The engine of complexity is one of five great books that have made a revolutionary impact of my thinking on nature. Other 4 books are `Pi in the Sky' by John Barrow, Order out of Chaos by IIya Prigogine `End of Time' by Julian Barbour, and `The Emperors New Mind' by Roger Penrose.
I was initially reluctant to purchase this book because of popular strong AI claims in computer science that algorithms is the be all and end all to intelligence and I was not certain about a relationship between Algorithms and Complexity.
However it was clear to me that automobiles have been evolving over the last 120 years in just the same way as life has been evolving over the past 3.5 billion years and for this reason I decided to purchase this book.
The foundation of the book is based on the distinction between two different types of physical structures.
1) Those that require a code of instructions (a recipe) in order to exist. (coincidence of structure is ruled out as exceedingly highly improbable).
2) Those that do not require a code of instructions to exist.
Fig 7.1 in the book illustrates this superbly and hopefully will become iconic as E=mc2.
Compare life forms/life fossils and man made structures (including economies and institutions etc) on Earth to all other structures on Earth.
The Engine that drives complexity in nature is when these instruction codes are processed in space/time while dictated to by the laws of nature and that these outputs are copied back to the inputs with a degree of error (ideally the error is a random element so the search is not constrained within the search space.)
It is the Laws of Nature that dictate this complexity search. It is not that Engine of Complexity searches for Code that create just any physical structure, but all the successful searches code for structure that take most advantage from the Laws of Natures. For example if the instruction code resulted in the construction of a lighter with no flint, then the complexity code for this lighter would be poor and not last long (ie. will not be reproduced unless another use was found for it).
However if over 3.5 billion years, results that codes for a body that holds a large brain with the right physical structure to utilise natures laws which allow 'thinking' to take place then this complex code has the potential to be very successful and regarded as complex in our universe.
In this book I feel that John Mayfield uses the term `Computation' too loosely.
A procedure that inputs several alternatives, processes these inputs and outputs only a selection of the inputs is basically the procedure for the Engine of Complexity.
The set of selected outputs are often more complex than the set of inputs. They are often more well-matched to utilising natures laws for durability and procreation.
However the processing part of the procedure may be a computational process or it may be a non-computational process.
In terms of natural selection, the processor is the environment and the inputs are DNA sequences. For the `Engine of Complexity' as a computational process one would have to obtain a code for the complex environment, the part that is relevant to the selection of the DNA inputs (i.e. possibly a subset of the universe). If this is not possible in principle then the process is a non-computational process.
The same engine drives complexity in those computer programs that utilise the same principles, but in this case the search space is more limited, constrained by the physical functioning of a computer and the coded selection factors. By virtue of being processed by a computer the selection factors must have a code and therefore the `Engine of Complexity' within the confines of computer processing is a computational process.
When we look about at the Grandeur of life of this refined 3.5 billion year evolutionary search, it is not necessarily so that the complexity generated over this time, that we experience about us, is simply the result of some Galactic Digital Computer churning through some Galactic digital algorithm, as the subtitle of this book may suggest.
It may have been more appropriate to subtitle this book as, "Evolution as Information Processing".
The Engine of Complexity is encapsulated within The Laws of Nature.
The fuel that drives the Engine of Complexity is an Entropy Gradient.
The Entropy Gradient of the Universe can be defined as all the different states of the Universe, from the Big Bang through to a state of uniform radiation in the extreme distant future. At this current instant the entropy of the Universe is much closer to the former than the later.
Molecular Motion and chaos is also encapsulated in the Laws of Nature
As we scribe through an Entropy Gradient, Motion and Chaos abound among the astronomical but finite number of particles in the Universe. The majority of these particles/molecules interact creating more complex arrangements via Quantum Mechanics and chance. (see
Life's Ratchet: How Molecular Machines Extract Order from Chaos
, by Peter M Hoffmann.) (Update 2019 - Must read Stuart Kauffman 'A World beyond physics - The emergence and evolution of life.' This book give a highly credible and experimentally verifiable theory regarding the emergence of life and pathway to complexity as referred to in No. 1 below.)
This is a source for `Novel Creations'.
So Nature has two Pathways to Complexity:
1) Chaotic chance ratcheted along a molecular entropy gradient - (slow and achieves only a low upper bound to complexity)
2) Engine of Complexity - (super fast and no limit to Natures complexity's upper bound)
A single cell life form on planet Earth is an example we have of the `Engine of Complexity' in the Universe.
When we envisage the tree of the `Engine of Complexity' on Earth we have things like Humans, dolphins, religions and economies and the arts at the top of the canopy while at the very base we have the original Engine from which all this complexity was later formed - (at least that is what all the data seems to indicate. As to date we have not found two fundamentally different life forms on Earth).
However even this primary Engine required a considerable amount of complexity in order to function. It is possible that the complexity of the original Engine on Earth developed via Path One. It could not have developed via path two.
Statistically it is interesting to imagine the odds for a potential life forming planet to form in the Universe since the time the Universe first started forming planetary systems several billion years ago. Is Earth one of only a few places in the Universe that has the potential to form life? That is; statistically should we expect look out into the sky and to see the billions of galaxies in space (as opposed to a handful of stars) from our vantage point because the odds are so tiny?
If this is the case and only a small limited number of potential life forming planets exist in the universe what are the chances of life actually forming? If planet Earth is the only life forming planet (at least in our vicinity of our galaxy) then if Earth were merely 1 km in diameter would we be here today discussing such matters? Would the search space for all the series of extraordinary coincidences necessary to materialise a life form on a planet, in the period to date, be large enough for life to initiate in the limited time that has passed?
What are the secondary ‘Engine of Complexities’ (path two) that have formed since the first replicating cell? Chance is responsible for the original engine. What complexity engine (or engines) has this original complexity engine formed in the 3.8 Billion years of life?
To date we know of only two (discounting biological immune systems): One appeared around 40,000 years old the other less than a decade old.
These secondary engines are so complex that chance alone could not have possibly produced them (even within the search space of the whole Universe covering the last 13.8 billion years).
Anthropologists witness evidence of Human creativity from artwork and culture dating back several thousands of years. On page 255 Mayfield states “One thing we can be for certain of is that without some sort of random input, creativity is an illusion”.
Creativity generates many different types of scenarios for a given challenge within the informational processing space of the brain. Nature generally arbitrates on the value of these scenarios that we create. The brain functions at a frantic pace compared to natural evolution, operating 24 hrs a day for several decades.
Today results of creative outcomes are recorded and collated in libraries and now the on the internet. When information is communicated between many millions of brains for further informational processing - complexity is generated at an extraordinary rate. Our biological memory and communication methods have been enhanced by our man-made physical communication and recording devices. These have now become essential components of the engine. The inclusions of the written language, recording and communication devices supercharge this ‘Engine of Complexity’. Hence the society, the arts, technology, sciences and culture we find ourselves in today compared to just 100 years ago.
The other secondary Engine of complexity is just starting to appear today. Computers are now sophisticated enough and our software development advanced enough that we are able to code computers using neural networks that emulate the way we predict how the human brain processes information. The code makes small changes to a problem and is programmed for what goals to aim for. In this way a computer is continuously computing information (learning) similar to our subconscious brain. This is how the computer code ‘Deep Mind’ recently defeated the AlphaGo World Champion.
If devices like these start to incorporate physical sensors and become networked together their influence on our planet could become quite substantial (and possibly scary).
The Engine of Complexity is optimised when:
1) the error size (mutation) is optimised approx (1/n where n is the length of the input code string).
2) The error is a random element (that is; not limited in the search space)
It is interesting that a genuinely random element is crucial for the Engine of complexity to perform effectively. This directly relates quantum mechanics to complexity as Quantum mechanics is the only generator of purely random elements in nature.
The book also points out that our laws of nature conform to the Engine of Complexity (which is obvious else we would not be around to discuss such matters). But anthropically (if there is such a word) it is interesting to note that the Engine of Complexity does not work for all mathematical structures (possible other universes' laws of nature).
John Mayfield is a master on this broad subject and has succeeded in explaining the ideas essential for an understanding of this concept to a layperson. A joy to read.
ASIDE:
Although Mayfield doesn’t go into it, it is interesting to ask how long this engine of complexity can run for and whether or not we are running out the necessary ingredients of fuel and sample space?
Fuel:
In terms of fuel, we are talking here about the abundance of free energy and how depleted the universe is at this stage in terms of entropy. In our current evolution of the universe the entropy is still very low and there is plenty of free energy to fuel the engine of complexity.
Sample space:
Well, this is different. Cells, neurons, ants and bees for example have been selected for under laws of nature. These selection criteria highly attract entities that are able to cooperate in groups and function as single whole - as meta entities.
During the last 200,000 years “group selection in terms of both time and space” has become more important as a selection criterion - a powerful complexity multiplier.
Sometime in homosapiens past homosapiens had built up a repertoire of beneficial frames within the network of the brain about comprehension of the senses and the environment that they operated within. The level of complexity of the brain, at this point in time, reached a threshold where the brain was able to build a frame about thinking itself - thinking about thinking. At this stage the entity had the capacity to look at itself as doing the thinking - as an individual with existence in the world that is doing the thinking.
At this point there must have become a strong selection pressure on these individuals to relate their thinking “now” with those thinking entities that share the same memories with those entities projected back into the individuals past - up to the time of the first memories sequenced backwards. The time of what the individual now perceives as their birth.
This entity was also able extrapolate its thinking (entities that will share a consistent memory sequence) into the future. The first instance when a human became vexed by its own finite mortality. This individual now views all of these thinking sequences as a whole unit - the sensation of a personal me flowing through the aether of time.
They now plan and prepare to maximise the survival for these future thinking entities that they consider to be themselves. Entities selected for with this frame of mind were more likely to survive, have and nurture their offspring, creating a positive feedback. It should be noted that this beneficial mindset had to be selected for - with other mindsets discarded and washed away by the Engine of Complexity.
The group time selection above, now, like ants, developed groups of family and tribal units where they could cooperate as a whole each individual with some speciality (mental and physical) within the group. This group, unlike ants, includes the time dimension as well. A group with a universal neural network that is able to reflect upon itself and plan ahead of time - able to adapt quickly to many different and chaotic environments. We spread across the world.
Once we developed techniques of farming and cropping, we developed into sophisticated cities, individuals and institutions with their own specialities with information past down from generation both orally by the printed word.
Countries grew naturally as communication by horseback allowed a given area to cooperate/trade as a beneficial whole. Countries and districts were subject to intra and inter community competition which prevailed for thousands of years. A threshold in intelligence, a mindset for group unity and stability were meet where certain communities on the planet were able to develop resilient, long-term science-based knowledge acquisition. These communities thrived on our planet at the expense of others as per the Engine of Complexity.
Today the media enterprise and educational establishment which drive and direct our democratic political systems in the West can now be viewed as one global unit. Countries generally now have open immigration policies blurring the diversity between countries and communities (cultural and intellectual). Most democracies have social systems that encourage/support single parents and the poor to provide generations with their respective future populations. With modern global satellite/cable communications, shipping and airline industries we have for all intents and purposes become one global entity with the life-support of our living planet dependent on our actions.
If intelligent life is extremely rare in our universe - perhaps the only space in our universe where this can be found is on planet earth - then the engine of complexity has run out of sample space. If the Engine of complexity runs out of sample space for competition, then it grinds to a halt.
It is possible that with the makeup our current democratic systems we have unwittingly bred general populations that are unfit (morally and intellectually) to evaluate and elect heads of government capable of navigating natures intricacies required for survival of our species along with the survival of many other life forms on earth. For example, we may not have the capacity to vote in leaders with the intellectual and moral sophistication to avoid catastrophic climate change or runaway nuclear war.
If this was to occur then it is very unlikely that some future advanced intelligent lifeform develops later on to carry on the advance in complexity, if not only because that we have used up the easily accessible free energy sources available on earth. It is unlikely that the new intelligence could jump from animal driven power directly to solar, wind or nuclear.
Consider the hypothetical situation where a similar earth type planet with similar advanced life forms and social structures exists. On this planet too social structures have formed into countries with communication channels and a transportation system utilising the horse. These countries also have developed the printing press and a science based educational system. However, the political systems are slightly different. These countries value their individual culture and celebrate their unique arts, culture, language and architecture. Each country has their peculiar idiosyncrasies and characteristics. They value and nurture their gene pool which they consider as a precious national asset - an asset that is vital for their group survival. They also have a national democratically elected government. In order to vote you must have a clean criminal record and have achieved a deemed level of university certification. Candidates that put themselves up for election must have proven past experience which is assessed for moral certitude and incorruptibility. They must have a degree in politics, a general knowledge degree which on this planet includes significant amounts of history, science and philosophy. The financial system tightly regulated and fair.
It could be that this life form manages to survive for hundreds of thousands of years and transform/inhabit other planets each planet evolving in their own uniquely diverse way.
If this were the case then the Engine of Complexity would not be depleted of sample space (although it did pass through an extremely narrow bottleneck) and continue well on into the far future.
無料のKindleアプリをダウンロードして、スマートフォン、タブレット、またはコンピューターで今すぐKindle本を読むことができます。Kindleデバイスは必要ありません。
ウェブ版Kindleなら、お使いのブラウザですぐにお読みいただけます。
携帯電話のカメラを使用する - 以下のコードをスキャンし、Kindleアプリをダウンロードしてください。
何か問題が発生しました。後で再度リクエストしてください。
著者について
著者をフォローして、新作のアップデートや改善されたおすすめを入手してください。
他の国からのトップレビュー
John McIntyre
A revolutionary book that must be classed as one of the classics of the 21st Century.
2013年11月24日にアメリカ合衆国でレビュー済みAmazonで購入
Pancho
Continuous improvement = evolution
2013年12月24日にアメリカ合衆国でレビュー済みAmazonで購入
Tim Tones
It is not easy, but it is highly recommended for anyone trying ...
2015年8月10日にアメリカ合衆国でレビュー済みAmazonで購入
wvfgolden
Chewy
2014年6月8日にアメリカ合衆国でレビュー済みAmazonで購入
Harley
An awesome read if you are interested in systems behavior
2014年11月8日にアメリカ合衆国でレビュー済みAmazonで購入
