Information Becoming Aware of Itself By: Brent Peters Published: March 20, 2014 The network of neurons in our brains somehow grants us the conscious ability to learn and grow, so the pattern itself is self-optimizing. As I discussed in the previous post, its branching structure – what I refer to as the Constructal Pattern – also seems to be a structural feature of the universe. Intelligence emerges from this same cosmic pattern – intelligence that has the ability to learn from (and adapt to) its environment… A living mind that emerges from a nonliving pattern. Understanding cognition means understanding the Constructal Pattern. Because of its unparalleled ability to produce emergent phenomena like consciousness, we really need to explore its potential — which is the purpose of Obama’s BRAIN Initiative, a scientific endeavor to research how we “think, learn, and remember” – secrets whose answers are locked within the constructal pattern. Consciousness: more than the sum of its partsClick to show The constructal pattern is essentially the fingerprint of life – and it’s found everywhere. To return to a question I raised earlier, does this make the universe one giant, self-optimizing neural net? Matter forms networks. Networks form patterns. Patterns spawn consciousness. Nature’s Ability to Self-Improve Nature improves with time. Powered by the process of natural selection, natural systems always deviate toward improved flow (better efficiency, survivability, resilience, etc). This means the best natural designs flourish, whereas outdated or less efficient designs dwindle with time. I discussed in the Exponential Series how the process of natural selection (specifically with regard to biology) has been circumvented by modern civilization and replaced with an information economy of competing ideas. Still, the one commonality that remains is improvement over time — a persistent evolution towards better flow. [1] This concept of natural selection is paralleled in technology as recursive self-improvement. Both natural and artificial systems self-improve: Nature does it through evolution, and technology does it through the process of iterative design. However, this second process must be consciously guided — our technology doesn’t just improve on its own (at least not yet). I began to describe in the last post how the interplay between nature and technology can be interpreted as the process of evolution itself. To rephrase, imagine this chain of events: Matter is subject to the laws of physics. Over time, natural forces shape matter into meaningful patterns. The Constructal Pattern that results eventually spawns consciousness. The intelligent life that emerges has the ability to make new technologies that further accelerate its own growth. (A second derivative of growth) Eventually, technology is designed in such a way that it gains the ability to self-improve (“The Singularity”). Designs become self-improving without the need for a designer, essentially recreating the conditions of Step 1 (“unattended”, or natural optimization). Here’s the cool part: To design a self-optimizing, intelligent network like the one I’m describing, we must peer into the origins of our own consciousness (i.e. the Constructal Pattern) and replicate the same designs. This means we must learn from the structure our own brains (nature, the Constructal Pattern) in order to build what we consider to be a highly advanced technology (AI).[6][7][8] The Technological Singularity is the postulated event when technology gains the ability to automate its self-improvement, bypassing the need for manual (human-controlled) design.[2] Imagine a technology that can continuously re-adapt to its environment and optimize itself – a technology that can learn, grow, and get better with time. Surprisingly, such a technology already exists: Nature. Because natural networks are self-improving and grow without any obvious form of guidance, they demonstrate a form of unsupervised learning. Nature and the Technological Singularity Because natural networks already demonstrate many key aspects of the Technological Singularity (conscious, self-improving, massively parallel), nature can be understood as a fully manifested Technological Singularity in its own right. In other words, the universe we live in is a self-improving technology; it has spawned a pattern that live, learn, and adapt to its environment. The self-improving technology we dream of already surrounds us. If the Singularity means self-improving technology – and if we view nature as a self-improving system – then at least one Singularity has already occurred, and we exist within it. Thus, I think we need to rethink our understanding of the Technological Singularity and what it means. Rather than being a purely “technological” event, I expect the Singularity to be a time of rapid convergence between nature and technology – a time when human technology fully optimizes itself for its environment (i.e. the laws of physics) – which implies mimicking the Constructal Pattern, a natural phenomenon.[4] In other words, we need to correct the notion that the Singularity represents a separation between technology and the natural world. In fact, it represents an even closer union. Footnotes More discussion here. Regardless of whether it’s biological evolution or the evolution of an idea. Terminology gets a bit funky here. Since the Constructal Pattern is natural, if we replicate the same pattern to build AI, then the consciousness that results isn’t “artificial” (having emerged from a natural pattern). This essentially means there is no such thing as “artificial” intelligence; only general intelligence. This begs the question; what defines technological complexity? Because we have so much to learn from the brain, it is in many ways a superior technology to anything we have today. This cycle does not necessarily repeat. Perhaps the fate of intelligent life in the universe is to merge with technology and enter a new state of being that we cannot fully comprehend. 2020 is the most popular prediction for when this might happen. Today we have machine learning algorithms that can be optimized for specific uses, but aside from specialized applications, self-improving technology is outside the scope of our present abilities. If the constructal pattern represents the best optimization given the laws of physics we experience, then as technology becomes more efficient with time, its design will approach the same pattern. ### Brent Peters attends the University of California, Santa Cruz, where he is a fellow of the Global Information Internship Program (now the Everett Program), which is an organization that focuses on bridging the digital divide through projects that empower communities with Information and Communications Technology.[1] He will graduate in 2014 with a degree in Global Information and Social Enterprise Studies (ICT/Sociology).
Posted on: Thu, 20 Mar 2014 17:40:56 +0000
Trending Topics
Recently Viewed Topics
© 2015