The Singularity
II . The Near Future 2029-2039
D R A F T

II. THE NEAR FUTURE 5 to 25 years , 2020 - 2039

We can now combine the topics of AI and robotics.

'Watson,' the natural language question answering system developed by IBM that wins at Jeopardy! On January 9, 2014 IBM formed the Watson Group with 2,000 people for $1 billion. The cloud-delivered enterprise-ready Watson has seen its speed increase 24 times over and its physical size shrank by 90 percent; it is now the size of three stacked pizza boxes. They expect Watson to generate $10 billion in annual revenue.

BLACK BOX. Imagine you are in a room behind a black door. You have a list of words, a dictionary, say in Serbian, and a computer. Somebody submits a question in Serbian thru a slot in your black door. You translate and output the question in English. You go Online and seek an answer. Using a word processor, you cut and paste and format an answer. You did not and still don't know Serbian. The person with the Serbian questions does not know, and doesn't need to know what goes on behind the black door. This is a black box, it takes input and generates outputs.

Will there ever be a black box that you sit down and it finds every defect in your house and repairs it? Possibly. Will there ever be a computer that can respond to the thoughts of Steven Hawking about the origin of time? Yes, with this observation. Many PhDs have lots of knowledge of things they have read. That does not mean any of these educated monkeys can have an original thought. How important is originality? I have little originality, but have done a lot of reading that can be applied to new or unusual situations that might be called intelligence. Is there anything new under the sun (or behind it) or just gradually expanding knowledge of facts? Facts are something that a machine can store, search, and retrieve better than a human. I am not belittling machine intelligence ; I am belittling what passes for human intelligence.

Novel situations might exceed any capabilities. Hence your stock broker's warning that "Past performance does guarantee of future performance." Same with forecasting from past data. I used to be a expert in forecasting and always built in a human override for the chance of novel situations - new products, international events. Yet, if the capabilities to draw up meaningful past experiences to apply towards novel situations is called intelligence, then machines will surpass humans.

Some Singlularitians want to get to Advanced General Intelligence as soon a possible owning to its potential to alleviate human suffering. Others hope AGI takes a long time because the likelihood of destroying ourselves may diminish with time due to more, better research. [Some feel the same way about GMOs, cloning, and daughters dating.]

The idea of AI has been around since 1956, the key has not yet been found, ... yet.

Human Mind.
200MM years of evolution have honed the human mind.
Yet it is composed of components selected for specific tasks.
And it is certainly not optimized.
The is a branch of AI attempting to reproduce the brain.
• 1 . Duplicating the brain functions -- how does the brain do it, neurons and chemicals.
  2 . Duplicating the functions of the brain -- what it does, sight, speech -- with other materials.

Dartmouth student won a contest to recognize an object by an algorithm 140 times faster than the existing. Older AI publications praise the brain's ability to recognize faces, not just objects by shape, size, color. But, now you see facial recognition on every TV detective show.

Watson.
IBM's Watson AI program won at Jeopardy. This means answering questions about just about everything, before a human competitor can, but also pickup clues, puns, similes, cultural references. IBM is re-tasking Watson to do medical diagnostics. Self learning by reading 8 million books. It learns like a child, through multiple references that differentiate a" roof" from a "basement". When asked if Watson is capable of thinking, the team leader asked, can a submarine swim?

AI will be necessary to protect the nation's cyber base : power, finances, communication, entertainment, et al.

MACHINE LEARNING. recommendations for you to buy -?-
Try to understand the black box by rational economic theory -- make decisions to be optimize its goals. But man is not rational, makes decisions on such things ad the weather, last meal, what learned at his mother's knee.
    What are robot goals? What are those of a self-improving machine?
1 . Efficiency - to better accomplish its goal -- win at chess, make more money by stock market trades, ?
2 . Self preservation -- needs energy, data, will duplicate itself, immortal so has long time span of view. Take action (preventive action?) to remove threats and to survive.
3 . Resource acquisition -- budget, build space ships to mine an asteroid for a mineral to build better circuits for winning at chess.

NELL (Never-Ending Language Learning system) is a semantic (meanings), machine learning system developed by at Carnegie Mellon University, and supported by grants from DARPA, Google, and the NSF, with portions of the system running on a supercomputing cluster provided by Yahoo!
  NELL was programmed to identify a basic set of fundamental semantic relationships between a few hundred predefined categories of data and now runs 24-7 to read the internet, to make new connections in a manner that is intended to mimic the way humans learn.

We are already controlled by machines : stop lights, monthly billing, One estimate is that 90% of humans would be dead within a year without electricity.

Do you watch the TV show "Person of Interest"? That started out as fiction, then we found that it is commonplace NSA activity. The self-improving, independence aspect may just be something that has not leaked to the press yet. Where is the next Snowden?

When I subscribed to "Futurist" magazine back in the 1960s, this was next century stuff. No more : We are in the next century with only 20-30 years to go. Twenty-six years ago was the end of the Reagan administration and the internet only connected research centers and had not yet gone public. Fifty years ago, at the NY World Fair, IBM was pushing Selectric typewriters, not computers.

It has been suggested that hacker software will provide the spark for salience. The software is advanced, uncontrolled, must operate independently while the human programmer goes off to do his homework. There are more malware programs than legitimate ones. We don't need to ask the rhetoric question if anyone will be greedy enough or malicious enough, or even loony enough to intentionally create chaos.

Funding and complexity will not hinder the singularity or advanced AI or progression to super-intelligence.

Then it might be with the full resources of the Chinese government. From 2007 to 2009 the U.S. government detected 47,000 cyber attacks. Note: detected, no idea how many were not detected. Dept of Defense now acknowledges warfare as land, sea, air, space, and cyber.

Energy grid maybe the largest system, but I doubt it is a candidate for initial spark to consciousness. I envision a data system, but electric systems is necessary for all other systems -- transportation, financial, even education needs electricity. It has been estimated 90% mortality if electric is off for a year.

Concerning national defense, without technology, we will be on a par with swordsmen on a horse. And we don't have horses any more and ceremonial swords used at weddings are dulled so nobody gets hurt.

Biologically based memory is currently expensive and used only on supercomputers, but offers advantages of speed without heat and memory when it is packed more densely. Is there a point when AI will have living components? Is a machine cyborg any different from a human cyborg : machine with human parts or a man with mechanical parts.
Cyborg.

Debates about AI technology have tended to focus less and less on possibility and more on desirability, to actually seek to build more intelligent successors to the human species -- selection, genetic modified, chemical additives, computer chips, ... Transhumanism is the interest in human enhancement technologies for intellectual, physical, and psychological capacities to become more than human. A common theme in fiction is the super-soldier.

Government without waste and corruption would be a gigantic boon : no self interest, bribes, favors, corruption or fraud. Reduce taxes to zero, or even provide a negative tax (get money back). Can government reinvent itself ? Will AI become an equivalent civil service system?

AI will give the Muslims something more worthy to fight about. The word of Allah against the godless Western robots.

Couple the technology of 3D printing with chemical fabrication of foods and you have the robotic cook. Speak your desired meal and it is presented fresh and warm from the printing microwave.

Print on demand cars are available today. The Strati is produced from fiber-reinforced thermoplastic strong enough for use in an automotive application. (remember the 1960's Corvette was Fiberglas.) The chassis and body (which goes on a drive train, wheels) weighs a scant 450 pounds and the completed car is comprised of just 40 components, a number that gets smaller with every revision. Fewer parts can mean fewer problems.

II. B. HUMAN ENHANCEMENT

Either natural or artificial, the subject can range from cosmetic eye lifts to letting the blind see.

GENE THERAPY
    The use of DNA as a drug is successful with several specific illnesses. Generally entails introducing a virus. Currently tends to be short lived an requires recurring treatment. Research seems to be concentrated on popular diseases such as AIDS.

GMO, Genetically Modified Organisms. Has had great success, but seems to be limited to plant life amidst great controversy, that seems to me to be based on economics rather than fear of catastrophic delayed reaction.

The Health department in England is allowing a fix for defective genes that introduces clean DNA for invitro fertilization. This created a controversy about three-parent and genetically modified children. Will this be a horror or god-sent?.

EXOCORTEX is the interesting work of augmenting the mind. State of the art is still mapping the brain with limited real life applications. Neural implants seem limited to treatment of Parkinson's disease. Leading edge work in artificial eyes and cochlear implants for hearing. An article in "Popular Science" years ago allowed 8x8 pixel vision which means the blind can see the light and dark of large objects and motion. Work is going on round the world with labs in 2014 planning to use 1040 to 1500 electrodes. Does this mean 32x32 pixels ? CGA (old TV) was 320x200 ; acceptable VGA 640x480 ; HiDef 1920x1080. All are steps in progress.

WEARABLE COMPUTERS. Trendy. Bluetooth everything. Worn over, in or under clothes. Long history. 1600s China saw a fully functional abacus on a ring. Progress thru the wrist watch to smart phone.

The "Future Combat Systems" project is to create a lightweight, fully integrated infantryman. Situational awareness with helmet-mounted 360° display, providing maps, routes, and wireless communications. Health monitoring ; heat control ; fuel cell and backup solar power.

MACHINE TO MACHINE, the so called INTERNET of THINGS or device to device networks. Everything from your toaster to home security system that respond to specific or general instructions from you. These commercial applications will rapidly adopt and move forward with artificial intelligence. Because every device will have a unique ID number, the size of internet address (IP) had to be expanded from four blocks of four numbers to a new standard which consists of 16 blocks of numbers and letters.

MIND to MACHINE Interface.
Three levels.
1 . Brain implants to control prosthetic limbs. Good work in getting the spinal cord injured to walk. You can imagine the expense of this research. But, further imagine this combined with robotic exoskeleton augmentation.
2 . Skull implants are promising with good signal quality and no scar tissue interference.
3 . Non-invasive (electrodes caps) Poor signal, less precise source of specific neurons. We see the host of science shows moving a cursor across a screen and even running a model car across the room.

Headline: "Man in motorized wheelchair robs store...". Is all progress good?

MIND to COMPUTER.
This seems more fruitful to me. Let one's mind control a computer and let computer control sensors and activators. Mind activity may need more work, but computer vision (input) and motion (output) are well developed. Is there any promise ? A teenage epileptic was able to play Space Invaders with his mind while waiting for some brain tests. I would say, Yes.

MIND TO MIND.   Army is interested in silent communication. Current level is sending and receiving Morse code. Code to words is easy next step. Imagine privacy issues as this field of mind to mind develops.

OUTLIER.
    Neurons from a rat were "cultured in a petri dish and rapidly began to reconnect themselves to form a living neural network. The cells were arranged over a grid of 60 electrodes and used to fly an aircraft simulator. "The study's focus was on understanding how the human brain performs and learns computational tasks at a cellular level." This strikes me as huge!
http://www.wireheading.com/misc/artificial-brain.html
http://neural.bme.ufl.edu/page12/page1/page1.html

DARPA has a project to develop brain implants to resort memory. 270,000 U.S. veterans have been diagnosed with traumatic brain injury. With the advent of body armor, GIs that used to die are left with wounds to the head and limbs. Both are subjects of Human Augmentation. Progress of sorts.

WHY IS ARAPA so successful? They have specific goals and don't have to be concerned with outside pressures such as outcries by people who like to be outraged by social concerns.

Golden Age: "People should not have to work so hard," The founders of Google said, they want two part time workers hired instead of one full time. [Do you see any problems with that?]

RoboMotion and appearance.
In addition to the robots with a mechanical look, android-type communication robots with highly human-like looks have been developed. A merit of android robots is that they give people a feeling of human-like presence as if people were interacting with a real human, not a strange looking machine. Robotic pets are on the market.

Researchers are working on improving the technology and expects these robots will infiltrate our daily lives in roles such as building security, house cleaning, and greeting customers at shopping centers. Humanoid robot is expected to have behavior like a human as customers can suppose from its appearance.

SEX With ROBOTS -- Fembots or Malebots
    The beginning of the 21th century was marked by the creation of the world's first realistic female robots. Japan and Korea, began production of anthropomorphic robots. It is impossible to say exactly which of them has officially made the first "realistic female robot", because the scientific and laboratory studies were conducted simultaneously in both countries.

A genius inventor named Le Trung who has never had a real girl-friend has successfully created his own perfect girlfriend, a robot. The fembot, Aiko, can do the cleaning, mix Trung's favorite drink and read him newspaper headlines. The fembot acts like any other woman will. If you grab her in a certain way she might react differently. Aiko may be even slap you if you squeeze her body parts. Aiko actually means "love child" in Japanese but the software engineer denies he has created the robot for sex. Yea, sure. ;^) Better yet. Aiko has been also programmed to simulate an orgasm. Creating Aiko has cost Le Trung around $21,000, with much of the money coming from credit cards and the sale of his car. Trung is currently working to improve his "girlfriend" in other ways.
http://www.sodahead.com/united-states/fembot-would-you-have-sex-with-a-robot/question-207300/?page=3

Realistic female dolls have been are the market for years, cost has come down from $7,000 to $3,000. The robotic models are listing for $100,000 to start. Like personal computers that price will come down, level off, while the performance will grow formidably.

E. FUTURE WAR

Robots and computer will make independent judgments. The current setup is too slow. The development of reconnaissance drones was initiated as a billion dollar investment to find Osama Bin Laden. It worked, he was detected twice. But the rules of engagement required that a higher authority including a legal team to review the evidence before authorizing a strike. It took time to setup for targeting a missile or launching an air strike. Even if there are bombers in the air, it takes travel and targeting time. The combined human confirmation process and response time of distance allowed him to get away each time. The next natural step was to arm the drones, to improve the response time. But manual authorization is still required by people concerned with the presence of non-combatants, specially children, around the target. The next logical step is to allow the drone to make that determination to strike. The same holds true with every weapons system : tanks, artillery, infantry weapons. This includes automating out of the human driver and combat soldier. Makes war a lot safer. Yes, no, maybe?

I am reminded of an original Star Trek episode in which computers alone fought wars according to rules that did not require destruction of property on either side . Only of people . The computers determined the successfulness of each attack, the number of casualties, and the population was conditioned for the proper number of thousands of persons to rush to their local annihilation centers for termination.
Computers can make war very efficient.