Given all this cause for pessimism, and the weight of scientific consensus apparently behind it, I was pleased to read that a prominent neuroscientist -- especially one who's been an outspoken critic of both Ray Kurzweil and IBM's artificial-brain projects -- had recently predicted a transhuman revolution. I was even more pleased when I realized the source of that revolution was analogous to a computing technology I'm studying for my master's thesis: general-purpose computation on graphics-processing units (GPGPU).
Transhumanists believe that the abilities humans will gain in the future, usually by becoming cyborgs, are as incomprehensible to "unenhanced" humans as most human behaviour today is to animals. Mark Changizi takes this analogy a step further -- he says superhuman powers will arise from the same phenomenon as the human-specific abilities that archeologists call "behavioural modernity" and Mark calls "Human 2.0".
That phenomenon is called
Apparently, this explains why activities such as language use more or less the same areas in all human brains. If a large part of the brain were "fully programmable" (the traditional explanation for why we have those abilities), Mark says the layout would depend on the individual and be more or less random. Instead, he argues that reading and writing make use of the same hard-wired processing in the visual cortex as hunting and gathering, explaining how so many people could become literate in the 4,000 years or so since the first alphabets were invented. (Apparently, his book Harnessed discusses this argument in more detail.)
If you're familiar with GPGPU, you'll probably notice the analogy as I did. Like neuronal recycling, GPGPU is about reframing hard processing tasks so that specialized hardware built for another purpose -- in this case, a video card -- can be programmed to do them more efficiently than more general-purpose hardware.
Just as many gaming computers and consoles turned out to have more processing power on the video card than on the CPU, so does the human brain turn out to have more processing power dedicated to sensory and motor processing than to the functions associated with "thinking". This is called Moravec's paradox.
This is why non-linear relationships between two statistical variables are hard to detect mathematically, but easy to describe once you see them on a scatter plot. Instead of the thousands of neurons you've trained for math, the scatter plot lets you use the millions in your visual cortex that recognize two-dimensional shapes. Scatter plots and other data visualizations, when they work well, are probably good examples of neuronal recycling, and I tend to think of them as a primitive form of mental GPGPU.
|Mental GPGPU: faster and more accurate than the correlation coefficient.|
If true, Mark's hypothesis about the transhuman potential of neuronal recycling will put an interesting twist on an old trope of science fiction: the myth that "you only use 10% of your brain". While neuroscientists have been disproving and discrediting the trope for years, they're now reconstructing it in a twist ending. The "other 90%" is being used... for unnecessary sensorimotor processing, and now you can use it for thinking instead. Sometimes truth is more Troperiffic than fiction!
|Daniel Tammet (Image via Wikipedia)|
Apparently, this gift is more common in blind people than others, so I bet it too is an example of neuronal recycling (and/or neural reuse, since autism affects both biology and receptiveness to culture) of the visual cortex. Blindness, or in Daniel's case a combination of autism and epilepsy, just happens to enable a more powerful form of recycling for math than anyone's been able to manage using charts and graphs. Maybe the transhuman revolution Mark's describing is one in which we'll all gain gifts similar to Daniel's, and analogous powers in fields other than math, without having to become blind or autistic first.
I guess it will be the dawn of a blue day.