How We Can Manage Our Way Through the Intertwined Promise and Peril of Accelerating Change
Ray Kurzweil
page 8 of 15
Conservative analyses of Freitas' design indicate that if you replace 10% of your red blood cells with these respirocytes, you could do 
an Olympic sprint for 15 minutes without taking a breath or sit at the bottom of your pool for four hours. He also has a white blood cell design. These are circa 2020 scenarios. I actually watched one of my red blood cells on a microscope. It took an hour and a half to destroy a bacteria; this robotic system would just do something similar in a few seconds.
If this sounds futuristic, I would point out a couple of things: there are already four major conferences on BioMEMS, biological microelectronic mechanical systems. I would not call them nanobots quite yet, but they are capsules, devices that are blood-cell size, several microns in size, with nanoengineered features, that are performing therapeutic functions in animals.
One scientist actually has a very sophisticated device, with nanoengineered seven-nanometer pores, that cures Type 1 Diabetes in rats. It lets out insulin in a controlled fashion and blocks the antibodies, because Type 1 is an autoimmune disease. We can see that the idea of blood cell-sized devices with nanoengineered features in the bloodstream performing therapeutic health functions is not quite as futuristic as it may sound.
We already have devices in the human body that can download software from outside the body. Consider the Parkinson's implant, the latest generation of which allows you to download new software to your neural implant from outside the patient. This is not a nanobot, but if you apply these exponential trends of doubling price-performance of computation and communication, the shrinking technology at an exponential rate, it is conservative to expect that these devices, which are all ready working at some level in animal experiments, will be quite sophisticated by the 2020's.
If we continue that exponential trend of computation through this century, $1,000 of computation will equal even my more conservative estimate: ten to the sixteenth calculations per second for functional emulation of the human brain by 2020. That was a controversial notion in 1999, but it is pretty much a mainstream view today that we will have plenty of hardware computation to emulate human thinking by around 2020.
Image 14: Exponential Growth of Computing
Now the controversy is, will we have the software? The ultimate source of the software of human-level intelligence is really to understand the best example we have of a complex, intelligent system, which is human intelligence. We are making much more rapid progress in doing this than people realize. Chapter 4 in The Singularity is Near is about this issue; why we can be confident in reverse-engineering the human brain. We have made more progress than people realize. Brain scanning is growing exponentially in spatial and temporal resolution. The latest generation can see non-invasively individual interneuronal connections. For example, there is an exciting new technology from the University of Pennsylvania that can see individual interneuronal connections signaling in real time. For the first time, we can actually see not only our brain creating our thoughts but our thoughts creating our brain. As we think about a subject, we are creating new spines and new synapses; we can actually watch that now. We are getting experientially more data about the brain.