Scientific American has published a letter from Ray Kurzweil in its January 2004 issue in response to an editorial in the September 2003 issue critiquing Kurzweil’s vision of reverse-engineering the brain and the future of machine intelligence.

“The ultimate goal is not for us all to become cousins of the Terminator or Max Headroom,” stated the editors. “Rather it is to correct neural defects and to take normal people (whatever ‘normal’ means) and make improvements from baseline …”

Ray Kurzweil’s response:

I appreciate your commencing your special issue on “Better Brains” with reference to my views on the upcoming “marriage of the biologic and the cybernetic” despite your skepticism. I will note that a primary source of our different outlooks on the prospects for brain reverse engineering is a different time frame.

The special issue describes well some of the neuroscience advances now in development, innovations that we can expect to benefit from during this coming decade. We need to ask, “what happens after that?”

Progress will not only continue, but its pace will continue to accelerate. Casual observers expect the current pace to continue (with a generation of technology now spanning on the order of five to eight years), but this view ignores the inherent quickening pace of technology. The reason for the acceleration is that each stage creates more powerful tools to enable the next stage.

Consider for example that the spatial and temporal resolution of brain scanning technologies is improving at an exponential pace (one of many examples is the in-vivo scanning system being developed at the University of Pennsylvania, which is designed to resolve individual neurons in a cluster of up to 1,000 simultaneous cells with sub-millisecond temporal resolution, a dramatic improvement over current systems). According to my models, we are doubling the paradigm shift rate (rate of technical change) approximately every decade, so we can reasonably anticipate a dozen generations of technology over the next three decades.

Scientists are trained to be conservative in their outlook and expectations, and this translates into an understandable reluctance to look beyond the next generation of capability. When a generation of technology was longer than a human generation, this orientation served society’s needs well enough. However, with the rapid acceleration of progress, looking ahead a single generation of technology is no longer sufficient. The public has a legitimate interest in informed opinion that looks forward 20 to 30 years.

If you do the thought experiment of considering the implications of multiple generations of technology, the availability over the next several decades of enormous increases in the capacity of our computational and communication tools, the advent of molecular nanotechnology, and far greater insight into the principles of operation of the human brain, I believe that our perspectives will converge.

– Ray Kurzweil