conformist youth.” Two performers, one black, one white, both talented dancers and acrobats, work through a wonderful slapstick routine which involves confrontation between styles and cultures. It’s a great bit and a great break from some of the heavy economics we’ve faced the past two days. The circus sounds like a tremendous institution, taking kids from all walks of life and giving them a chance to perform together.

Kwabena Boahen tells us that he encountered his first computer as a teenager in Accra. He fell in love: “I could play games, I could program in BASIC.” He became fascinated with how the machine actually worked and started reading about the internal systems. He was fascinated that all the data moved through the CPU, and that the CPU had to work like crazy to keep all the data moving in and out of short and long term storage.

People weren’t very worried about this bottleneck, because computers were so fast. In 1946, Alan Turing was able to predict, “In 30 years, it would be as easy to ask a computer a question as to ask a person.” Why aren’t we seeing computers challenge the brain? This is the focus of Boateng’s research.

Boateng tells us that we pay a huge price for that speed, especially in terms of power. Blue Gene, the IBM supercomputer, uses 120,000 processors to process 10 quadrillion bits per second. In the process, it consumes as much power as 1,200 Us households. By comparison, the human brain proceses a similar amount of data in a similar timeframe, but uses only 10 watts, about as much as a laptop computer.

Why is the brain so much more efficient, using a hundred thousand times less energy than a computer? Boateng believes it’s all in the wiring. Computers put all data through a central bottleneck, either a CPU or an I/O processor. Brains, on the other hand, are networks – each neuron connects to hundreds of other neurons, and there is no central bottleneck. Computers are serial, rigid, while the brain is parallel, fluid. Or as Brian Eno said in 1995, “The problem with computers is that there is not enough Africa in them.”

Boateng looks closely at how transistors work, in comparison to ion gates in the brain. At current levels of miniaturization, roughly 12 electrons can get through the gate of a transistor. But as miniaturization increases, we’ll hit a point where only a single electron can pass by 2015. At this point, we end up with highly unreliable signals – the gate wants to close and deliver zero current, but an electron is trapped in the gate, or vice versa. This sort of gate unreliability happens all the time in the ion gates of the brain – it’s possible for ions to jump a channel even when it’s closed. But the brain functions because there’s massive redundancy that compensates for the imperfection of gates.

Boaten wants to build a computer that works more like the brain. Specifically, he’s researching a system to replicate the function of the retina, “a bit of the brain that lines the back of the eye.” The system tracks light and dark contrast, as well as movement, and creates a visual image that’s more similar to the way the human brain works than most existing processing systems.

He ends with a picture of a steel drum – what happens when you put Africa in the piano. Boaten’s goal is to put Africa in the computer, generating thought, creativity and dynamism. It’s likely a long road from the artificial retina to an African computer, but Boaten is walking that road.

2 thoughts on ““The problem with computers is that there is not enough Africa in them.””

  1. I like the title.. Although computers these days are quite powerful they are no match for the brain. Maybe in the not so distant future, computers would work as fast as a brain would work. Anyway, it was nice reading this article. Thanks for this.

Comments are closed.