#The_Singularity_is_Near: When Humans Transcend Biology - Ray_Kurzweil - #Review

     I have read this book nearly 20 years after its publication and more after finishing a second volume to this sequel (when we merge with AI) before the first - also note the change in the title from "Near" to "Nearer". Some of Kurzweil’s technology trends have clearly come true, but the central “Singularity” has not yet happened. His predictions that came out true are in computing, networking, mobile devices, and the accelerating usefulness of AI. His timelines for nanotechnology, brain–machine integration, radical life extension, and fully immersive virtual reality appears to be much more speculative still.

     His claim that roughly US $1,000 of computing would approach human-brain-level computational capacity around 2020 is not verified yet, because “human-brain computation” is not a single agreed-upon measurable quantity. The trend was directionally right, but the exact comparison is uncertain.  He had also predicted that machines would surpass humans in increasingly broad areas. That has happened for many bounded tasks, including image classification, some language benchmarks, visual reasoning, and several forms of game-playing or pattern recognition. Occasionally, a few prestigious institutes have released studies connected with this topic that are still to reach out to the public. Like this one: University of Innsbruck has built a quantum computer using a single atom of calcium that has the potential to simultaneously encore dozens of qubits.

     Kurzweil predicted that AI would become powerful in medical image analysis. A notable example is his claim that neural networks would match human doctors in radiology image analysis; a Stanford milestone reportedly achieved that around the time he predicted.

    The book also predicted medical nanobots that would travel through the body, repair disease, interact with cells, and eventually connect with the nervous system – including a bot that would release a virus inside. We have nanoparticles, targeted drug-delivery systems, molecular machines, and some nanomedicine, but not general-purpose autonomous medical nanobots of the kind described.

     Kurzweil has also maintained a prediction that AI will reach human-level general intelligence or pass a meaningful Turing test around 2029. Current systems are impressive, but they still show weaknesses in long-horizon reliability, novel reasoning, calibration, and real-world action. The book’s underlying computational trend has proved remarkably influential and partly accurate; its more dramatic biological and nanotechnological promises have not yet been demonstrated.

    Evolution results in better answers, not necessarily more complicated. Noise is inherently unpredictable but carries no information, however, is also unpredictable if we can predict future data from past data that feature data stops being information. Perhaps his second volume takes cognizance of the fact that the predictions have not yet materialised and would take some more time and the way to it more linear than exponential not to mention about the quality of human mind that has become addicted to the technology.

Here are some excerpts from his narration, without any rationalisation of the views:

    There are only 800 million bits of information in the entire human genome and only about 30 to 100,000,000 bits after data compression is applied. Compression not achieved.

    We are able to see only a very small portion of the visual field with high resolution. The portion imaged by the fovea in retina is focused on an area about the size of a single word. At 22 inches away outside the fovea we have very little resolution. Thus, a VGA array is detected by our eyes as motion and causes constant moment of the fovea. This slows down reading speed, which is one reason that reading on screen is less pleasant than reading in a printed book. So flat panel do not flicker – Flat panels?

    Elsewhere in the book- Hans Moravec has analysed the transformation performed by neural image processing circuit in the retina, which is 2 cm wide and 1mm thick. Most of the retina’s depth is devoted to capturing image, but 1/5^th of it is devoted to image processing. The human brain is about 75,000 times heavier than the 0.02 g of neurons in this portion of the retina so that there are about 1014 instructions per second for the entire brain. Our emotions are closely linked to areas of brain that contain maps of the body.

    All of biology operates through linear sequence of two-bit DNA base control the sequencing of only 20 amino acids in proteins.  In so far in the human brain would require and the order of 1 billion billion bits which is compact human and genome. The entire genome consists of 800 million bites, but most of it is redundant. Only 30 to 200 million bits is useable which is smaller than MS word program.

    Basic neural transactions are several million times slower than contemporary electronic circuits. Non-biological portion of our intelligence will be trillion of trillion of times more powerful than unaided human intelligence and hence can perform repeatedly at high speeds with optimal accuracy without tiring. Human progress is exponential rather than linear. So?

    Two proteins zonolin and zot have been discovered the react with receptors in the brain to temporarily open the BBB at select sites and also in small intestine where they allow digestion of glucose and other nutrients. Unique patterns of proteins found only in the presence of cancer. Actin is the chemical in the brain, that shapes, structures, and regulates communication throughout neurons. Can these be evaluated from the blood samples alone?

    Wolfram provides evidence to show how increasing complexity can originate from universe that is at its score, a deterministic algorithmic system. There are three great philosophical questions. What is life? What is consciousness and thinking and memory, and all of that, and how does the universe work? The evidence is like there is an animal to find.

    IBM has developed about 250 GB of data per square inch in 2003 demonstrated working memory device using polymers that self-assemble in 20 nm wide signal structures. The classic example of a practical use of quantum computing is in factoring very large numbers. Factoring numbers with more than 512 bits is currently not achievable on a digital computer. Factoring is finding which smaller numbers when multiply together result in the larger number. We cannot create a 20-qubit computer simply buy company 2 10-qubit machines. All of the qubits have to be Quantum entangled together, and that has proved to be challenging.

   Processor speed is related to voltage and the power required is proportional to the square of the voltage. So, running a processor at slow speed significantly, reduce power consumption, parallel processing is advisable. Human brains use about 100 trillion computers in the form of neuronal connections where most of the processing takes place, but these processors are very slow in computational Power and thus run relatively cool. What is the present situation with regard to the data centres now?

     When a bit of information is erased, that information has to go somewhere. According to laws of thermodynamics, they erased bit is essentially released into the surrounding environment, thereby increasing its entropy, which can be viewed as a measure of information. This results in higher temperature for the environment. The reverse is also true, so logical operations such as NOT could be performed without putting energy in our taking heat out.

     Charles Bennett showed the any computation could be performed using reversible logical operations. The fundamental concept is that if you keep all the intermediate results and then run the algorithm backwards when you finished your calculations, you end up where you started, have used no energy and generated no heat, however, you have calculated the result- Uncertainty?

     An ordinary rock may appear silent, has nothing much is going on inside, but approximately 10 trillion atoms in a kilogram of matter are actually extremely active and atoms are all in motion sharing electrons changing spin, moving electromagnetic fields. All of this activity represents computation if not meaningfully organised!

    Just like you can't make a boy and girl fall in love with each other simply by pushing them together, you cannot make a process in chemistry occur as desired between two molecular objects with simple mechanical motion. This way enzymes do!

    Solar panels in space could convert space sunlight to electricity and can be sent back to Earth by microwave through satellites. Self-assembling electronic devices will take less energy to manufacture and may produce few toxic by-products than conventional semiconductor manufacturing methods. Indication towards the present energy crisis.

    If you understand something in only one way, then you don't really understand it at all. This is the basis of strong AI. Robotic telescopes are able to make their own decisions on how to look and optimise as they are able to detect very subtle phenomena, such as star blinking for an hour or a second, which may indicate a small asteroid in the outer ranges of our solar system.

    Anthropic Principle: only in our Universe that allowed evolution we would be able to ask questions.  We can have a drug in 50 years that will lowest us to eat as much as we want without gaining weight. This has been accomplished in mice. So, let’s borrow some time.

    Equivalence Principle: we should not use the brains in efficient coding methods to rate one’s memory efficiency. Analytical voice versus neuromorphic modelling of brain, chess problem. Humans recognise pattern while machine is built over huge logic, countless of moves and counter moves – machines are learning still and have a look at self-driving cars first.

    In a hologram information is stored in a different pattern throughout an extensive region. Even if 3/4^th is destroyed the image is retrievable. This is similar to the reason how memories can remain intact. Memories are dynamically distributed throughout region of neurons memory don't fade.

    If protein molecules can fold improperly, they must be destroyed, else they will accumulate. Accumulation of these misinformed proteins, aggregate into particles called Proto fibrils, which lead to Alzheimer disease.  Ubiquitin, a tagged carrier molecule breaks down into component amino acids. Haemoglobin is created 500 trillion times each second in the human body with more than 500 amino acids in each molecule, it comes to about 15 billion billion read operations every minute and only four bases provide the digital storage for all the complexity in human life – Thus we are still copying.

    Biological systems can fly, but flying at 30,000 feet and thousand per hour uses modern technology. Likewise, brain can remember, but to do large calculations technologies is required. As we get older, the heart is unable to replace its cells at sufficient rate, so it compensates by making cells bigger using fibrous material. Why is this there?

    A relevant example is the 20 years it took to derive clinical benefit from monoclonal antibodies -- the rate-limiting steps had little to do with computational complexity. So, the notion that, in the future, completely real biological, physiological, and ethical constraints will simply melt under the blaze of increased computing power is fundamentally misguided.

Some sharp liners:

# As soon as an AI technique works, it's no longer considered AI.

# AI is the study of how to make computers do things at which at the moment people are better.

# Fat insulin receptor (FIR) is the reason why animals remain slim.

# Photons measured twice the speed of light as a result of quantum uncertainty.

# The principle role of all religions has been to rationalise death.

# Modern medicines, progress depends on chemistry, physics, biology, computer science.

# I understand English, but none of my neurons do.

# Theoretical calculations to produce 100 W from blood of one person has been evaluated

# Technology is a way of organising the universe, so that people don't have to experience it.

# Common sense is not a simple thing. Instead, it is an immense society of hard and practical ideas of multitudes of life, learned rules and exceptions, dispositions, and tendencies.

# My body is temporary. Its particles turn over almost completely every month. Only the pattern of the body and brain have continuity

# Seth Floyd estimates 1080 particles in the universe and the info in the universe in its surface area divided by square of Blank constant is 10120 bits which cannot be decoded

# You wouldn't find consciousness, even if you blow mind to the size of universe, our legal system is based largely on the concept of consciousness and attention is given to action that cause suffering

# Moore's law is running up against the quantum limit, so micro-processor size is exponential up to a fast-approaching limit.

# Fredkin, believes that the universe is very literally a computer and that it is being used by someone or something to solve a problem- God

# A phonebook has a lot of information, but the level of order of that information is quite low. Human beings represent and extremely intricate and elaborate hierarchy of other patterns.

# The reasonable man adapts himself to the world, the unreasonable one persists in, trying to adapt the world to himself. Therefore, all progress depends on the unreasonable man!

    The major claim in the book is that brains will merge with computers and that our only problem need be to produce an intelligence greater than our own. Once this is accomplished, all other tasks can be up to that greater intelligence.

Nevertheless, this book can open you up to several issues on various fronts that require attention and then you can be with this ‘singularity’.