Reflection on Feed Forward

I have this little browser plugin of my own creation. It filters out content I don’t like. It substitutes terms I don’t care for with the ones I feel comfortable with. Most importantly, it continues to learns my taste each time I click the small, barely noticeable button that says I don’t like this. Behind the curtain, it operates on several ordinary machine learning techniques that would, with every ounce of effort, extract features from the preference I’ve explicitly supplied. The result is most satisfying. No more pusky nonsence or eye-rolling schnanigans. No more horrible Mandarin loan words slapping me right at my face. If peace has ever existed, I am with it now.

This plugin, as Hansen describes, has directly altered the sensory experience I gain from online matter (and their causal efficacies), and therefore influced my moves and decisions. Even though my consciousness is aware of such intervention of the evident kinds, it may and is more than likely remain ignorant of some more subtle and nuanced effects. My consciousness may realise that it feeling glad is a result of absence of disagreeable content, whilst having no knowledge of it dispising a certain product being a consequence of failed replacement of abhorring Mandarin words. I have no access whatsoever to the process inside the plugin even though I designed it. The flow of sensory data creeking under the table is simply non-existent to me, yet it has affected almost every aspect of my online, even offline moves and decisions.

This could be even more unsettling in the case of some widely installed systems that operates on the same principle. Most social media and news agencies have recommendation systems that learn users’ taste and decide what to show them. Similar systems utilise the data, sensory and experiential, from these twenty-first-media, study, manipulate and exploit them so that they could control the final moves and decisions entailed by those data. Consciousness doesn’t even stand a chance now that the mere existence of such intervention (or more precisely dictation) is lost to it, let alone the process and the data they operate upon: the browsing history, the time lingered on one page and how these lingerings are distributed, everything that is recordable piling up so high. These are all inaccessible to us either in scale or in sensory capacities. They are not only the loyal scriber of our past and current conscious experience, but also, in ambitious hands, fed into and shape our future conscious experience. They have opened a window to the future, predicatable and forseeable future founded on the solid, copious sensibility captured as machine-processible data, collected, catagorised and studied.

Facebook’s advertising system

Another thing that pops into my mind whilst thinking about such systems are the famous brain in a vat thought experiment. To some extent it is the ultimate form of such system, where everything outside the carrier of consciousness itself are opaque to it. The brain still feels, experiences and interprets the decisions and moves afterhand, only that none of the sensibility are real. But then what would matter if the efficacies and the data are real? After all they are just media through which sensibilities evoke sensory experience and ultimately moves and decisions that consciousness is aware of. Twenty-first-century or twenty-second-century, print or electronic, no longer matters.

CMT: Unthougth Reflection – TA

The idea of silicon-based life has been present in popular culture (particularly in science fiction) for the better part of a century.  The forms this type of life has been depicted by has been varied, from crystline geodic entities to robots and other synthetic life.  To help sift through these many conceptions of silicon-based life, I will use the ideas from Catherine Hayles’ recent book Unthought to select the most probable form of silicon-bsed life.  Specifically, this post will focus on cognitively-sophisticated life at least on the scale of humans and other animals, and how these sophisticated forms of cognition “emerge from underlying material processes” (p65).

Material Processes
One of the most common occurances of silicon in nature is in quartz crystals.  Interestingly, quartz has an interesting electrical property in that it can generate an electric current when put under tension and pressure, and conversely it will vibrate when an electrical current is put through it (how’s that for vibrant matter!).  This is known as the piezoelectric effect, and the implications are impressive; a specific arrangement of quartz crystals under the right conditions could produce an interesting pattern of electrical charge, even responding to pressures from its environment.  In fact, the bones in a human skeleton use the same peizoelectric effect sense forces being applied to it.

On the other hand we have the synthetic silicon technology that makes computers possible: transistors.  Transistors cannot concievebly occur naturally, but they provide a silicon-based platform for cognition.  I will not go into the details of how transistors operate, but the various applications of digital technology are evidence enough for the power of these microscopic components.  It should be noted that neither quartz crystals nor transistors are cognitive in themselves (and certainly not living), but sophisiticated arrangements of these two silicon-based components allow much more sophisticated phenomenon to emerge.

There are plenty of definitions of life that have been proposed, but I will go with a common sense definition and modify it with a stipulation of Hayles: life is the “ability of organisms to endure through time, construct as well as interact/intraact with their environment, and deploy agencies that are not merely emergent but also intentional, even when nonconscious” (p70).  Neither computers nor macroscopic robots can persist through time because neither of them can repair or reproduce.  This is one of the greatest disadvantages of our silicon-based techhnology today, and is likely to remain unchaged for the foreseeable future.  However, at the microscopic scale things are different.

Nanorobots may be the most likely route to artificial silicon-based life.  In the same way that humans do not appear fully-formed on the surface of the earth, nor are they assembled piece-by-piece by an external force, neither must nanorobots be create by external intervention.  Nanorobots can be the silicon-based analogue of biological cells, duplicating and organizing into complex systems and even cognitive and concious entities.  This ability to replicate and self-organize is also what elevates nanorobots beyond the sum of their material parts, or as Hayles describes: “The differences between material forces whose actions are deterministic and hence can be calculated precisely as the sum of the relevant forces, and those that involve self-organizing, chaotic, and complex dynamics and whose actions can lead to the emergence of increasingly complex outcomes, including life and cognition” (p81).

An important consideration for this future possibility is how communication will occur between humans and these new lifeforms.  Although nothing might be known about what form such a lifeform might manifest in, there is a method by which we might estimate our mutual capacity for meaningful communication.  The Sentience Quotient (SQ), proposed by Robert Freitas in 1984, takes the information processing speed of a cognitive assemblage divided by the mass of that asssemblaged, transfered to a logarithmic scale.  To illustrate, we can say that the human brain achieves a +13 SQ, and is 10 times more efficient than IBM Watson with +12 SQ.  Most animals, even insects like ants, fall within a few SQ points of human beings.  Considering this, any lifeform with an SQ beyond few points from human SQ (in either direction) would have difficulty having meaningful communication with us in the same way that we cannot have meaningful conversation with an ant.  As I said, IBM Watson is currently only 1 SQ point behind us, and by the time we achieve silicon-based life it is reasonable to suppose such life will be capable of SQ’s much higher than that.  It might never be possible for us to communicate with other kinds of life forms, even if we have created them ourselves, and that leads to a whole new set of questions about synthetic life-forms.