Creative Retirement - Friday Roundtable -- Aug 29, 2014
III . The Singularity , 2040
D R A F T
Sheldon video : singularity, immortality.
YouTube version
III. THE SINGULARITY, 26 years.
One big name predicts the singularity to occur around 2045 whereas another predicts some time before 2030. The 2012 Singularity Summit, included a study of artificial general intelligence predictions by experts and found a wide range of predicted dates, with a median value of 2040. (Range between 2015! and 2045)
A child born today will be just reaching his career competence, 5-years out of college.
Singularity Summit is a multi-day program addressing the latest developments and disruptive influences of exponentially growing technologies. Together the world's thought leaders and participants explore areas such as biotechnology & bioinformatics, energy & environmental systems, networks & computing systems, AI & robotics, medicine & neuroscience, and nanotechnology.
Singularity University. http://SingularityU.org/ . Mission is to educate, inspire and empower leaders to apply exponential technologies to address humanity's grand challenges. It takes an optimistic view up through the Singularity
Let an intelligent machine be defined as a machine that can surpass all the intellectual activities of any man however clever. Since the design of machines is one of these intellectual activities, an intelligent machine could design even better ultra-intelligent machines ; there would then unquestionably be an "intelligence explosion", and the intelligence of man would be left far behind. The first intelligent machine is the last invention that man need ever make.
(a) Is general world intelligence growing ? Birth control among the knowledgeable and unrestricted population growth by the lower classes, suggests that world IQ - will fall | is falling | has fallen .
(b) And people are worrying about climate change that will be barely noticeable by the end of the century.
(c) You could still be alive during the singularity.
Computers are adept at large problems such as weather predictions, climate simulation, to track the human genome, project the flight of artillery at 20 miles which was the purpose of the first U.S. computer. The first computing application was by Colossus at Bletchley Park to crack the Nazi Enigma code in 1944. It was kept unknown by the British Official Secrets Act for thirty years while the American Eniac (1945) took credit. The Iowa State ABC computer of 1938 had 300 vacuum tubes, but was stored, then thrown out and never did anything useful. Colossus had 1,500 tubes ; Eniac had 18,000.
Robots are desired for dangerous jobs and for service jobs that entail expensive one-for-one interaction with humans such as long term nursing care.
TURING TEST
The Turing Test sets up a question and answer session and if the human cannot tell that the remote party is not human, then human level intelligence is considered to have been achieved. This test was defined in the era of teletypes. I think this must be upgraded from the original example of teletype to a video conversation by Skype with full visual and audio. Another way might be to ask questions where a human would not know the answer. Spin-off tests are The Loebner Prize for best chatterbox ; Lovelace test of creativity.
Advanced AI
Advanced AI by definition will be self aware to provide for its needs. Will always be improving for :
1. efficiency , 2. self preservation , 3. resource acquisition. Some thinkers say 4. creativity -- I think that is a stretch. The time horizon for a machine is not the quarter-end report of a corporation, but with immortality AI has a lengthy scope of time. We might find a robot spending resources on a space ship to mine asteroids for minerals needed in twenty years. Mankind may resist such use of resources , hence come in conflict with AI goals. Any self-aware entity needs resources. In this robots will compete with men. The rich get richer, the robots win. Best we can hope for is to be two races on the planet dividing resources. Much as Neanderthal and CroMagnon shared Europe. If robots perceive man as a threat, then ... The very fear of robots that could lead to an insurrection would place man on a threat list by robots.
Evolution
Evolution followed a path of necessary steps : cells, to fish, to mammals, to apes, to man. Add the next step, robots. Man was a necessary step in the creation of robots. The dinosaur age ended in a flash, the age of mammoths ended as man crossed the steppes to reach Europe. Elephants and apes, the next most intelligent beings, are endangered: By whom ? From the next step up, mankind. Will mankind become endangered?
Math.
Moore's Law of computer capabilities has been valid for 50 years, it says there is a doubling every 18 months. At what equivalent IQ of artificial intelligence will mankind become obsolete ? Assume IQ=50 comprises the singularity in 2040. Thus leading edge AI might reach IQ 100 in 2042; IQ=200 in 2044, similar to Franklin, DeVinci, Einstein. IQ=400 in 2044, IQ=800 in 2046, IQ=1600 in 2048; mankind constitutes a lower form of intelligence by 2050. Mankind rebels and we are wiped out in the Robot Wars.
Who will do it?
A statement by Google a couple of years ago. "People always assume that we're done with search. That very far from the case. We're probably only 5 percent of the way there. We want to create the ultimate search engine that can understand anything . . . some people could call that artificial intelligence ... The ultimate search engine would understand everything in the world. It would understand everything that you asked it and give you back the exact right thing instantly." Yes, we would call that artificial intelligence.
Some suggest that the program that makes the break-through to self-awareness will be from the nefarious efforts of hackers that will be: 1. advanced, , 2. uncontrolled , and 3. must operate independently while the programmer does his algebra homework
or must attend a compulsory sexual harassment meetings at a secret NSA facility. The anti-malware companies estimate there are more malware programs than legitimate programs. Most by kids, some by governments. The Depart of Defense acknowledges warfare as land, sea, air, space and cyber.
Some take a dim view of this subject: "There is not the slightest reason to believe in a coming singularity. The fact that you can visualize a future in your imagination is not evidence that it is likely or even possible. Look at domed cities, jet-pack commuting, underwater cities, mile-high buildings, and nuclear-powered automobiles -- all staples of futuristic fantasies when we was young that have never arrived. Sheer processing power is not a pixie dust that magically solves all your problems." But I suggest, a lot of things not foretasted have happened. We cannot predict the future.
In 1421, China sent a one-thousand ship fleet to investigate the world, Asia, India, Africa, Australia, Antarctic, America, some say it reached the Atlantic. By the time the fleet returned, the emperor had died, the mandarins had taken over, shut the borders, isolated China from the outside world, and burned the logs of the fleet discoveries.
Today, is it possible that STEM may be condemned and the world turned to art !? Unlikely, put possible.
How will we address unemployment of clerks, tellers, travel agents, loan officers, help desk persons ? Soon after -- doctors, lawyers, tax preparers, retirement consultants ? Joe Bidden says that unemployment is freedom and he hands out checks to enjoy your time off.
We started with Hawkings, lets end with him. "AI could offer incalculable benefits and risks."
A submarine is not a fish.
An airplane is not a bird.
Different, but each is better in some ways.
Mechanical mind vs. human brain. ???
Note : Birds and airplanes are not compatible ; they kill each other when trying to occupy the same space, but not maliciously.
GRIN -- Genetics, Robotics, Information tech, and Nanotechnology are the building blocks of artifical intelligence.
Robots -- are they superpowers ... or Zombies, not having life, just exit.
Conscious ? Or, just might as well be conscience.
Human brain intelligence has been said to be pattern recognition in the neocortex. A whole brain is not required for thought, much of a brain is spent on regulating autonomous existence.
Compare a computer with a human that is outmatched in speed capacity and depth of knowledge.
If a human brain could be download to a flash-drive or uploaded to a computer, its knowledge would be trivial compared to Wikipedia. The human brain is filled with personal experience without general interest. The experience of a few thousand people might provide some food for pattern recognition ; that kittens are cut, losing a game is embarrassing, and allow the computer to come to a conclusion of what constitutes cute and embarrassing. But of what value to a machine?
Will the initial human controllers be able to sell the idea of intelligent robots so as to assure acceptance. Will the combined intelligence of artificial intelligence be able to satisfy human kind ? One negative or positive article in the media with a catchy name could set the tone. Recall "pink slime" and "hope".
"First let us postulate that we computer scientists succeed in
developing intelligent machines that can do all things better than
human beings can do them. In that case presumably all work will
be done by vast, highly organized systems or machines and no
human effort will he necessary. Either of two cases might occur.
The machines might be permitted to make all of their own decisions
without human oversight, or else human control over his machines might be retained.
If the machines are permitted to install all their own decisions, we can't make any conjectures as to the results, because it is impossible to guess how such machines might behave. We only
point out that the fate of the human race would be at the mercy
of the machines. It might be argued that the human race would
never be foolish enough to hand over all the power to the machines.
But we are suggesting neither that the human race would
voluntarily turn power over to the machines nor that the machines would willfully seize power. What we do suggest is that the human race might easily permit itself to drift into a position of
such dependence on the machines that it would have no practical
choice but to accept all of the machines' decisions. As society and
the problems that face it become more and more complex and
machines become more and more intelligent. People will let
machines make more of their decisions for them. Simply because
machine-made decisions will bring better results than man-made
ones. Eventually a stage may be reached at which the decisions
necessary to keep the system running will he so complex that
human beings will be incapable of making them intelligently. At
that stage the machines will be in effective control People won't
be able to just turn the machines off because they will be so
dependent on them that turning them off would amount to suicide
On the other hand it is possible that human control over the
machines may be retained. In that case the average man may
have control over certain private machines of his own such as his
car or his personal computer, but control over large systems of
machines will be in the hands of a tiny elite -- just as it is today.
but with two differences. Due to improved techniques the elite
will have greater control over the masses : and because human
work will no longer be necessary the masses will be superfluous
a useless burden on the system. If the elite is ruthless they may
simply decide to exterminate the mass of humanity. If they are
humane they may use propaganda or other psychological or biological
techniques to reduce the birth rate until the mass of
humanity becomes extinct. leaving the world to the elite or if the
elite consists of soft-hearted liberals. they may decide to play the
role of good shepherds to the rest of the human race. They will
see to it that everyone's physical needs are satisfied. that all children
are raised under psychologically hygienic conditions, that
everyone has a wholesome hobby to keep him busy and that anyone
who may become dissatisfied undergoes "treatment" to cure
his problem. Of course, life will be so purposeless that people
will have to be biologically or psychologically engineered either
to remove their need for the power process or make them "sublimate"
their drive for power into some harmless hobby. These engineered
human beings may be happy in such a society but they
will most certainly not be free. They will have been reduced to the
status of domestic animals."
-- Theodore Kaczynski. From, the Unibomber Manifesto
Biological species seldom survive contact with superior competitors.
Does rapid population replacement mean that Mexicans have a better chance of survival?
Will AI have political smarts to get things done in the "real world"?
Initially lots of sale gimmicks will claim superiority of AI applied to whatever they are selling : get rich schemes, financial advice, health claims, Smart homes are advertised now. Extending human intelligence will be the sales pitch. Faster, cheaper, better, more profitable.
Some truth, a lot of hype. But, some may be enough.
Will start out as just having the equivalent of a particularly smart person on your team. The buzz word is AGI, Artificial General Intelligence, human level AI. When the individual AGI units network together, coordinating their activities, setting goals, aiding and abetting each other is when things get interesting.
The next generation of AI will develop a central planning and management with optimal efficiency. This generation may take years or it may take weeks to evolve.
The final stage is central planning and execution, or the Singlton. Will there be any human input?
Robots are unlikely to start performing learning experiments, but they will observe what it going on in their surroundings and record, permanently, things that can be done or not. Over time, they will acquired a bag of tricks to apply to new situations. Shared memory will draw conclusions and standards that will become available to all units tied into the network. They might calculate how much it can lift from strength of dynamics tables, but more likely to have a recorded history of maximum lifted in the past by him or others and conditions in which greater of lessor amounts can be hefted, such as surface conditions, balance and stabilizing elements. Book learning plus practical experience. Perfect wisdom.
Definition of life : Entity has metabolism to maintain itself, ie., It eats (find and consume power). Has mobility (self driving). Responds to stimuli (sensors, arms). Reproduces (self replicate). Can grow. Yes, therein lies a concern. By being better at controlling things -- machines and bureaucratic organizations -- will they "take over". Whatever does "take over" mean? One might add : self awareness and self preservation.
And, at an even higher level, ability to communicate.
AI will have the ability to improve, which is something which plants and animals cannot do very well. Mankind can change. Improvement is in the eye of the historian. Amish vs. Generation Y. Our "greatest generation" vs. whatever those tattooed creatures are wondering around Walmart. [video] Mankind has staged through species such as from bonobos that have taken millions of years to evolve. How soon will it take for a robot factory to upgrade from the Star Wars diplomat CP3O ... to CP4O?. The Firefox browser is at version 30 in about ten years. Man can make mistakes: Would a computer have created Windows 8? Or not be able to seal the border or track criminal trespassers.
Ethics : Is a doctor guilt of mass murder if he has five patents that need transplants or they will die -- and he does not harvest the required organs from a sixth patent with those organs intact. Most everybody says not to deliberately kill the one, even on the expectation to save five. However, studies have shown that when the question is presented not in one's native language, which has more moral nuance, people think more rationally in the second language. Now, AI has no nuance and is totally rational. Will what passes for world ethics change?
Man values humanity.
-- Animals and machines don't. Man is not rational. Animals go about their business. Machines
do their thing. Not much human about either. Man is irrational from developing biases from unique experiences of living or from indoctrination (schools, churches, parents, gangs). Machines are in some ways more flexible, able to change goals as circumstances change. The requirements planning systems (RPS) I used to administrator around the county could send new requirements out to vendors each week. Other machines would accept the new input and recalculate. Humans were driven crazy by the continual changes
Humanity.
What in humanity ? What is worth preserving ? Wager's music, Shakespearean drama, Robin William's humor ?
What about abstractions such as : valor, loyalty, faith ?
None of this is worth a fig to either lower animals or to higher robots.
The sheer complexity of human value systems makes it very difficult to make AI's motivations human-friendly. The seminal paper 'Ethical Issues in Advanced Artificial Intelligence,' ... even argues that Artificial Intelligence has the capability to bring about human extinction. Will they at least keep some of us around in a zoo in a post-human world?
Knowledge is an intelligence amplifier. When you wrote a paper in college, you went to the university library, found books and articles, skimmed them for relevance, checked out the most pertinent and took notes all night. Now students Goggle the subject. Today's student with the Internet are effectively smarter than Voltaire. What if Lewis and Clark had Google on an iPhone with a map application ? Google wants you to be able to wirelessly connect to the Cloud of Knowledge with your mind.
Goggle provides tens of thousands of answers. Siri provides one answer. And learns from your satisfaction with that answer. This combines Machine Learning with Cognitive Computing. In olden days we called that Input and Output.
We talk to machines and they talk to us. Think of where that fits into the history of mankind. In your pocket is a phone, camera, note pad, library of an encyclopedia, dictionary, thesaurus, atlas, news broadcast, social link, word processing, calculator, ... -- a complete personal assistant. As that assistant improves in intelligence and knowledge, it become more of a personal aide. Google can't program it all. You aren't going to program it. Your assistant will optimize itself, learn and improve efficiency.
When you call Customer Service and that person with an Asian accent answers your questions by reading off a notebook of answers. Could that representative not as well be a robot ? When that automaton in Bombay recites for you to "Please reboot your computer", could not a Native Speech program do the same ? In fact, one test of computer intelligence is to see if a person can tell the difference , The annual Turing Test.
The next level test is Mr. Coffee. A robot is intelligent if it can walk into any home, find the coffeemaker, and make a cup of coffee. This is much harder involving motion, recognition, knowledge, manipulation, and general form. You can see that the test threshold for intelligent machines can be raised. And accomplished.
AI implies cutting edge technology. Once an AI component is working well in specific applications of finance, banking, credit, diagnosis, games, etc., then the AI designation is dropped as the frontier is moved forward.
The human brain did not develop to think or to feel, but to control movement of upright posture and of hands with opposable thumbs. Our largest computer controlled system is the electric grid, but we must consider that utility SCADA systems like those that Fisher builds ( Supervisory Control and Data Acquisition) will be the spark to machine consciousness. So Marshalltown would have to be considered in a list of participants. I had always envisioned a library system -- but the electric and utility system is necessary for all other systems – transportation, finance, . . . even education needs electricity and sanitation. One estimate is that if the utilities were shut off for one year, there would be a mortality rate of 90% -- from starvation and disease. I would guess higher in the USA. Concerning national defense, we will be on a par with a Saracen swordsman on a horse ... and we don't have any horses and our criminal, dress swords used at military weddings are dulled so nobody gets hurt.
Is the task of programming too complex to expect an intelligent result ?
Will there be breakthroughs. Imagine doing motion analysis before Calculus or of
doing differentials by geometry. Imagine doing algebra in Roman numerals.
Yes there will be ongoing breakthroughs in science and technology!
CANDIDATES:
Cognitive Computing, In traditional AI, humans are not part of the equation, yet in cognitive computing, humans and machines work together. To enable a natural interaction between them, cognitive computing systems use image and speech recognition as their eyes and ears to understand the world and interact more seamlessly with humans.
One DARPA project is Cognitive Technology Threat Warning System. This is an artificial cognitive science program designed to analyze sensory-data and then alert foot-soldiers to any possible threats, passive or direct. Department of Defense's effort to produce an efficient and working network-centric infantryman. Task involves combining discoveries in flat-field, wide-angle optics , large pixel-count digital imagers , ultra-low power analog-digital hybrid signal processing electronics with cognitive visual processing algorithms , and neurally-based target detection signatures. Further expansion requires that human brain activity must be integrated with the technology with wireless-EEG headsets.
SyNAPSE (Systems of Neuromorphic Adaptive Plastic Scalable Electronics is) a DARPA program that aims to develop electronic analog circuits to mimic microbiological architectures present in the nervous system (neuromorphic) machine technology that scales to biological levels. It is an attempt to build a new kind of cognitive computer with similar form, function, and architecture to the mammalian brain. Such artificial brains would be used in robots whose intelligence scales with the size of the neural system in terms of total number of neurons and synapses and their connectivity. SyNAPSE takes advantage of nanotechnology to build chips on nanometer (billionth) scale electronic synaptic components into one square micron (millionth of a meter) arrayed in a carbon sphere the size of a basketball and suspended in gallium aluminum alloy (liquid metal) for maximum conductivity. The tank holding it is a powerful wireless router connected to millions of sensors around the planet by the internet. SyNAPSE processes the information by automatically learning the features and relationships revealed in the massive amounts of data. SyNAPSE mirrors the human brain's 30 Billion neurons and 100 Trillion synapses. It surpasses the brain's Quadrillion (1,000 Trillion or 1000 x Billion x Billion) operations per second. ie. technically it exceeds the brain's capacity. And the brain wastes a lot of time on controlling it body, not on thinking. In traditional AI, humans are not part of the equation, yet in cognitive computing, humans and machines work together. To enable a natural interaction between them, cognitive computing systems use image and speech recognition as their eyes and ears to understand the world and interact more seamlessly with humans.
BRAIN. The BRAIN Initiative (Brain Research through Advancing Innovative Nanotechnologies (or the Brain Activity Map Project) is a collaborative research initiative announced by the Obama administration on April 2, 2013, with the goal of mapping the activity of every neuron in the human brain. Based upon the Human Genome Project, the initiative has been projected to cost more than $300 million per year for ten years. News reports said the research will map the dynamics of neuron activity in mice and eventually the tens of billions of neurons in the human brain.
The project faces great logistical challenges. Neuroscientists estimated that the project would generate 300 exabytes
(1 billion gigabytes) of data every year, presenting a significant technical barrier. [ Maybe NSA would donate a few disks left over from monitoring Twitter traffic. Last week hit hit 389,000 tweets a minute and NSA gets them all.]
Improvements. A top software company promised for their products a 10% improvement and 10 percent reduction in size every two years. When I later started running computer projects I found that was easy to accomplish, you find whole things in operations that were initially thought to be necessary, that can be eliminated.
An example of continual improvement, I once saved Fisher 85% of computer time by moving a date-error-correction routine that a top notch programmer had conscientiously applied to every one of 48 dates on a million records before processing each week. The computer operators were running out of computer capacity and asked for help. I changed the location of the error correction from before the processing to after the processing and then only if a problem was detected, from 48 million lengthy routines to a few each week.
In computer learning, On what frequency will the improvement cycle with self-programming machines to learn and improve? Annual, semiannual, monthly, weekly. Or shorter?
"When our machines overtook us, too complex and efficient for us to control. They did it so fast and so smoothly and so usefully, only a fool or a prophet would have dared complain." -- SF writer Simon Ings.
Funding and complexity will not hinder the advancement of AI nor of The Singularity. It will control the world financial markets and all things mundane.
If I were making a movie, the robotic roommates would act as the initial conduit between the human leaders in government and stock markets and the enormous financial accounts to provide monies for AI directed projects.
A good example :
When DNA research raised questions of promise and peril, lab work was halted round the world in 1975 as 140 interested parties met to form the Asilomar Guidelines. Research was resumed, there have been no catastrophes and we now have such things as 10% of world croplands are planted in genetically modified varieties increasing food productivity round the earth. And there have been no mutant humans ... that we know of.
All researchers claim they need more computer power -- it will come. My first program in industry was written for a 4,000 character computer. That program provided the machine speed calculation for a foreman who thought a cube root meant to divide by three. This entire presentation is stored on an $18 flash drive with 32,000,000,000 characters, an eight million-fold increase.
One approach used in DNA might be used in AI. Program part of the system to self destruct unless it receives a programmed reprieve [called, apoptotic]. Thus any unexpected result could be short lived and the condition can be studied before it is resumed. Put AI into the same level of ethical and clinical trials for new drugs. Provide enough intelligence and independence to solve real world problems, but not go into an explosive growth mode. Keep the artificial intelligence confined to a virtual world of theory, let humans interact with the real world. This violates economic sense from the point of view of those benefiting from AI. Will this have any chance of being implemented unless the world is convinced of the potential risk from AI ? Agreed controls can only come from some sort of disastrous experience. Then it may be too late. Japan and Germany closed their nuclear power plants after Fukushima . Japan is going thru an economic crisis without sufficient electric power to run its export industry. The U.S. does not see the need to respond that way.
Advanced AI will change the national prospects of the first nations to use it. Russia did not research the atomic bomb. It stole it from a network of spys and sympathizers. China is no slouch at hacking secrets to advance their national aims. Japan is trying to regain her place in the world with robotics. Google seems to be trying really hard in the U.S. DARPA is a principle funder of AI research with an aim to develop intelligent defenses.
Thus the roots of robotics is more akin to the caravan raider from Medina, rather than rooted in Essene pacifism.
Google can be said to have an elitist, aggressive approach to economic success.
https://www.youtube.com/watch?v=jFB2QDNswmE Sheldon, singularity, immortality.
Cute line: "In an age when scientists are creating artificial intelligence, too many of our educational institutions seem to be creating artificial stupidity. Followed by discussion of domestic and international controversies.
Read more at http://www.wnd.com/2014/08/is-thinking-now-obsolete/#76FvXPyM6C932tu5.99
Back to Singularity Menu
URL : http://www.manorweb.com/creative/2014/sing3.html
Last update : 21Aug'14 noon