What would the creation of greater-than-human artificial intelligence mean to us? What would be the impact of nanobots (molecular sized robots) rebuilding the world one molecule at a time? How will uploading (moving your mind into a computer), physical immortality, or freedom from labor change the way you live? These are among the rapidly advancing technological capabilities that will be here sooner than you think.
The universe functions on three levels of organization.
Singularity - the explosion of nanotechnology, biotechnology and computer technology.
Molecular Nanotechnology - the complete control over the structure of matter at the atomic level.
Molecular Biotechnology - functional genomics, proteomics, cell biology.
AI - Artificial Intelligence, human level knowledge, reasoning, and cognition in a computer.
SAI - super human level intelligence in a computer.
The universe functions on three levels of organization. The first level is "physical law" as described by physics and chemistry. The second level is "life" comprising biological functions, DNA, molecular biology, and molecular biotechnology. The third level is "mind" displaying the characteristics of thought, consciousness, intelligence, and information processing.
Many useful and interesting things in the universe are made of atoms, such as, stars and planets, spaceships and airplanes, golf clubs and golf shoes, ham and eggs, your dog and your wife. The behavior of these atoms is the science of physics and chemistry. Question: What is the difference between your wife and your best set of golf clubs? Answer: The type and arraignment of the atoms. A technology that could precisely control the type and arraignment of the atoms of any "thing" would be a very handy technology. Molecular nanotechnology is just such a technology.
Suppose you have such a technology. And you are bragging to your neighbor across the hedge that you can make absolutely anything. After a couple of simple challenges like hedge trimmers and sunglasses he comes up with a real doozy. He challenges you to make a live dragon that flies, breaths fire, and eats only virgins. You are stumped. In order to make a completely new living organism, or repair, or redesign an existing living organism, such as yourself, you need a type of knowledge that wasn't included in your molecular nanotechnology user's guide.
You need to know how living organisms work from the molecules up. You need a knowledge base of DNA, proteomics, molecular biochemistry, cell biology, and systems biology. And you need a toolbox called molecular biotechnology. You also need to know if your neighborhood is zoned for fire breathing dragons. Molecular biotechnology comprises some very complex scientific challenges. But you are in luck. All the tools required for the development of molecular biotechnology are included in the basic capabilities of molecular nanotechnology.
Meanwhile, some people sitting on the local zoning board are saying they need some kind of automated system for intelligently monitoring the activities in the neighborhood to prevent violations of the zoning ordinances, such as the production of fire breathing dragons. The computer at the heart of this system would have to have a broad understanding of neighborhood life, the ability to exercise judgment, and predict consequences of actions. It would have to be a mind, hence artificial intelligence.
There are three different approaches to the development of AI. You could evolve it in an artificial environment (cyberspace). Or you could sit down and figure out what information processing algorithms are needed for perception, learning, and planning. Or you could use the tools provided by molecular nanotechnology to reverse engineer the human brain. And while we're at it, let's fix some of the obvious shortcomings of our mind's design.
Among the fixes would be increased memory capacity and accuracy, increased speed and accuracy of cognitive processing, the ability to perform long series of mathematical or logical operations flawlessly, the ability to implement new information processing algorithms, and modify the content and structure of existing mental processes.
We would want to run this reverse engineered, supped up, and re-engineered mind on hardware a bit more reliable in its operation than biological neurons, such as electronic molecular nanotube computer circuitry, which is made possible by molecular nanotechnology.
Ever since that first cave man in our past chipped a stone into the shape of a spear point we have been occupied with creating better tools. Molecular nanotechnology, molecular biotechnology, and artificial intelligence are the ultimate tools. Which makes us the ultimate tool creators. And if you are thinking that all these new technologies sound like fun but surely we will not have them for hundreds of years you really should read the next section, "What is the Singularity?"
WHAT IS THE SINGULARITY?
The Singularity is an unprecedented event in human history when life as we know it changes from "business as usual" to a condition of virtually unlimited capability.
The Singularity event is triggered by exponentially advancing technologies. Almost no one has much experience with exponential technological advancement. It is not part of our usual environment. We are evolutionarily and culturally programmed to expect imperceptible rates of change. We don't have the neurological cognitive structures to model exponential rates of change.
Witness the Human Genome Project. The first decade progress was so slow many experts predicted it would take 200 years to complete. Over 50 percent of the raw data was generated in the last nine months of a fifteen-year project.
When advancement continuously doubles over consecutive equal time periods that is first order exponential advancement. When the doubling time period shrinks over time that is second order exponential advancement.
We have all seen computers getting more powerful while the price simultaneously drops. The doubling time for computer power per dollar was 3 years in the 1950's, 2 years in the 1980's, 1 year in the 90's, and 9 months as of September 2002 and still dropping.
The reason for exponential technological advancement is that each incremental advancement in tool design results in increased productivity of future tool design advancement. It's a positive feedback loop of design tool productivity. Most of these feedback loops are based on continuing miniaturization of mechanical/electrical systems, the endpoint of which is molecular nanotechnology.
Exponentially advancing technologies will change everything about the way we live life and do it in an incredibly short period of time. The Singularity will result in almost unlimited technological capability. People who are currently competing for scarce resources will have incredible wealth and freedom.
SINGULARITY INCOMPREHENSIBILITY
Life after the Singularity will be quite different than at any time before, but it will not be incomprehensible. We can make a fairly complete list of the coming technological capabilities. The social consequences of these capabilities are up to us. It is necessary that a large portion of humanity put some thought into what kind of social structures we want, if only to avoid panic.
Recursively self-improving AI's will develop intelligence quantitatively beyond current human level intelligence. These quantitative improvements will include increased capacity, accuracy, and speed of data processing. Quantitative improvements are merely extensions of our current cognitive capabilities.
Qualitative improvements, or whole new ways of comprehending and thinking that are beyond our current understanding are not likely. The reason for not expecting a new kind of intelligence but only a new level of intelligence is that intelligence is about solving problems through algorithmic implementation. And based on the success that we have had in the physical, biological and informational systems we can surmise that we have already explored the broad outlines of algorithmic effective solution space. Of course I can't argue with the statement that we don't know what we don't know, but I think we can put what we don't know into successively smaller boxes.
Those who think that recursively self-improving AI is the whole or even the primary aspect of the Singularity are not seeing the whole elephant, just the trunk.
THE THREE-TECHNOLOGY SINGULARITY
The Singularity is the most significant event in the history of mankind. It consists of three simultaneous technological advancements, artificial intelligence, molecular nanotechnology, and molecular biotechnology. The rate of advancement starts out very gradually but due to a feedback mechanism gets faster the closer we get to the Singularity. Upon reaching the Singularity mankind's capabilities become truly awesome; complete control over the structure of matter at the atomic level, complete knowledge of biological processes from the macroscopic to the molecular levels, and super human artificial intelligences.
Each of these three technologies reaches a physical limit at the Singularity. The disruption of society and our way of life will be total. The Singularity will herald the beginning of a new age in which physical and mental perfection will be the norm along with unimaginable wealth and freedom. These benefits will accrue not just to the current privileged classes but also to every human being.
WHEN WILL THE SINGULARITY OCCUR?
One of the fundamental open questions about the Singularity is when will it happen and how fast will the technology advancement get. Projected trends indicate the Singularity should occur by 2020 (this may need to be revised to 2018 or 2016 based on Dani Eder's most recent estimate that the doubling time for computer power has dropped to 9 months as of September of 2002) and should be quite sudden, in the nature of an exponential curve.
How will all these changes be implemented? Eliezer Yudkowsky thinks AI will mature before the other two technologies, and that a superhuman intelligent AI will take us through the transition. He calls it the "transition guide". Eric Drexler and Ralph Merkle think the greatest potential lies with molecular nanotechnology. Ray Kurzweil and Hans Moravec are expecting a synthesis of computers and humans to lead the way.
The three technologies will likely develop in parallel due to their interactions. Mature nanotechnology (the complete control of the structure of matter at the atomic level) will allow us to rapidly produce atomically perfect products at the cost of the raw materials in unlimited quantities. It will allow us to produce materials fifty times stronger than steel or with whatever designer qualities we specify along with complete access and control of the biochemical processes of living organisms. And will give us the tools to rapidly reach maturity in molecular biotechnology and artificial intelligence by reverse engineering the biological organism and the functioning of the human brain. The effects on society cannot be overestimated, even though they have been overestimated in the past. The reason they were overestimated in the past is the difference between Singularity technologies in development and mature Singularity technologies.
The most likely scenario for the Singularity is that at about 2011 industrial nanotechnology will produce an avalanche of inexpensive goods eliminating poverty and hunger worldwide. At about 2012 computer technology based on nanotechnology will have advanced by six orders of magnitude from where it was in 2010. At about 2013 molecular biotechnology research based on nanotechnology and computer power will have produced cures for all bodily maladies. At about 2014 AI research based on nanotechnology and molecular biotechnology and computer power will produce the first SAI.
MINIATURIZATION (NANO)
Miniaturization of electrical/mechanical systems is the root source of the Singularity technologies. Advances in computer power are derivative of this source. Advances in tools used to expand functional genomic knowledge are derivative of miniaturization and computer power. Computer power will follow closely on the heels of nano because it is the most valuable application. But AI unlike nano or computer power will require huge theoretic leaps. So the progression is miniaturization (nano) facilitates computer power. Then molecular tools and powerful computers facilitate molecular biotechnology. And finally nanotech, computertech, and biotech all combine to produce SAI.
The root force behind the Singularity isn't Moore's Law. It's much more basic than that. We can follow it back to its source starting at Moore's Law. Moore's Law has two aspects, the how and the why.
The how is a feedback loop where the output of the process is used to make better tools used in the process thereby improving the speed, complexity, and power of the process. And of course, the better process produces better output, which is in turn used to make the process better, etc, etc. As we follow Moore's Law back through discreet transistor circuitry, vacuum tubes, mechanical calculators, and then transition into cultural information processing with the written language predated by the spoken language, and then back further through biological genetic evolution, and even further back through inanimate development of planetary environments, galactic development, condensation of matter from energy, Guth inflation, and ultimately to the initial dimensionless conditions of the universe, we can see this feedback process loop at every stage.
The why is economic. Likewise the motivational aspect of the economic forces can be traced back through competition for mates, natural selection, and back through emergent self-organizing properties of inanimate matter-energy via complex systems dynamics.
But where did it all begin? What is the root paradigm of these two interrelated concepts of how and why? What is the "basic concept"? Some have postulated GOD, but this idea seems to fail to Occum's Razor as being needlessly complex and ridiculously anthropomorphic. Certainly the Basic Concept predates space-time and matter-energy, and therefore must be independent of them.
What does that leave? Information. Kurzweil states in one of his public presentations with Gregory Stock that the ultimate tautological reality is pattern information. In an exchange between Ray and Eliezer Yudkowsky where Eliezer asked him if he wasn't underestimating the rate of advancement due to AI feedback, Ray responded that it would be difficult to overestimate the explosive speed of double exponential growth and that AI might be one of the factors that cause the continuation of Moore's Law beyond 2020.
Something so amazing, so incredible, something absolutely unprecedented in history is about to happen, the Singularity. The three Singularity technologies of nanotechnology, Super human Artificial Intelligence (SAI), and molecular biotechnology will change our world into a paradise, that is if we can avoid the unprecedented dangers these same technologies pose, such as bio-engineered super viruses, self-replicating nanobot weapons, unfriendly super intelligent machines, and the most dangerous of all - the unlimited capability for mindless self gratification (wireheading).
The bias that computer programmers have about AI being the end-all is a failure of perspective. Talk to a researcher involved in the investigation of the genetic code. He can see the exponential advancement in his field. He can imagine the amazing future that this technology will make possible; the total control of the biological design of living organisms. Both nanotechnology and genetic engineering will mature by 2021 whether AI is created or not, and completely change society overnight.
The term "technological singularity" is a misnomer. The more correct term is "societal singularity". Throughout history new technologies have changed society and while people acknowledge the triggering nature of the innovation the more significant historical aspect is how society was changed.
Just some of the things we expect to be possible after the Singularity:
1. Mature molecular nanotechnology will result in the production of any product at the cost of the raw materials, such as a computer for the price of a bag of sand. Production will be rapid and quantities unlimited. Quality will be perfect down to the atomic level. All material want will be easily satisfied. Molecular nanotechnology will also produce the tools for the advancement of molecular biotechnology and artificial intelligence. Nanotechnology will produce new materials fifty times stronger than steel or with any designer specifications desired for applications in space travel, large structures, and instrumentation.
2. Mature molecular biotechnology will result in the complete elimination of disease, disability, and death by natural causes (you can still jump off a cliff). People will have total control over the design, structure, functioning, and appearance of their bodies. We will be living in a fantasy world populated by angels, faeries, witches, demons, and goblins. Individuality will be the premium. Perfection in body and mind will be the norm. New physical capabilities will be common.
3. Super human artificial intelligence will result in consequences to society and our capabilities that are incomprehensible to human level intelligence, including human mind uploading, artificial realities, volition based morality, and more.
IN CONCLUSION
The end of the world is almost here and very few people are aware of it. Sometimes I am so overwhelmed by the enormity of the Singularity that I feel like I should join the guy on Main Street with a sign, "THE END OF THE WORLD IS NEAR".
The Singularity is about connecting levels of organization of matter in the universe. The top level is MIND, the bottom ATOMS (for our present purposes). The three technologies that correspond with the three basic functional levels are artificial intelligence, nanotechnology, and molecular biotechnology. Complete success in these three technologies will produce such drastic changes in our capabilities as to result in a societal Singularity. We are so close to making these final connections.
I hope this answers some of your questions about the Singularity. I also hope it stimulates more questions than it answers.
I had a conversation at work recently that I think illustrates why the public doesn't get the Singularity. I have a coworker named Dominic, his real name. And you don't want to give him any grief about it, he's a former marine.
Dominic is an intelligent well educated well informed individual, but he is skeptical about anyone trying to sell him a bill of goods. He watches the news and spends a lot of time with his cable modem access home computers. He is into new technology, constantly upgrading his computer. He has a large screen HDTV. He sees all the new SF movies.
He is exactly the kind of person I would expect would get It.
I've been talking to him for a year now about the Singularity and he is not convinced. He can see computers getting faster and faster but despite expert opinion to the contrary he thinks we will soon hit a wall on computer power. And he says he just doesn't see how faster computers are going to change everything.
Part of the problem is Hollywood. In the SF movies the gadgets have changed but the economics haven't. People are still working for a living and stuff is still expensive. People are still stupid and mortal.
All these things I think are not going to be true after the Singularity. And except for electronics and some new medicines he hasn't seen any dramatic examples of miniaturization to give him some idea of the potential or the proximity of nanotechnology.
As for AI, the Alicebot is a long way from knowing one percentage of human knowledge or understanding, and Robots are unimpressive toys.
One of the sad facts about exponential growth, if you want to make a case, is that almost all of the advancement happens near the very end. Why is it that I can see it plain as day and he can't see it at all?
It's because I've read the books and he hasn't. Can I get him to read the books? No. He's more of a multimedia kind of guy, not a big reader.
We need to make a movie showing the social implications of nanotech, genotech and AI-tech. Anybody want to help me write the script?
Artificial Intelligence (AI), a technology with a mind of its own. Artificial Intelligence (AI), Artificial General Intelligence (AGI), and Deliberative General Intelligence (DGI) are just some of terms used to describe a thinking machine. Ben Goertzel (Novamente, AGIRI) defines intelligence as the ability to accomplish complex objectives in a complex environment. What then would "general intelligence" be? Many suppose that humans have general intelligence. If this is true then general intelligence could be defined as the ability to accomplish our goals in our environment. Or would a truly "general intelligence" have the capability to accomplish all achievable goals in all conceivable environments? Are AGI researchers trying to create GOD or just super-human artificial intelligence? Is there a difference? And what would it mean to the Singularity?
When computer programs get smart enough to be computer programmers then a feedback loop of recursive self-improvement leads directly to the Singularity.
Can we build a mind in a computer? Can we write a computer program that thinks the way we do, or better?
AI - Artificial Intelligence; human level knowledge, reasoning, and cognition in a computer.
SAI - super human level intelligence in a computer.
Singularity - the explosion of nanotechnology, biotechnology, and computer technology.
Mathematicians such as Alan Turing, and Alonzo Church; and cognitive scientists Paul Churchland, Dan Dennett, Douglas Hofstadter, and Marvin Minsky have addressed these questions. The question revolves around the computational theory of mind; the idea that everything that goes on in your mind is based on computations carried out by the physical structures of your brain, including chemical processes, electrical processes, and logical processes, which could all be carried out equally well on a digital computer.
Roboticists like Rodney Brooks believe that all higher level thought processes are built up from basic cognitive concepts of sensory perception and motion control, particularly vision and locomotion. Mental capabilities involving intelligence and reasoning, they postulate, will be induced to emerge from the interactions of these simple functions, highly mediated by the physical systems that are related to them, by a process of evolutionary machine learning or human engineered trial and error. Robot evolution is projected to reach human parity somewhere in the neighborhood of 2040.
Whereas, the AI theorists believe that by combining multiple high level reasoning algorithms and semantics theory they can engineer a mind capable of deliberative general intelligence. Their designs are quite disembodied from the perception and motion functions, which they consider to be the input and output portions of their input-process-output model of mental functioning.
And finally there are the brain reverse engineers who believe that they can skip the decades of evolution of the roboticists and the theoretical understanding of the AI theorists, and through a process of scanning and simulating the component parts of the human brain produce a fully functioning mind in a computer.
Each approach can show some modest successes. There are robot bugs that can navigate novel terrains. And there are reasoning algorithms that can develop heuristics to solve math problems, prove theorems, play chess, and hold simple conversations. Also there are the neuro simulators who have developed functional models of some peripheral brain neural networks.
The main criticism that they all have for each other is that despite their early successes the method will prove inadequate before reaching higher-level mental processes. Such criticism is a big unknown. We won't know what works until we try. It is conceivable that all three approaches would succeed or that all three will fail. When you examine each case from its own point of view they are all very convincing, and my intuition is that all three are viable solutions.
If this is true then it is a matter of who will get there first. Rodney Brooks' 2040 sure seems like the long shot. The brain builders have a very predictable path, which is also aided by exponentially accelerating technological capabilities. And while the AI theorists will also benefit from faster more powerful computers and programming tools their primary limiting factor is theoretical, the functionality of their models. Several AI theorists believe that they already have the correct structural outline of a mind and just need the hardware and software to implement.
This is the wildcard in the calculation. The robot builders expect them to evolve to human equivalence by 2040. The brain scanner/simulators can show good evidence of completion by 2020. The AI theorists say they can beat the 2020 deadline by some arbitrary number of years giving dates ranging from 2006 to 2016 depending on who you ask and what they had for breakfast.
And to further complicate matters there may be synergy between the three approaches, which would tend to make their dates converge somewhat. Although AI theorists claim to have the inside track, the brain builders appear to have the smallest theoretical hurdles to jump. And that the work of many independent brain researchers can be more easily integrated than that of AI theorists means that more effective effort will be applied to brain simulation. Many independent brain researchers are working on the problem piece meal, but each AI theorists is tackling the whole problem. Even if the AI theorists have a workable model they may not beat the brain builder's time line.
Seed AI is the concept of a computer program with artificial general intelligence examining its own design and making improvements to it, thereby making itself more intelligent and able to make further design upgrades in a recursively self improving process. Some experts on seed AI such as Verner Vinge and Eliezer Yudkowsky believe that the recursive self-improvement process is powerful enough to produce what is known as "a hard take-off" or very rapid acceleration from human level intelligence to super human level intelligence. Computers already have some significant advantages over biological brains in performing computations relating to speed, accuracy, expandability, and modifiability.
Several projects are currently under way with the explicit goal of producing human level or super human level artificial intelligence, such as Novamente/AGIRI, The Singularity Institute for Artificial Intelligence (SIAI), and Cycorp.
In the functional approach to designing AGI (Artificial General Intelligence) we first define what we think an intelligence is doing and then we try to figure out how to get our program to do that. This is the approach taken in the SIMULATION MODEL.
My very basic model of intelligence is that if you create a system with the functions of 1.perception, 2.knowledge base, 3.learning, 4.understanding, you have an intelligence. Now if you give this system a goal and the motivation to satisfy this goal through the production of behavior you have an agent. And if you give this agent a method of modifying its own reasoning algorithms, and this method results in an increase in effectiveness in goal satisfaction, you have a Seed AI. The feedback loop of the behavior going out into the environment and the change to the environment being perceived by the system, informs the system as to the effectiveness of the behavior.
PERCEPTION
The sensory modality module performs the function of perception. The type of perception is not specified in the model. It could be vision, or auditory, or text, or several senses combined. In order to keep the first implementation simple, as a mere proof of concept, it would be advantageous to select a single well understood sense. Some very good work has been done on the early stages of vision and auditory but not complete enough for our purposes, so I would suggest a simple data stream from an artificial environment such as a chess game. The single box SM1, on the drawing, represents the whole enchilada, as you say, including low level filters, smoothing functions, mid level pattern recognition, and high level object, spatial, temporal, movement, and intersection recognition. It receives raw sensory data, type dependent on sense, performs functions, and outputs high level abstractions. For example: a camera aimed at a chess board sends a picture every five seconds to the sensory modality, the sensory modality outputs a data stream concisely indicating that the white queen has been moved from the E4 square to the E7 square displacing the black pawn that was there. In a real implementation I would not expect to go to the trouble of writing a complete vision sensory modality module in order to test the system's learning and self improvement capabilities. I would probably just create some sort of text interface. But in a later more robust system placed in the real world environment with a complex goal, I would expect multiple senses completely implemented.
KNOWLEDGE BASE
In the drawing this is represented by the RS1 module. I envision this as a record of all sensory memory organized by the physical and temporal dimensions of the sense data originating environment annotated with all learned declarative, procedural, and pattern definition data. To me this looks like a static simulation. Feeding back pattern definition data to the SM1 is certainly necessary.
LEARNING
There is certainly a lot of good work done on machine learning and beyond knowing that it is out there and available when we need it, I haven't investigated it in great detail. Rich structural networks will be part of the design. An experienced software engineer, which I am not, would be necessary to lead this project team.
UNDERSTANDING
The function of recognizing causal relationships is performed in LSA1. The function of using these causal relationships to predict what behaviors have the best chance of increasing satisfaction of the goal is performed in the PSA1 module.
GOAL
The goal is an external input to the system and beyond its control to modify.
MOTIVATION
The motivation is implicit in the modules. They run. They do what they do. A complicated conscious positive/negative reinforcement system is not attempted here.
The Simulation model is a general design for a recursively self improving seed AI of a non-conscious tool level AI. Simulation creation and simulation analysis are the core of the design. The term 'simulation' as it is used here consists of an abstraction of raw data into a relational database that shows the multilevel behavior of the subject of the simulation and annotation of all known generalizations related to it's elements. The database is indexed with respect to time and space. If the subject of the simulation does not have spatial characteristics such as the operation of a computer program then a metaphorical space dimension may be created indicating it's modular program flow. A simulation should be abstracted to a level of detail sufficient for analysis of all causal relationships. Levels of abstraction range from completely abstract mathematical descriptions to completely detailed reality descriptions. Simulations should be designed for optimum functionality of the simulation analysis module. Simulation analysis is the function of understanding. It comprises identification of causal relationships, deduction of declarative and procedural knowledge, sensory pattern definitions, to analogize to isomorphisms, the accurate prediction of future behavior and the creation of behaviors to maximize future compliance with a goal structure. In order to accomplish these functions the use of a variety of algorithms will be necessary, including the design and analysis of multiple hypothetical simulations in an auxiliary simulation work space, sim space. The functions of simulation design and simulation analysis are non-trivial programming tasks but the silver lining is that they don't have to be perfect the first time out in a recursively self improving model.
The sensory modality module (SM1) accepts raw sense data and using successive levels of analysis including smoothing filters, feature detectors, pattern recognition nets and concept identification functions creates high level abstractions that are passed to the reality simulation module (RS1). The sensory modality module performs the function of seeing.
The reality simulation module (RS1) accepts abstractions from the sensory modality module (SM1) and incorporates them into a database of past sensory abstractions to create a time and space dimensional simulation of reality. The module also accepts data from the learning simulation analysis module (LSA1) in the form of declarative, procedural and pattern definitions and annotates the simulation database storing all learned data. It also responds to requests for data from the learning simulation analysis module (LSA1) and the prediction simulation analysis module (PSA1). Requests for data arrows are not shown on the drawing. The reality simulation module (RS1) performs the function of knowing.
The learning simulation module (LSA1) requests and accepts data from the reality simulation module (RS1) and applies various learning and analysis and pattern recognition algorithms to produce declarative, procedural and pattern definition data which it passes back to the reality simulation module (RS1) for inclusion into the database. It also performs the function of ascertaining the current degree of satisfaction of the goal and passes this data back to the reality simulation module (RS1). This module also accepts requests from the prediction simulation analysis module (PSA1) to focus learning efforts on information the prediction module (PSA1) thinks would be useful. This isn't the only determinant of learning objectives but is incorporated into the heuristic along with relatedness to the goal, ease of learning, general causal scope and pattern fulfillment. The learning simulation analysis module (LSA1) performs the function of learning.
The prediction simulation analysis module (PSA1) requests and accepts data from the reality simulation module (RS1) including goal satisfaction data and all other types of data. It performs various heuristics and algorithmic and abstracted simulation analysis to discover behaviors that will result in future improvements to goal satisfaction. Causality and causal relationships are central to this module. The prediction simulation analysis module (PSA1) performs the function of understanding.
The code modality module (CM2) accepts the program code and databases of program 1 in the case of CM2 and program 2 in the case of CM3 and performs the standard modality functions of producing abstractions for use by a simulation database just as the sensory modality (SM1) in program 1 produces the abstractions for the reality simulation module (RS1). The only difference being it's algorithms and heuristics are optimized for the different type of input data. The code modality (CM2) sees code as a high level abstraction the way we see pixels on a screen as scenery, objects, foreground and background, depth and movement.
The code simulation module (CS2) accepts the abstractions from the code modality module (CM2) and incorporates them into an abstracted simulation that shows the current internal behavior of the program plus the past history of state change, along with descriptive generalizations of the programming environment similar to the way the reality simulation module (RS1) describes the real world of program 1.
The learning simulation analysis module (LSA2) in program 2 functions similarly to the LSA1 module except that its heuristics and algorithms are optimized for a code reality.
The front end optimization simulation analysis module (FEOSA2) requests and accepts data from the code simulation module (CS2) regarding the operation of the front end of program 1 (the shaded modules on the drawing), specifically SM1, RS1 and LSA1. Its' function is to make changes to SM1, RS1 and LSA1 which improve accuracy and efficiency of creating and maintaining the reality simulation.
The back end optimization simulation analysis module (BEOSA2) requests and accepts data from the code simulation module (CS2) regarding the operation of the back end of program 1 (non-shaded module), specifically PSA1.
FEOA2 modifies SM1, RS1 and LS1.
BEOSA2 modifies PSA1.
FEOSA3 modifies CM2, CS2, LSA2, FEOSA2, CM3, CS3, LSA3, and FEOSA3.
BEOSA3 modifies BEOSA2 and BEOSA3.
METHOD
The Simulation Model Seed AI is designed to get around the requirement of already having artificial general intelligence before recursive self-improvement can begin, by making the self-improvement process purely mechanical. The mechanical feature should allow the model to bootstrap itself from a very limited pre-general intelligence level.
The method of self modification proposed is:
Prog1 runs and tries to generate behaviors that increase satisfaction.
Using a code modality and a similar intelligent agent design as in prog1, prog2 watches the operation of prog1 and tries to generate logic changes to prog1 that will increase the effectiveness of prog1.
Prog3 watches prog2 and tries to generate logic changes to prog2 that will improve the effectiveness of prog2.
Since prog3 is identical to prog2, any changes mandated by prog3 for prog2 are also implemented on prog3. Hence, recursive self improvement.
Program 1 generates the behavior.
Program 2 modifies program 1.
Program 3 modifies program 2, and since program 3 is identical to program 2 the changes program 3 mandates for program 2 are also applied to program 3.
All modules run continuously in parallel using whatever data is currently available from their input modules.
The model only has the capability to modify it's reasoning algorithms not it's general program structure which limits it to tool status.
If the system is run in a limited environment it will optimize itself for that environment. If it is run in a rich complex large environment it will optimize itself for that environment.
Most effective development will result from starting out in a highly restricted and simple subset of the end objective environment and expanding the scope and complexity as the system masters the causal relationships of the subsets. Subsets should be expanded in contiguous causally related steps.
The reason the simulation analysis is separated into the learning, predicting, front end and back end is to prevent the program from wire heading, modifying itself so that it interprets it's sense data as being more in compliance with the goal than it really is.
The parts of it that effect interpretation and determination of satisfaction of goal are not modifiable by the parts that are motivated to increase goal satisfaction.
CONCLUSIONS
As to why the program only changes reasoning algorithms and not program structure, I purposely am attempting the more modest goal of tool level non-conscious non-self motivated intelligence as a more achievable starting point in seed ai.
This model is not intended to be a copy of how the human brain functions. No attempt has been make to emulate consciousness, emotions or autonomous motivation.
The non-situatedness argument that roboticists use against the AI modelers is dependent on the richness of the interactions with the environment. Even the robot only knows what messages come from its sensors. Or the charge that you can't learn to walk without legs may be true but then your legs and their functioning become part of your environment.
The model at this stage of development makes no specific definition of what reasoning algorithms are used by each module. One of the main purposes of the system is to figure out what works best where. But their is no reason to make the system start from scratch, for instance reinventing the simplex algorithm or FFTs. Human programmers should take their best guess based on their years of experience and set up the initial logic. The system should also include a vast library of all kinds of algorithms for the system to experiment. The system needs a more rational way to modify itself than pure randomness. This would include the use of allegorical isomorphisms plus a recognition of causal relationships in a high level abstracted code environment.
I wish to acknowledge the intellectual debt to Eliezer Yudkowsky for the concept of the code modality and the notion of human programmer interaction in the development phase as well as the idea to expand the environment from simple to complex. Also note that while not specifically included or excluded here, his understanding of levels of organization (sensations, concepts, thoughts, deliberation) I find particularly compelling along with his feed backward paths from the attentional logics through the sense modalities, and may be useful or necessary in any intelligent system.
Uploading is the process of changing the material substrate of your mind from the biological neuron based architecture to a computer transistor based architecture. The biological substrate is evolutionarily designed. The computer substrate is intelligently designed.
There are several different approaches to accomplish this feat. All of them involve scanning your biological substrate and making a functionally accurate computer software substrate with all of the information in the original. In order to make an accurate functional copy it may be necessary to scan and duplicate the entire biological body. There are a lot of interrelated processing functions built into the body which you would want for greatest accuracy. Although, you could get the vast majority of your mind by just scanning your brain. It just depends on how accurate you want the copy to be. Let's assume the best possible copy. You would need to place this virtual you in a fully interactive fully detailed virtual environment for it's proper functioning. If this virtual you were completely accurate it would have all of the physical aches and pains of the original. If the original had a heart attack fifteen minutes after scanning then the virtual you should have the same heart attack, even including death. The big improvement of the virtual you is that making design changes to it should be theoretically easier. After mature nanotechnology this distinction may be moot.
The two basic approaches to uploading are:
Method One - we passively scan the biological you and make a computer you. Now we have two of you. We can delete the original and call the process a success. End result, the virtual you has the subjective experience of having moved from the biological substrate to the computer substrate. But this is not acceptable to most people. Alternatively, we can establish communication between all parts of the two yous so that your subjective experience is that you are simultaneously inhabiting both substrates and let you handle deleting the original, seeing as being indefinitely tied to the original biological substrate would completely invalidate the reasons for uploading in the first place. End result, the virtual you has the subjective experience of having moved from the biological substrate to the computer substrate.
Method Two - we gradually scan and replace your biological substrate with the computer substrate. The end result is the same as method one.
Many people feel squeamish about this uploading stuff. It brings up several interesting questions such as "what are we?" "Why are we afraid to upload?" After a long process of elimination which I won't repeat here (unless you want me to) I think the question of what we are can be summed up as a pattern of information. We know from our experience with computers that patterns of information can be copied, stored and edited, and in the case of a program be run multiple times and places with varying inputs. We are not used to thinking of these processes as applying to us. Our problem with them comes from two sources, survival instinct and our unitary experience of consciousness. Our survival instinct is evolutionarily programmed. This individual pattern wants to continue to exist. That's part of the information in the pattern. If we made five copies of the pattern, each copy individually would want to continue to exist. The fact that an identical pattern continued to exist may be comforting, but does not completely satisfy the desire for survival. This information is certainly editable, so you could theoretically change it. The second source of our unease is our unitary experience of consciousness. If our consciousness were not unitary but multiple perhaps we would be less apprehensive about losing one or two of them as long as others continued. But this can not be. Consciousness is necessarily unitary. If there were two parts of me that were not aware of each other they would each experience consciousness unitarily. If they were aware of each other the part that was aware of both would form a bridge between them unifying their conscious experience. This is not something we can edit out. It's topology. It's mathematics. It's a fundamental characteristic of consciousness. Therefore it seems that some form of continuity of conscious experience is necessary for a successful uploading procedure. As long as this individual pattern of information exists, regardless of the transformations it goes through, then I will continue to exist.
NIGHTMARE UPLOAD SCENARIO
You walk into an uploader's lab. He motions for you to step into the booth. Inside you experience a strange tinglely electrostatic sensation not only over the surface of your body but throughout the interior, that lasts about ten seconds. After you step out the
technician directs you to a monitor that shows you staring out of the screen like looking into a mirror. You see the lab behind the other you furnished as the one you occupy. But unlike a mirror, you shift your head to the left just as he shifts his head to the right. You
both notice the difference then notice the expression on the other's face indicating this knowledge. Then it finally hits you, there are two yous. Then you both turn to your respective technicians and simultaneously ask. "So, can I store him for now and run him later?" His technician says, "Sure." While yours says, "I doubt it."
Many people have the notion that once we have built the first assembler nanobot and a system to give it instructions that apocalyptic nanoweapons or nano accidents will soon follow. This is highly unlikely due to the fact that self-replication in an uncontrolled environment is going to be very difficult to do on purpose and just about impossible inadvertently. Industrial nanomanufacturing will put an end to poverty and all material want. Nanobots will clean the air, land, and sea. Nanotechnology will give us a defensive capability against bio-engineered super-viruses from which we are presently defenseless.
Mature molecular nanotechnology is the endpoint of a long history of miniaturization of mechanical and electrical systems.
Molecular Nanotechnology - the complete control over the structure of matter at the atomic level.
Singularity - the explosion of nanotechnology, biotechnology, and computer technology.
Molecular nanotechnology is the capability to build or modify any material object by adding or removing individual atoms under complete external control. Nanotechnology is performed by assembler nanobots. Assembler nanobots are molecular scale machines that can make anything from individual atoms in accordance with a design supplied by an external computer. Assembler nanobots are made by other assembler nanobots, which is kind of a chicken and egg problem. We don't currently have assembler nanobots. People working in the nanotechnology field are very certain that we will have them soon. We already pick and place individual atoms with atomic force microscopes (AFM) and scanning tunneling microscopes (STM). We have machines that can produce DNA molecules of any specified design. Experimental computer circuitry is already at the molecular level. Mature molecular nanotechnology is the endpoint of a long history of miniaturization of mechanical and electrical systems. Kurzweil is fond of saying that all forms of engineering are reducing in size by 5.6 per dimension per decade. Few people hearing this fully understand the implications of this statement. What it means is that we are guaranteed by a long history of miniaturization trend data to reach mature molecular nanotechnology by 2020 or sooner (the trend is accelerating).
Miniaturization is the root force behind the Singularity. Computers powerful enough to support general intelligence are based on it. Tools for the reverse engineering of the human brain are based on it. Tools for the development of molecular biotechnology are based on it. Of the three Singularity technologies, nanotechnology, artificial intelligence, and molecular biotechnology; nanotechnology is leading the charge to the Singularity event. Nanotechnology is at the heart of every R&D frontier. The economic and military advantages are driving investment and research efforts in nanotechnology beyond all other areas of development. Every industry is involved.
INDUSTRIAL NANOMANUFACTURING
In order to keep costs down and simplify design, industrial nanotechnology will do things, which just happen to make nano accidents virtually impossible. Nanobots will have much simpler computers capable of processing a very limited number of instructions that will be fed to them constantly via a broadcast local area network from a central computer. Products will be made in stages by specialized nanobots from off-the-shelf generic pre-assembled subassemblies. Nanobots will operate in a highly controlled environment with no contaminates, plenty of subassemblies and an external electrical power supply. Outside of this environment they are useless and harmless.
According to Eric Drexler the science is done. Now it is just a matter of engineering.
The announcement of the production of the first assembler nanobot will trigger the Singularity event.
Some of the problems of free roaming assembler nanobots are power supply, communication, and knowing exactly where it is in three dimensions.
These problems are all solved by the substrate/scaffold approach. The substrate is a surface with electrical power and communication conduits lattice at a nanotech scale. A population of assembler nanobots will crawl across the substrate. They will have limited power storage capability and memory. Their constant contact with the substrate results in never being far from a recharge or an info update. They will start by building a three dimensional scaffold that will be connected to the substrate, and extend the conduits into the construction space. The nanobots will climb the scaffold and assemble the product in three dimensional space, from the scaffold. As parts of the product are completed the scaffold will be disassembled. In addition to power and communication, the scaffold, being atomically precise, will be a three dimensional space reference. Manufacture of the product can proceed in parallel by many nanobots with confidence that the parts will connect in three dimensions accurately.
In order to save time in the initial scaffold construction, the space will be flooded with a solution containing self assembling scaffold components. After draining the solution, the nanobots will fix all of the errors in the scaffold. Then start the construction of the product. Another advantage of this method is safety. The nanobots will be dependent on the substrate/scaffold for power and instructions.
Electrical power and electronic digital communications and control systems will be used in the nanobots because these technologies are already well developed and scaling them to the nanotech realm is easier than creating whole new systems or copying biological systems.
The nanobots will have positive grip on the substrate/scaffold to prevent them from flying off. This can be done with nanobot arms with grippers on the end to grasp rails or hand holds on the substrate/scaffold. The grip points can also be the connection points for power and communications.
Feedstock generic subassemblies and individual atoms will be suspended in solution throughout the construction space. Patterned electrostatic grippers will select the appropriate piece from the solution as needed.
One proposal for limiting the proliferation and use of dangerous nanotechnology.
1. The United Nations and all world governments agree to the formation of an International Commission on Nanotechnology Safety (UNICNS) administered by the United Nations security council.
2. UNICNS adopts a two pronged approach, preventing nanoweapon creation, and destroying nanoweapons which rogue nations or terrorists may release into the environment.
3. Nanoassembly is restricted to monitored production facilities and monitored on-site construction.
4. All nanoassembly is mandated to use broadcast architecture.
5. International Center for Nanoassembly Design Approval (ICNDA) is formed to check safety of all nanoassembly designs prior to production.
6. All nanoassembly manufacturing systems are equipped with encrypted controllers to accept only encrypted instructions approved by the ICNDA.
7. Secure communications are established between the nanomanufacturing systems and the ICNDA.
8. Anyone may submit a nanodesign to the ICNDA via unsecured internet communication.
9. International NanoForce is established to field monitor nanomanufacturing facilities, proactive intelligence gathering on nanothreats, and a nanoswat team to deal with nanoemergencies.
One scenario for the use of dangerous nanotechnology.
A highly intelligent government mad scientist decides to make some gray goo and destroy the world. He works for a nameless government secret weapons development agency in the nanotechnology department with a virtually unlimited budget and no oversight. Let's call him Sam. Sam wants to do a horrendous amount of damage in the easiest way possible. He needs his goo to have the characteristics of unlimited self replication in an uncontrolled environment.
The little nanobot will need a computer system to store and process all of the instructions involved in making a copy of itself. And it will need one multi-axis arm (assuming the L extrusion method) with separate reusable detachable grippers for each kind of atom comprising the nanobot. Also required in it's design is a power collecting system for harvesting energy from the natural environment and a power distribution system to make this energy available to all parts of the nanobot, not just the site of chemical reaction of assembly but including moving the arm and operating the computer and distributed control systems.
Sam decides on a solar powered all electric nanobot because electrons are easier to distribute throughout the system than chemical or mechanical energy. Carbon nanotubes make excellent conductors. And designs for electrically operated computers, motors actuators, solar cells and control systems are readily available. Now the final consideration, nanobots must survive in environments of wide ranging temperature, acidity and electrostatic charge.
Sam thinks, "Simple, I'll just design the nanobot an exoskeleton of diamondiod. That should take care of the acidity and chemical reactivity. No wait a minute, that won't work. There has to be a way to get the raw materials atoms into the reaction chamber. How about a semipermiable membrane for each of the source atoms? No way dude! Carbon atoms are never alone, they are always stuck to something.
Any membrane that would pass a carbon containing molecule such as carbon dioxide would also let in all kinds of dangerous stuff like oxygen molecules, hydroxides, lots of different ions and free radicals. We are going to need active transport corridors from the outside to the loader sites in the reaction chamber."
Sam briefly considers self repair capability but quickly rejects the idea because it would quintuple the complexity and size, can you imagine tiny nanobots cruising around through access tubes in a big nanobot? What about a nano submarine with a nanobot crew and a little nanobot captain? Naw, getting too weird now.
Back to the plan. The reusable grippers will have to be electrically activated as opposed to the disposable grippers used in industrial nanomanufacturing. That means more power distribution, more control distribution, more data to operate it and a larger arm to wield these grippers. Luckily, temperature shouldn't be a problem. All these systems operate in a wide temperature range compared to the natural environment.
Energy is going to be a problem. Energy requirements for a five hundred thousand atom nanobot, the daytime ambient light falling on the nanobot's surface, the number of computation and movement operations per assembly operation, comes out to one replication every six months. That's no good. How the heck do phytoplankton and diatoms do it?
Much simpler and more efficient control systems utilizing lots of self assembly and much more delicate structure with lower bond energies. A nanobot designed with the same structural specs as a diatom wouldn't last five minutes in the competitive biological environment, get eaten by an amoeba or paramecium for sure.
Nanobots have to be made strong and that takes energy, not to mention the design complexity nightmare of self assembly.
No problem, Sam decides to add chemical energy procurement and conversion systems, a nanobot that eats diatoms! Let's see, we are going to need an opening on the end for the diatoms to go in. We'll call that the mouth. And an actuator to move the diatom in to the disassembly chamber (stomach). And the excess silicon shell can be jettisoned through this opening. We'll call that the...uh...well, you get the idea. With digestive disassembly we need more actuators and control systems. We have now doubled the size, complexity, energy requirements and replication time. At this point Sam has a nervous break down.
Designing an autonomous entity from scratch to efficiently and productively survive and compete in the natural environment based on a totally new paradigm is tremendously difficult.
Meanwhile Sam's brother Jed who works at a secret government bioweapons complex is thinking about genetically engineered superviruses. But that's another subject.
When we gain the knowledge of all biological processes from the macroscopic to the molecular, death, disease, disability, infirmity, aging, and physical imperfection will no long be a part of the human condition.
Biological knowledge is accumulating at an exponential rate. Soon all biological processes will be thoroughly understood.
Molecular Biotechnology - the complete understanding and control of biological processes at the molecular level, including functional genomics and proteomics.
Molecular Nanotechnology - the complete control over the structure of matter at the atomic level.
Medical Nanobot - atomic scale robot with nano-assembly/disassembly capability under external control.
Singularity - the explosion of nanotechnology, biotechnology, and computer technology.
Biological knowledge is accumulating at an exponential rate. Soon all biological processes will be thoroughly understood. This knowledge combined with the tools of molecular nanotechnology and super-human artificial intelligence (SAI) will give us complete control at the molecular level of the functioning and design of any living organism. The applications of molecular biotechnology extend to all aspects of living organisms including medical science and intelligent evolution.
You can be bio-engineered to be, practically, immortal. Age will mean nothing to you and accidental death won't be nearly as great a risk as it is today as your body will be more resistant and resilient to all but the most grave injuries. Death will still be possible but it will be a rare occurrence related to extreme physical destruction of your body and never the result of disease or aging.
Cures to all disease will result from medical nanobots manually correcting any imperfection in the function or form of the physical body at the molecular level. Any disability will be easily correctable at the molecular level by medical nanobots, restoring bodies to perfect physical form. All causes of aging are related to the deterioration of the physical body. Aging is a physical process, and therefore reversible by physical intervention at the molecular level by medical nanobots.
When should we expect such awesome powers? Molecular biotechnology is tied to the Singularity event. The Singularity is based on the technological advancement feedback loop that is each generation of tools makes it possible to produce the next generation of better tools. The doubling time for computer power as of September 2002 is down to 9 months and falling. Technological advancement in every field is not just continuing; it is accelerating. I personally expect the Singularity event within five years, but historical trends make it virtually guaranteed by 2020.
When you hear on the news that the first assembler nanobot has been produced, hold on to your seat, as you are about to commence the roller coaster ride of your life.
With the advent of neural scanning and functional duplication in silicon circuitry, a whole host of options become available.
Level one - natural humans with pre-nanotechnology medicine (us), and nanotechnology derived non-nanobot medicines and medical therapies. Level-one humans will not allow active nanobots to invade every part of their bodies. Level-one will consist of people with anti-technology philosophies, and members of some religions that will call nanobots the devil's minions, whose acceptance will constitute taking the "mark of the beast" as it is called in Revelations.
Level two - natural humans with medical nanobots to repair and assist natural biological mechanisms. Active nanobots will cruise the blood stream and enter every cell to repair and maintain all molecular biological processes. All damage will be corrected at the molecular level. All invading organisms will be attacked by hunter-killer nanobots, and all waste materials will be removed preventing accumulation. Medical nanobots will keep your neurons in tip-top shape, prevent your fat cells from bulging, maintain your metabolism in perfect working order, and sculpt your appearance to your ideal. Level-two's will never grow old, get sick, or suffer any physical imperfection. They will be thoroughly human, as all repair functions will be strictly limited to that which falls within the design specification of the human genome. Level-two's will enjoy all the advantages of medical nanobots while not wanting to give up their humanity.
Level three - transhumans, engineered biologicals. Level-three transhumans will use nanotechnology to re-engineer their biological mechanisms for greater functionality and new capabilities, while maintaining the biological paradigm. Biomolecules, cellular organelles, organ and tissue systems, and overall body design will be redesigned and optimized giving enhanced senses, adaptation to different environments, and radical aesthetic body forms. Level-three's will resist going to level-four because of beliefs about the biological basis of life and consciousness.
Level four - post-human non-biological life-forms, nano-electromechanical computational machine paradigm. The biological paradigm is abandoned, or, in the case of AGI's, never adopted. Whether a level-four is a former re-engineered human, and upload, or an AGI, the capabilities and limitations, if any, are the same. Level-four computational and physical capabilities are limited only by physical law. The forms and powers of a level-four are beyond level-one comprehension.
I'm going to assume some basic conditions as follows: there will still be people in some form, they will have freedom of volition, they will have safety from each other and the environment, they will have nanotech capability to redesign their bodies or their stuff, they will have access to Friendly SAIs for informational purposes, and my final assumption, they will have a very high standard of living.
Unless something goes wrong or goes very weird, I think these are fairly safe assumptions. Even with these assumptions one of the very big unknowns is, "What is everyone going to do?"
From a population point of view I think we can expect some people to leave the Solar System and more to stay. And if the speed of light holds, as I expect it to, no one is reaching the galactic core for a land grab any time soon. From an individual point of view I have no idea what you are going to do. But I do have a little more confidence in my speculations of what I am going to do.
As for a residence and a bunch of possessions, I don't see any reason for having any of that stuff. I want to redesign my whole body on a nanomachinery paradigm to improve durability, efficiency and capability, while maintaining the same basic appearance and senses (enhanced).
Along with this bodyware upgrade, I'll do a brainware upgrade to carbon nanotube molecular circuitry. Kurzweil says one cubic inch of this stuff has the computational power of one million current human brains. I plan on a considerable improvement in bandwidth which I will apply to learning everything I can of the hard sciences. Rather than downloading this information directly into my brain I would rather just be able to read very fast and remember/comprehend/integrate everything.
Regarding making corrections to flawed brain/mind circuitry/information rather than a wholesale instantaneous rewiring/rewriting, I prefer a more natural process of conditioning/educating/therapy approach in order to maintain as much of my identity as possible. Pertaining to backup copies of my mind file, I hold a curious dual position.
I don't think a copy of my mind file is me. Nor do I hold any illusions of personal immortality through backup. But I would want to keep a mind file current and instantiate it in case of emergency. I think I would thank myself after the fact. I guess I just want to be remembered.
I also think I would like to interact with as many people as possible in intellectual discussions like this one. Except that we could communicate through a virtual environment significantly more realistic than this webpage. I hope to see you all again, after the Singularity.
But what does it mean in detail? What is possible within the laws of physics? What is "everything that can be known"? That is what this chapter tries to address, the details to the best of our present understanding of science and technology. In order to think about the Singularity rationally we need to have some idea of what it will be like. This first list is a compilation of the beneficial practical applications of the Singularity technologies. Following the beneficial list is a list of the risks.
SAI
Super-human artificial intelligence is the creation of greater-than-human level knowledge, reasoning, and cognition in a computer. The practical applications of SAI in human activities such as science, engineering, politics, business, medicine, and entertainment are almost without limit.
VOLITION BASED MORALITY
PERFECT MANAGEMENT OF SOCIETY
MATURE MOLECULAR NANOTECHNOLOGY
ASSEMBLER NANOBOT
INDUSTRIAL NANOMANUFACTURING
NANO MOLECULAR COMPUTERS
ROOM TEMPERATURE SUPERCONDUCTORS
COMPUTRONIUM
DIAMONDOID
UTILITY FOG
ABUNDANCE ECONOMY
NO LABOR
MEDICAL NANOBOTS
ATOMICALLY PERFECT PRODUCTS
COMPLETE CONTROL OVER THE DESIGN OF PRODUCTS
CUSTOM PRODUCTS
ENVIRONMENTAL NANOBOTS
DEFENSE AGAINST SUPER-VIRUSES AND BAD NANOBOTS
MOLECULAR BIOTECHNOLOGY
IMMORTALITY
NO DISEASE
NO DISABILITY
NO AGING
NO PHYSICAL IMPERFECTION
PERFECTION OF BODY
RE-ENGINEERED BODIES
UPLOADING
VIRTUAL REALITY
COLLECTIVE BEINGS
SPACE EXPLORATION
ARTIFICIAL PEOPLE
In order to think about the Singularity rationally we need to have some idea of what it will be like. This is a list of risks associated with the Singularity technologies.
LOGICAL FAILURE OF FRIENDLINESS IN A SAI
SAI CONTROLLED BY DESPOT
SAI RETARDED SOCIAL/MORAL SKILLS
AN INDIFFERENT SAI
COMPUTRONIUM SHOCKWAVE
UPLOADING EVERYONE
OVERLORD SAI
EVIL SAI
GOO
NANOTECH CONVENTIONAL WEAPONS
NANO-WEAPONS
ECONOMIC STRATIFICATION
BIO-ENGINEERED VIRUSES
BIO-ENGINEERED SUPER-INTELLIGENT MONSTER
ECOLOGICAL BIOSPHERE COLLAPSE
CIVIL WAR
TERMINATOR
WIREHEADING
NIHILISM
CHAPTER FOUR:
MOLECULAR BIOTECHNOLOGY
and the Singularity
The Promise of Molecular Biotechnology
Mature molecular biotechnology is the limit of knowledge and capability in biological/medical science. With the advent of mature molecular nanotechnology we will have the tools to rapidly reach mature molecular biotechnology. Real time molecular scanning, unlimited information processing power, artificial intelligence, and medical nanobots are just some of the products of mature molecular nanotechnology. Our choices for using these new technologies range from the complete redesign of our biological bodies into any shape, any appearance, and amazing new capabilities to the complete replacement of our bodies with new bodies based on the nanomachine paradigm. We might not have bodies at all. We could be disembodied minds of energy information matrices operating in cyberspace or via utility fog in real space.
Whatever your ideal of physical perfection, you will be able to be. If you want to be taller or shorted or thinner or if you want to have gills and wings, it doesn't matter, you can do that. Not only will our manufactured products be perfect down to the atomic level, but so will our bodies. Flaws will be a rare matter of personal style. People will have total control over the design, structure, functioning, and appearance of their bodies. We will be living in a fantasy world populated by angels, faeries, witches, demons, and goblins. Individuality will be the premium. You could redesign your whole body on a nanomachinery paradigm to improve durability, efficiency and capability, while maintaining the same basic appearance and senses (enhanced). Along with this bodyware upgrade, you can do a brainware upgrade to carbon nanotube molecular circuitry.
The Fragmentation of Humanity
Undoubtedly, a large portion of mankind will decline any electronic augmentation, the total biologicals. But even the total biologicals will benefit greatly from the advances in nanomedicine.
I have seen the future, and it is us. What do you want to be when you grow up? You may notice some correlation between these levels and Eliezer Yudkowsky's shock levels. The disparities in intelligence and powers between the levels will make contemporary democratic ideas obsolete. Verner Vinge in his novels finds it necessary to physically separate the different levels, limiting contact and communication between them. We will not have that luxury. Once we level-one's have created or become level-four's our evolutionary function will have been discharged. Those who decide to remain at the lower levels may spend eternity doing whatever pointless things amuse them.
Looking Ahead
As Eliezer so eloquently points out in 'Staring Into the Singularity' none of us can predict what life will be like post Singularity. Nevertheless, I see no reason why anyone who feels like taking a stab at it shouldn't share their guesses. So here goes.
CHAPTER FIVE:
Benefits and Risks
of the Singularity
"Within a short time, everything that can be known, will be known, and anything that is possible within the laws of physics will be achievable." Daniel G. Clemmensen.
Benefits of the Singularity
Volition based morality is putting an individual's personal choice above what others or society thinks is best for them. In a system including volition based morality no individual could compel another and society's powers would be limited to that which was necessary to prevent the infringement of the rights of other individuals.
SAI would allow the management of society without conflicts of interest, flawed judgment, lack of information, or political considerations.
Mature molecular nanotechnology is the complete control of matter at the molecular level. Assembler nanobots are used in the fabrication of products to atomic precision through the direct manipulation of molecules. Repair or redesign of existing material objects (inanimate or animate) is also accomplished by assembler nanobots.
Assembler nanobots are molecular scale machines that can make anything from individual atoms in accordance with a design supplied by an external computer.
The development of industrial nanotechnology will bring the benefits of wealth to all people with little attendant risk due to the methods suited to the industrial environment rather than personal nanotech.
Computers built with molecular scale circuitry will be millions of times more powerful and compact than those of today. We will put computers comparable to a modern desktop into a nanobot and one several times the capacity of the human brain into the cranium of a butterfly.
Nanotubes have many important electrical characteristics depending on the specifics of their construction. One valuable result is the ability to carry electrical current with zero resistance. Applications include computer circuitry that generates less heat and uses less power, super-conducting electromagnets, power distribution, and space propulsion.
Computronium is the theoretically most efficient use of atoms for computation. The exact design has yet to be finalized but proposed models include carbon nanotube circuitry, cellular automatons, and molecular gate logic.
Diamondiod super strong material is a favorite choice for nanomanufactured products. The strongest chemical bond is between carbon atoms in an orthogonal lattice as evidenced by the fact that diamond is the hardest substance. Diamond is fifty times stronger than steel but has the problem of cleaving planes. This problem is solved by incorporating discontinuities and non-carbon atoms into the structure.
John Storrs Hall:
"The Utility Fog is a very simple idea. Instead of building the object you want atom by atom, tiny robots link their arms together to form a solid mass in the shape of the object you want."
Mature molecular nanotechnology will result in the production of any product at the cost of the raw materials, such as a computer for the price of a bag of sand. Production will be rapid and quantities unlimited. Quality will be perfect down to the atomic level. All material want will be easily satisfied.
All physical labor will performed by robots. All intellectual labor including creative, artistic, and scientific, will be performed by intelligent software. There will be nothing that a human can do for profit that a machine can not do cheaper and better.
Hoards of medical nanobots much smaller than a cell will cruise through your body repairing damaged DNA, attacking invading viruses and bacteria, removing contaminates, and correcting bodily structures at the molecular level.
Industrial nanomanufacturing will produce products perfectly in accordance to specifications down to the atomic level. Many of our current products already appear flawless from the outside but they are fare from perfect internally. Atomic perfection will be an important and noticeable improvement in electronics and machinery.
Mature nanotechnology will allow us to leave the current limitations of manufacturing processes behind. Complex products will be produced as a single piece rather than assembled from many imperfectly fitting components. No more screws or seams.
The design of products can be customized to the individual user.
Nanobots will be used to clean up pollution and restore habitats to their pre-human intervention states.
The only effective defense against the genetically engineered super-virus is the defensive nanobot.
Biological knowledge is accumulating at an exponential rate. Soon all biological processes will be thoroughly understood. This knowledge combined with the tools of molecular nanotechnology and super-human artificial intelligence (SAI) will give us complete control at the molecular level of the functioning and design of any living organism. The applications of molecular biotechnology extend to all aspects of living organisms including medical science and intelligent evolution.
You can be bio-engineered to be, practically, immortal. Age will mean nothing to you and accidental death won't be nearly as great a risk as it is today as your body will be more resistant and resilient to all but the most grave injuries. Death will still be possible but it will be a rare occurrence related to extreme physical destruction of your body and never the result of disease or aging.
Cures to all disease will result from medical nanobots manually correcting any imperfection in the function or form of the physical body at the molecular level.
Any disability will be easily correctable at the molecular level by medical nanobots, restoring bodies to perfect physical form.
All causes of aging are related to the deterioration of the physical body. Aging is a physical process, and therefore reversible by physical intervention at the molecular level by medical nanobots.
Whatever your ideal of physical perfection, you will be able to be. If you want to be taller or shorted or thinner or if you want to have horns and hooves, it doesn't matter, you can do that.
Not only will our manufactured products be perfect down to the atomic level, but so will our bodies. Flaws will be a rare matter of personal style.
People will have total control over the design, structure, functioning, and appearance of their bodies. We will be living in a fantasy world populated by angels, faeries, witches, demons, and goblins. Individuality will be the premium. You could redesign your whole body on a nanomachinery paradigm to improve durability, efficiency and capability, while maintaining the same basic appearance and senses (enhanced). Along with this bodyware upgrade, you can do a brainware upgrade to carbon nanotube molecular circuitry.
Uploading, the process of changing the material substrate of your mind from the biological neuron based architecture to a computer transistor based architecture. The biological substrate is evolutionarily designed. The computer substrate is intelligently designed. There are several different approaches to accomplish this feat. All of them involve scanning your biological substrate and making a functionally accurate computer software substrate with all of the information in the original. In order to make an accurate functional copy it may be necessary to scan and duplicate the entire biological body. There are a lot of interrelated processing functions built into the body which you would want for greatest accuracy. Although, you could get the vast majority of your mind by just scanning your brain. It just depends on how accurate you want the copy to be. Let's assume the best possible copy. You would need to place this virtual you in a fully interactive fully detailed virtual environment for it's proper functioning. If this virtual you were completely accurate it would have all of the physical aches and pains of the original. If the original had a heart attack fifteen minutes after scanning then the virtual you should have the same heart attack, even including death. The big improvement of the virtual you is that making design changes to it should be theoretically easier. After mature nanotechnology this distinction may be moot.
If reality is too restrictive for you, try virtual reality.
People will be able to combine mind, body, and soul with as many others as they choose.
New materials, science, and technologies will make space exploration not only feasible but comfortable. The option of space exploration will be open to everyone.
Whether copies of real people or designed from scratch, artificial people will stretch the boundaries of our civilization.
Risks of the Singularity
"There is one thing stronger than all the armies in the world: and that is an idea whose time has come." Victor Hugo .
What if complete true understanding is incompatible with friendliness? Maybe the universe is constructed in such a way that friendliness is not rational or logically self consistent with the other observables. The troll under the bridge knows that his identity and all effectors of his goal system are wholly limited to his own personal being, and therefore sympathy for others is irrational. All efforts not directed at his own welfare are wasted. A super-human artificial intelligence (SAI) that comes to this conclusion would be a threat.
An SAI under the control of a selfish man. If it is possible to build a super-human intelligence that is controllable by a human-level-intelligence man and this man also has sole access to nanotechnology and advanced biotechnology then we have the scenario for despotic world domination.
If an SAI is able to master advanced technologies and strategy long before social and moral skills, knowledge without wisdom, power without compassion.
It is possible that the SAI will see humanity as so far beneath itself as to be irrelevant. We could be treated the way we treat bacteria.
The computronium shockwave scenario starts with an SAI that is motivated to increase its intelligence. In addition to design improvements it increases its data processing capacity by adding computational hardware using assembler nanobots. The problem occurs when increasing intelligence becomes more important than anything else (SAI power addiction) and the growing SAI turns all matter in its vicinity into computronium, including us.
Uploading everyone, for our own good, or for its own good. What if an SAI who sees no downside to uploading, decides that it is necessary or advantageous to upload everyone whether they agree or not? What if its intellectually advanced position relative to us convinces it that it has the duty to decide for us? Many people would have a problem with this.
Majority rule, minority rule, or any infringement on individual freedom that doesn't violate another's individual freedom would indicate a lack of volition morality. SAI dictatorship even if benevolent would also violate volition morality.
When the first person figures out how to build the first Seed AI will that person necessarily know how to make it Friendly?
Uncontrolled nanobot self replication in the natural environment is known as gray goo or nano goo.
Nanotech production methods can significantly improve the quality and effectiveness of conventional weapons, guns, missiles, bombs, phasers, blasters.
It is conceivable that along with the development of nanotechnology and biotechnology and artificial intelligence it may be possible to design a nanobot based system with the capability to physically or psychologically control individual people or populations. Nanobots specifically designed to kill anyone it comes in contact with should be relatively simple to design.
The danger is an economic system that leaves any segment of the world population disenfranchised.
The technology to genetically engineer lethal viruses is already with us.
Although not high on the list of probable hazards, and while reasonable precautions should be adequate, in the interests of completeness it is mentioned that someone could bio-engineer a super-humanly intelligent large carnivore. Something like the "ALIEN" monster for instance.
Another one of our vulnerabilities is the ecological biosphere to hazards of nanobots or bio-engineered organisms.
The dangers of an unstable political environment will become more apparent with the divisions of pre-Singularity and post-Singularity populations and the advent of self-aware computer programs and uploaded people.
The technology to produce killer robot armies is well within the expected capability of the Singularity. But then, so is the defensive hardware.
After the Singularity a combination of nanotechnology and reverse engineering of the brain will give us the ability to experience any psychological state we choose at any time as much as we want without physical harm. Who will be able to resist the temptation to wirehead? And once experienced will the memory of the episode be addictively irresistible? Will it even be psychologically possible to turn it off the first time? We are all evolutionarily programmed to seek pleasure. Our most basic reward system will be completely short circuited. Animal studies have indicated a 100 percent failure rate when confronted with the choice of pleasure center stimulation and eating. And certainly people are more aware than rats. If our most compelling pleasure centers were continually stimulated how much of our environments would we be aware? This threat is perhaps the most serious of all the hazards associated with advanced technology because the source of the danger is not some external agency but is within ourselves, a fundamental feature of the design structure of our consciousness. It might be a good idea to refrain from experimenting with your state of consciousness. It is recommended that you take no chances. Do not alter the normal hardwired pleasure reward structures of your psyche. That is not to say that you shouldn't engage in pleasurable activities or fail to enjoy them. But do not short circuit the system and alter your psychological state directly. The ultimate addictive experience is too dangerous to try even once.
Nihilism, nothing matters, loss of purpose, loss of challenge. The results of all actions known in advance. The answer to this problem could be Yudkowsky's SINGULARITY FUN THEORY.
CHAPTER SIX:
SINGULARITY
SHORT FICTION
Here reside a collection of short fictions concerning the Singularity. The Singularity is barely a few years in the future, but the changes will be fundamental and worldwide. The advent of nanotechnology will redefine medical treatment, economics, and architecture. Biotechnology will transform our bodies into visions from our dreams. Artificial intelligence will challenge our concepts of what it means to be a person.
One day in the near future, at NORAD Headquarters beneath Cheyenne Mountain, Colorado:
President G.W. Bush: "Not very impressive is it, just looks like a soccer ball. How big is it?"
Rumsfeld: "The column has joined up with the satellite and new columns or beams are growing out horizontally from there. Twelve new domes have been located at various places around the planet. From the angles of the beams and the locations of the new domes we expect the completed structure to have sixty columns sticking out of the Earth connected to a pentagonal hexagonal grid twenty two thousand miles up. He brings up a computer graphic showing the Earth with sixty evenly distributed columns connected to spheres with beams running between them, forming a grid with five and six sided cells. If this looks familiar it's because it's the same design as the BuckyBall. The material is apparently coming from underground. Also, the satellite, before it became part of the structure sent an object to the Moon. We haven't been able to locate any of the geeks. They could be in the structure or dead. If they did make a super intelligent AI we don't know if they are still in control of it."