Multisense Realism

A great presentation on computing with hardware and software which are homomorphic to each other. Logical automata (explained 11:51 to 15:50) is just this sort of a WYSIWYG architecture where the software is executed from a bulk raw material, i.e., not Gb of memory or number of processor cores, but square feet or pounds of programmable matter. Atoms, or groups of atoms, are here used directly as bits.

Gershenfeld compares how our current approach of computing requires multiple stages of compiling disparate formats, but that to more closely match nature, we should strive to imitate nature in the sense of having a consistent zoomable format on every scale.

In nature, the map is the same graphically regardless of the scope of magnification. Marrying these two concepts, the idea is to design shapes which are themselves instructions, i.e. physical interactions with other physical shapes, so that computation is not encoded but rather, embodied. As in biochemistry, the output is a material product, and the machine is itself is a self fabricating machine tool as opposed to a manufacturer of inert objects.
I think that he is probably right that this is the direction of the future in general, “computing aligned with nature” which brings computation into matter. It is compelling to imagine that this kind of embodied computing could be the Holy Grail of nano-engineering, giving us control over virtually anything eventually.
At the same time I can see that there is something which has been overlooked. To quote Deleuze:

“Representation fails to capture the affirmed world of difference. Representation has only a single center, a unique and receding perspective, and in the consequence a false depth. It mediates everything, but mobilizes and moves nothing.”

– Gilles Deleuze, Difference and Repetition, p56
(source http://sleepinginthegreenery.blogspot.co.uk/)

To understand why this universe of embodied computation is not the universe that we live in, Difference is the key. An overhead map is only so useful to us. Even if we can zoom down to the human scale level, we really need to switch to a first person street view to make the transition from outside-looking-down to inside-looking-around. To get further into the subjective view, we would have to have access to feelings and thoughts, so that at some level of description the model of zoomable shapes is less than useless. In our personal awareness, the appearance of neurological structures in our view would be hallucination.

The assumption then of this uniformly computational matter, while fantastic for our purposes as human beings, would be a catastrophe for the universe in general. It would be the ultimate monoculture, with everything and anything reinvented as collections of positionable nano-Legos. The problem of conformity to a generic, universal structure is not that it won’t work, but that if it does, there will be no Difference possible.

Given Deleuze’s assertion that representation moves nothing, this intention of “reuniting bits with atoms’ seems to presume that they were united to begin with but doesn’t address why they were ever separated. Indeed, if this method of embodied computation is the way of nature, how and why could it ever seem otherwise? Where have these Different perspectives on different levels emerged from, and for what purpose?

I think that on closer inspection, even though this new approach is brilliant and revolutionary in some important ways, it still is founded on a sloppy assumption. It presumes that hardware which looks and acts like the software is identical to it. If this were the case, instead of images we would see the shapes of ganglia and retinal cells, but we would see the same thing instead of smells, sounds, and feelings. We could not feel dizzy, but rather be informed of some vestibular condition by means of these same shapes – which we think of visually or tangibly, but without the Difference, there really is no reason to assume anything perceptual at all.

Once again, even though I am impressed with the futuristic thinking, it still takes us away from the missing piece in physics – the privacy and interiority and qualia. Buying into the universe as undifferentiated plenum of self-machining bubbles we are betting that there is no difference between biology, chemistry, and physics. It’s all physics, all surfaces and volumes in public spaces. Is that really what the universe is though? Could our experience be understood that way if we didn’t have our own familiarity with it already? I think if we lived in a universe that was really all about universal computation, then we would never have separated bits from atoms in the first place. We never would have approached bits as encoded abstractions because we would have been comfortable already with the universal format.

Instead, the universe appears to be the opposite. On every level, even though there are repeating themes and forms, it is never exactly the same presentation. A whirlpool would not be mistaken for a galaxy and a grain of salt is not the same thing as a cube of ice. In the universe we actually live in, the only thing which seems truly universal is Difference. More universal than mathematics and physics is the variety of sensory qualities and modalities. It is not just formations or embodied information, but direct experience.