EdEarl

What's scary is the FBI, CIA and NSA having AI to monitor people using connected cameras, WIFI emissions to monitor people inside buildings, phone calls, other spy tech, and internet data. Soon AI will be able to track many people from cradle to grave. Perhaps Google has the best AI, but governments will have very good AI.

Google's AlphaGo (AlphaX as it has learned many games in addition to Go) is an example of a primitive level of cognitive autonomy. They let it learn games like space invaders with its programmed instinct being to read the game score and try to maximize it. It did not know how to play any of the games it learned. The set it playing and during the first hundred or so games, it lost badly. After 500 or so games it knew how to play and win. See: https://www.youtube.com/watch?v=Qvco7ufsX_0. It even developed strategies that none of the developers had ever seen or imagined.

We program individual features of a neural net that do not change. However, neural nets learn and change their own functionality as they learn, and sentience may emerge as neural nets learn. That does not mean we know how to write a sentient program that cannot learn.

A little bit of graphene goes a long way, about 2630m2/g.

My childhood associations were mainly conservative southern baptist; thus, my sex education was pathetic. They preach abstinence and taboo; basically they want children to be ashamed to think or talk about sex until they become legal adults. Then, they expect everyone will have a healthy attitude about sex, which is absurd. This kind of sex education is so bad I find it difficult to believe watching porn can do more harm; although, both are perverted. Most porn (IMO 99+%) is pure fantasy. Men crave novelty in porn, and the porn industry creates fantasy situations to satisfy that need for novelty. I think people will survive more or less OK, in spite of bizarre sex education. I think realistic education for adolescents would make the transition from sexually naive to sexually active easier, I think some people suffer poor sex attitudes for years because of poor sex education, but IDK how to improve the situation.

To program sentience, we have to understand it mathematically, but we don't. People agree we are sentient, but we cannot precisely define what it is. If it emerges from large neural nets, then we don't have to know what it is, we only need to build large neural nets.

I think it will take much less than 100 years; although, there will be arguments about whether AI is sentient for a very long time. We have no consensus on whether great apes, dolphins or elephants are sentient, and that discussion has been on-going for a long time. Just recently psychologists have discovered that some birds name their young, and that dolphins, elephants, great apes, and many other animals communicate among themselves. I think sentience is not on-off; rather, there are degrees of sentience. Clearly a person who is brain-dead is not sentient. Moreover, when people sleep they are not sentient; although, we might consider someone in REM sleep who is dreaming to have a small degree of sentence. Brain damage can make a person into a vegetable, who we would not attribute sentence, or leave them severely disabled, but with sentence. However, brain damage might be so severe that we could not assess whether a person is sentient, for example if a person could not talk, hear, see or move. A robot might be able to do anything that humans can do, even better than we can, yet we may not credit them with sentience, because they are simulating a human. Just as some people believe astronauts did not go to the moon, there will be people who do not accept sentient AI. In fact, I think it doesn't matter, because I think there is nothing special to do to give AI sentience; it will occur or not.

Rather than saying what consciousness is not, or guessing what it is; I'd like to discuss characteristics of consciousness. We are conscious of some external stimulus, such as visual images, odors, sounds, etc., and internal feelings caused by neurochemicals, such as adrenalin, oxytocin, and serotonin. In addition, we are conscious of our thought process or train of consciousness. Moreover, our brain filters out both internal and external sensations. For example, we can focus on a single sound (e.g., someones voice) in a chorus or cacophony. On the other hand, our brain does things subconsciously that do not affect our conscious thought process, for example moving our hand away from a fire that burns. We are aware after the the event, but we move about the same time our conscious realizes we have been burned. When we are first learning to read, we focus on each letter or each phoneme. With practice we focus on words and phrases, without sounding out each word. This process trains us, including the conscious part of our brain. To learn the letters we train hundreds of thousands of neurons to recognize the various letters of the alphabet. We start by looking at examples of each letter and try our hand at writing each one. Our writing is not exactly the same as our example, which means our brain learns a dozens or hundreds of variants of each letter. Much of how we learn is subconscious, but our conscious brain controls some things we learn. Eventually, as we read other people's writing, we will see thousands and thousands of variants of each letter and train millions of neurons (a neural net) to recognize them. Ray Kurzweil says there are roughly a hundred thousand neurons that recognize the horizontal bar connecting the two sides of the letter A. There will be a similar number for each side of the A, and for the A as a whole. We might write an A without closing the two lines at the top, in which case the brain must decide whether the letter is an A or and H. It uses additional clues to distinguish between the two, for example by testing whether the letter makes sense within the word currently being read. For example, if the word is Aaron, an H doesn't make sense (Haron) so inhibiting signals will be sent from some neurons to tell the H recognition neural net that the letter is not an H. Our brain is organized hierarchically. Neurons in the eye process individual "pixels" to recognize simple line segments and movements. They send outputs to higher level neurons that recognize more complex images until a single neuron when stimulated with a tiny electric shock can cause a person to recall a complex image, such as a person. Similarly, the inhibiting signals filter a chorus or cacophony to let us hear a specific sound; from this process emerges what we call consciousness, that allows us to make better decisions, albeit a bit slower, than a simple stimulus-response. It is hard to understand and define, because it can use all sensations and memories to make decisions; it requires complex neural circuits.

You are right. It is the graphene enhanced LED that is 10% more efficient with a longer life span. http://www.geek.com/chips/graphene-enhanced-led-light-bulb-is-cheaper-more-efficient-longer-lasting-1619454/

You got that right. I expect to see better batteries and supercapacitors that will change transportation from petrol to electric. Perhaps graphene can be used in 3D printed motors; if so the weight and cost of motors should fall as graphene prices fall. It is a great armor. https://www.newscientist.com/article/dn26626-bulletproof-graphene-makes-ultra-strong-body-armour/ It makes a desalinization membrane. http://cleantechnica.com/2015/04/03/methane-rescue-new-energy-efficient-graphene-desalination-membrane-99/ Graphene light bulbs are more efficient than LED and last longer. http://physicsworld.com/cws/article/news/2015/jun/18/graphene-light-bulb-shines-bright And many more.

First produced in 2004, graphine production per year is now 205 tons and growing. See also: https://graphene-supermarket.com/new-products http://www.independent.co.uk/news/science/first-graphene-light-bulbs-to-go-on-sale-this-year-10142026.html

Each video is followed by problems you solve to show proficiency. AFAIK you cannot jump to the exercises. Although, I just entered and it gave me exercises to find out my level of math.

If there is one, it must be the programmer who created this simulation we call the Universe.

Lots of practice at kahnacademy.

Have you looked at https://www.khanacademy.org/; it has math and physics lessons for beginners through some university level lessons.

It might be possible to have two wheels be coasters, and only one steerable and powered; in which case, a cart would have only three motors, which is less expensive and less maintenance. A fixed mirror would require two motors I think.

A picture in that article shows Nevada's Crescent Dunes concentrating solar system that generates 110 Megawatts and stores heat for overnight generation. It appears the mirrored area has been meticulously prepared; moreover, there might be a design that minimizes money spent for site preparation. It requires a swarm of mirrors on a cart with three independently steered wheels. The cart can point the mirror in any direction around the compass; thus, the mirror must position angular elevation. The site must be cleared of trees and brush, should be relatively flat and mowed. Parts for the carts may be 3D printed and assembled by robot. The mirrors may be covered in graphine or gold to minimize deterioration. Electronics must include sensing neighbors, position of target, and a computer, which are common on drones and robots. Each member of the swarm can drive itself to a cleaning station periodically, and to an inspection station for a periodic check up. A maintenance drone might be able to replace a defective wheel without moving the broken mirror cart from service; such service is possible but it may be more practical to haul the defective unit into a maintenance facility. Fixed solar requires access paths between PV or mirrors. A swarm does not; thus, it is possible to save about 30% on land. If one of the swarm is broken, the others can move to make an access path. Solar PV requires between 4 and 6 acres for 1MW. Solar thermal would require about 0.25 acres (10,000 sq ft = 0.1 hectare) of mirror according to http://www.solarmango.com/scp/area-required-for-solar-pv-power-plants/. If each mirror is 10sq ft (1 m 2), the swarm needs 1000 carts. If each cart cost $250, that's $250,000. An acre in my area can cost $250,000. I think this design should be considered by solar thermal engineers.

A water molecule is about 0.275 nm. A neuron is about 100 nm, and an axon is about 20 nm diameter.

I believe people have similar feelings because their brains are similar. I believe making an AI with similar feelings would be much more difficult than making a sentient AI. However, a sentient AI may be able to study human physiology and physiology; armed with this information and personal observation, an AI may develop a sense of how human emotions affect us. Subsequently, the AI may be able to mimic human-like emotions, for example anger and love. It might have a sense that loosing power would either kill it, but would it have a fear of death? Would it do anything to prevent being turned off, would it say, "Please don't turn me off," or would it just not care?

Can an AI really feel? Is it possible an AI understands our feelings, and reacts as if it had feelings but doesn't really? What is the difference between simulated feelings and real feelings?

Within the next 5-10 years anyone will be able to consult an AI with their phone. The AI, with access to the internet, will be a polymath, an expert at many things. Will people consult with AI to make political decisions? If someone can pervert an AI for their own purpose, will such efforts be effective propaganda? Note that OpenAI is a project to insure unperverted AI is available to everyone.