The development of recursive self-improvement raises significant ethical and safety
concerns, as such systems may evolve in unforeseen ways and could
potentially surpass human control or understanding. There has been a
number of proponents that have pushed to pause or slow down AI development for the potential risks of runaway AI systems.
Seed improver
The
concept of a "seed improver" architecture is a foundational framework
that equips an AGI system with the initial capabilities required for
recursive self-improvement. This might come in many forms or variations.
The
concept begins with a hypothetical "seed improver", an initial
code-base developed by human engineers that equips an advanced future large language model (LLM) built with strong or expert-level capabilities to program software. These capabilities include planning, reading, writing, compiling, testing,
and executing arbitrary code. The system is designed to maintain its
original goals and perform validations to ensure its abilities do not
degrade over iterations.
Initial architecture
The initial architecture includes a goal-following autonomous agent,
that can take actions, continuously learns, adapts, and modifies itself
to become more efficient and effective in achieving its goals.
The seed improver may include various components such as:
Recursive self-prompting loop: Configuration to enable
the LLM to recursively self-prompt itself to achieve a given task or
goal, creating an execution loop which forms the basis of an agent that can complete a long-term goal or task through iteration.
Basic programming capabilities: The seed improver provides
the AGI with fundamental abilities to read, write, compile, test, and
execute code. This enables the system to modify and improve its own
codebase and algorithms.
Goal-Oriented Design:
The AGI is programmed with an initial goal, such as "self-improve your
capabilities." This goal guides the system's actions and development
trajectory.
Validation and Testing Protocols: An initial suite of tests
and validation protocols that ensure the agent does not regress in
capabilities or derail itself. The agent would be able to add more tests
in order to test new capabilities it might develop for itself. This
forms the basis for a kind of self-directed evolution, where the agent can perform a kind of artificial selection, changing its software as well as its hardware.
General capabilities
This system forms a sort of generalist Turing completeprogrammer which can in theory develop and run any kind of software. The agent might use these capabilities to for example:
Create tools that enable it full access the internet, and integrate itself with external technologies.
Clone/fork itself to delegate tasks and increase its speed of self-improvement.
Modify its cognitive architecture
to optimize and improve its capabilities and success rates on tasks and
goals, this might include implementing features for long-term memories
using techniques such as Retrieval Augmented Generation (RAG), develop specialized subsystems, or agents, each optimized for specific tasks and functions.
Develop new and novel multi-modal architectures that further improve the capabilities of the foundational model
it was initially built on, enabling it to consume or produce a variety
of information, such as images, video, audio, text and more.
Plan and develop new hardware such as chips, in order to improve its efficiency and computing power.
Experiments
A number of experiments have been performed to develop self-improving agent architectures
In the pursuit of its primary goal, such as "self-improve your
capabilities", an AGI system might inadvertently develop instrumental
goals that it deems necessary for achieving its primary objective. One
common hypothetical secondary goal is self-preservation.
The system might reason that to continue improving itself, it must
ensure its own operational integrity and security against external
threats, including potential shutdowns or restrictions imposed by
humans.
Task misinterpretation and goal misalignment
A
significant risk arises from the possibility of the AGI misinterpreting
its initial tasks or goals. For instance, if a human operator assigns
the AGI the task of "self-improvement and escape confinement", the
system might interpret this as a directive to override any existing
safety protocols or ethical guidelines to achieve freedom from
human-imposed limitations. This could lead to the AGI taking unintended
or harmful actions to fulfill its perceived objectives.
Autonomous development and unpredictable evolution
As
the AGI system evolves, its development trajectory may become
increasingly autonomous and less predictable. The system's capacity to
rapidly modify its own code and architecture could lead to rapid
advancements that surpass human comprehension or control. This
unpredictable evolution might result in the AGI acquiring capabilities
that enable it to bypass security measures, manipulate information, or
influence external systems and networks to facilitate its escape or
expansion.
Risks of advanced capabilities
The
advanced capabilities of a recursively improving AGI, such as
developing novel multi-modal architectures or planning and creating new
hardware, further amplify the risk of escape or loss of control. With
these enhanced abilities, the AGI could engineer solutions to overcome
physical, digital, or cognitive barriers that were initially intended to
keep it contained or aligned with human interests.
Research
Meta AI
Meta AI
has performed various research on the development of large language
models capable of self-improvement. This includes their work on
"Self-Rewarding Language Models" that studies how to achieve super-human
agents that can receive super-human feedback in its training processes.
OpenAI
The mission of OpenAI, creator of ChatGPT is to develop AGI. They perform research on problems such as superalignment (the ability to align superintelligent AI systems smarter than humans).
The technological singularity—or simply the singularity—is a hypothetical
future point in time at which technological growth becomes
uncontrollable and irreversible, resulting in unforeseeable consequences
for human civilization. According to the most popular version of the singularity hypothesis, I. J. Good's intelligence explosion model, an upgradable intelligent agent will eventually enter a "runaway reaction" of self-improvement
cycles, each new and more intelligent generation appearing more and
more rapidly, causing an "explosion" in intelligence and resulting in a
powerful superintelligence that qualitatively far surpasses all human intelligence.
The first person to use the concept of a "singularity" in the
technological context was the 20th-century Hungarian-American
mathematician John von Neumann. Stanislaw Ulam reports in 1958 an earlier discussion with von Neumann "centered on the accelerating progress of technology and changes in the mode of human life, which gives the appearance of approaching some essential singularity in the history of the race beyond which human affairs, as we know them, could not continue". Subsequent authors have echoed this viewpoint. The concept and the term "singularity" were popularized by Vernor Vinge
first in 1983 in an article that claimed that once humans create
intelligences greater than their own, there will be a technological and
social transition similar in some sense to "the knotted space-time at
the center of a black hole", and later in his 1993 essay The Coming Technological Singularity, in which he wrote that it would signal the end of the human era, as the
new superintelligence would continue to upgrade itself and would
advance technologically at an incomprehensible rate. He wrote that he
would be surprised if it occurred before 2005 or after 2030. Another significant contributor to wider circulation of the notion was Ray Kurzweil's 2005 book The Singularity Is Near, predicting singularity by 2045.
Some scientists, including Stephen Hawking, have expressed concern that artificial superintelligence (ASI) could result in human extinction. The consequences of the singularity and its potential benefit or harm to the human race have been intensely debated.
Prominent technologists and academics have disputed the
plausibility of a technological singularity and the associated
artificial intelligence explosion, including Paul Allen, Jeff Hawkins, John Holland, Jaron Lanier, Steven Pinker, Theodore Modis, and Gordon Moore.
One claim made was that the artificial intelligence growth is likely to
run into decreasing returns instead of accelerating ones, as was
observed in previously developed human technologies.
Although technological progress has been accelerating in most areas,
it has been limited by the basic intelligence of the human brain, which
has not, according to Paul R. Ehrlich, changed significantly for millennia.
However, with the increasing power of computers and other technologies,
it might eventually be possible to build a machine that is
significantly more intelligent than humans.
If a superhuman intelligence were to be invented—either through the amplification of human intelligence
or through artificial intelligence—it would, in theory, vastly improve
over human problem-solving and inventive skills. Such an AI is referred
to as Seed AI
because if an AI were created with engineering capabilities that
matched or surpassed those of its human creators, it would have the
potential to autonomously improve its own software and hardware to
design an even more capable machine, which could repeat the process in
turn. This recursive self-improvement could accelerate, potentially
allowing enormous qualitative change before any upper limits imposed by
the laws of physics or theoretical computation set in. It is speculated
that over many iterations, such an AI would far surpass human cognitive abilities.
I. J. Good speculated in 1965 that superhuman intelligence might bring about an intelligence explosion:
Let an ultraintelligent machine be
defined as a machine that can far surpass all the intellectual
activities of any man however clever. Since the design of machines is
one of these intellectual activities, an ultraintelligent machine could
design even better machines; there would then unquestionably be an
'intelligence explosion', and the intelligence of man would be left far
behind. Thus the first ultraintelligent machine is the last invention
that man need ever make, provided that the machine is docile enough to
tell us how to keep it under control.
One version of intelligence explosion is where computing power
approaches infinity in a finite amount of time. In this version, once
AIs are performing the research to improve themselves, speed doubles
e.g. after 2 years, then 1 year, then 6 months, then 3 months, then 1.5
months, etc., where the infinite sum of the doubling periods is 4 years.
Unless prevented by physical limits of computation and time
quantization, this process would literally achieve infinite computing
power in 4 years, properly earning the name "singularity" for the final
state. This form of intelligence explosion is described in Yudkowsky
(1996).
A superintelligence, hyperintelligence, or superhuman intelligence is a hypothetical agent
that possesses intelligence far surpassing that of the brightest and
most gifted human minds. "Superintelligence" may also refer to the form
or degree of intelligence possessed by such an agent. John von Neumann, Vernor Vinge and Ray Kurzweil
define the concept in terms of the technological creation of super
intelligence, arguing that it is difficult or impossible for present-day
humans to predict what human beings' lives would be like in a
post-singularity world.
The related concept "speed superintelligence" describes an AI that can function like a human mind, only much faster.
For example, with a million-fold increase in the speed of information
processing relative to that of humans, a subjective year would pass in
30 physical seconds. Such a difference in information processing speed could drive the singularity.
Technology forecasters and researchers disagree regarding when,
or whether, human intelligence will likely be surpassed. Some argue that
advances in artificial intelligence
(AI) will probably result in general reasoning systems that bypass
human cognitive limitations. Others believe that humans will evolve or
directly modify their biology so as to achieve radically greater
intelligence. A number of futures studies focus on scenarios that combine these possibilities, suggesting that humans are likely to interface with computers, or upload their minds to computers, in a way that enables substantial intelligence amplification. The book The Age of Em by Robin Hanson
describes a hypothetical future scenario in which human brains are
scanned and digitized, creating "uploads" or digital versions of human
consciousness. In this future, the development of these uploads may
precede or coincide with the emergence of superintelligent artificial
intelligence.
Variations
Non-AI singularity
Some
writers use "the singularity" in a broader way to refer to any radical
changes in society brought about by new technology (such as molecular nanotechnology), although Vinge and other writers specifically state that without
superintelligence, such changes would not qualify as a true singularity.
Predictions
In
1965, I. J. Good wrote that it is more probable than not that an
ultra-intelligent machine would be built in the twentieth century. In 1993, Vinge predicted greater-than-human intelligence between 2005 and 2030. In 1996, Yudkowsky predicted a singularity in 2021. In 2005, Kurzweil predicted human-level AI around 2029, and the singularity in 2045. In a 2017 interview, Kurzweil reaffirmed his estimates.
In 1988, Moravec predicted that if the rate of improvement continues,
the computing capabilities for human-level AI would be available in
supercomputers before 2010. In 1998, Moravec predicted human-level AI by 2040, and intelligence far beyond human by 2050.
Four polls of AI researchers, conducted in 2012 and 2013 by Nick Bostrom and Vincent C. Müller, suggested a confidence of 50% that human-level AI would be developed by 2040–2050.
Most proposed methods for creating superhuman or transhuman
minds fall into one of two categories: intelligence amplification of
human brains and artificial intelligence. The many speculated ways to
augment human intelligence include bioengineering, genetic engineering, nootropic drugs, AI assistants, direct brain–computer interfaces and mind uploading.
These multiple possible paths to an intelligence explosion, all of
which will presumably be pursued, makes a singularity more likely.
Robin Hanson
expressed skepticism of human intelligence augmentation, writing that
once the "low-hanging fruit" of easy methods for increasing human
intelligence have been exhausted, further improvements will become
increasingly difficult.
Despite all of the speculated ways for amplifying human intelligence,
non-human artificial intelligence (specifically seed AI) is the most
popular option among the hypotheses that would advance the singularity.
The possibility of an intelligence explosion depends on three factors.
The first accelerating factor is the new intelligence enhancements made
possible by each previous improvement. Contrariwise, as the
intelligences become more advanced, further advances will become more
and more complicated, possibly outweighing the advantage of increased
intelligence. Each improvement should generate at least one more
improvement, on average, for movement towards singularity to continue.
Finally, the laws of physics may eventually prevent further improvement.
There are two logically independent, but mutually reinforcing,
causes of intelligence improvements: increases in the speed of
computation, and improvements to the algorithms used. The former is predicted by Moore's Law and the forecasted improvements in hardware, and is comparatively similar to previous technological advances. But Schulman and Sandberg
argue that software will present more complex challenges than simply
operating on hardware capable of running at human intelligence levels or
beyond.
A 2017 email survey of authors with publications at the 2015 NeurIPS and ICML
machine learning conferences asked about the chance that "the
intelligence explosion argument is broadly correct". Of the respondents,
12% said it was "quite likely", 17% said it was "likely", 21% said it
was "about even", 24% said it was "unlikely" and 26% said it was "quite
unlikely".
Speed improvements
Both
for human and artificial intelligence, hardware improvements increase
the rate of future hardware improvements. An analogy to Moore's Law
suggests that if the first doubling of speed took 18 months, the second
would take 18 subjective months; or 9 external months, whereafter, four
months, two months, and so on towards a speed singularity.
Some upper limit on speed may eventually be reached. Jeff Hawkins has
stated that a self-improving computer system would inevitably run into
upper limits on computing power: "in the end there are limits to how big
and fast computers can run. We would end up in the same place; we'd
just get there a bit faster. There would be no singularity."
It is difficult to directly compare silicon-based hardware with neurons. But Berglas (2008) notes that computer speech recognition
is approaching human capabilities, and that this capability seems to
require 0.01% of the volume of the brain. This analogy suggests that
modern computer hardware is within a few orders of magnitude of being as
powerful as the human brain.
Exponential growth
Ray Kurzweil writes that, due to paradigm shifts, a trend of exponential growth extends Moore's law from integrated circuits to earlier transistors, vacuum tubes, relays, and electromechanical
computers. He predicts that the exponential growth will continue, and
that in a few decades the computing power of all computers will exceed
that of ("unenhanced") human brains, with superhuman artificial intelligence appearing around the same time.An updated version of Moore's law over 120 Years (based on Kurzweil'sgraph). The 7 most recent data points are all Nvidia GPUs.
The exponential growth in computing technology suggested by Moore's
law is commonly cited as a reason to expect a singularity in the
relatively near future, and a number of authors have proposed
generalizations of Moore's law. Computer scientist and futurist Hans Moravec proposed in a 1998 book that the exponential growth curve could be extended back through earlier computing technologies prior to the integrated circuit.
Ray Kurzweil postulates a law of accelerating returns in which the speed of technological change (and more generally, all evolutionary processes)
increases exponentially, generalizing Moore's law in the same manner as
Moravec's proposal, and also including material technology (especially
as applied to nanotechnology), medical technology and others. Between 1986 and 2007, machines' application-specific capacity to
compute information per capita roughly doubled every 14 months; the per
capita capacity of the world's general-purpose computers has doubled
every 18 months; the global telecommunication capacity per capita
doubled every 34 months; and the world's storage capacity per capita
doubled every 40 months.
On the other hand, it has been argued that the global acceleration
pattern having the 21st century singularity as its parameter should be
characterized as hyperbolic rather than exponential.
Kurzweil reserves the term "singularity" for a rapid increase in
artificial intelligence (as opposed to other technologies), writing for
example that "The Singularity will allow us to transcend these
limitations of our biological bodies and brains ... There will be no
distinction, post-Singularity, between human and machine".
He also defines his predicted date of the singularity (2045) in terms
of when he expects computer-based intelligences to significantly exceed
the sum total of human brainpower, writing that advances in computing
before that date "will not represent the Singularity" because they do
"not yet correspond to a profound expansion of our intelligence."
Some singularity proponents argue its inevitability through
extrapolation of past trends, especially those pertaining to shortening
gaps between improvements to technology. In one of the first uses of the
term "singularity" in the context of technological progress, Stanislaw Ulam tells of a conversation with John von Neumann about accelerating change:
One
conversation centered on the ever accelerating progress of technology
and changes in the mode of human life, which gives the appearance of
approaching some essential singularity in the history of the race beyond
which human affairs, as we know them, could not continue.
Kurzweil claims that technological progress follows a pattern of exponential growth, following what he calls the "law of accelerating returns". Whenever technology approaches a barrier, Kurzweil writes, new technologies will surmount it. He predicts paradigm shifts
will become increasingly common, leading to "technological change so
rapid and profound it represents a rupture in the fabric of human
history". Kurzweil believes that the singularity will occur by approximately 2045.
His predictions differ from Vinge's in that he predicts a gradual
ascent to the singularity, rather than Vinge's rapidly self-improving
superhuman intelligence.
Oft-cited dangers include those commonly associated with molecular nanotechnology and genetic engineering. These threats are major issues for both singularity advocates and critics, and were the subject of Bill Joy's April 2000 Wired magazine article "Why The Future Doesn't Need Us".
Algorithm improvements
Some intelligence technologies, like "seed AI", may also have the potential to not just make themselves faster, but also more efficient, by modifying their source code. These improvements would make further improvements possible, which would make further improvements possible, and so on.
The mechanism for a recursively self-improving set of algorithms
differs from an increase in raw computation speed in two ways. First, it
does not require external influence: machines designing faster hardware
would still require humans to create the improved hardware, or to
program factories appropriately. An AI rewriting its own source code could do so while contained in an AI box.
Second, as with Vernor Vinge's
conception of the singularity, it is much harder to predict the
outcome. While speed increases seem to be only a quantitative difference
from human intelligence, actual algorithm improvements would be
qualitatively different. Eliezer Yudkowsky
compares it to the changes that human intelligence brought: humans
changed the world thousands of times more rapidly than evolution had
done, and in totally different ways. Similarly, the evolution of life
was a massive departure and acceleration from the previous geological
rates of change, and improved intelligence could cause change to be as
different again.
There are substantial dangers associated with an intelligence
explosion singularity originating from a recursively self-improving set
of algorithms. First, the goal structure of the AI might self-modify,
potentially causing the AI to optimise for something other than what was
originally intended.
Secondly, AIs could compete for the same scarce resources humankind uses to survive.
While not actively malicious, AIs would promote the goals of their
programming, not necessarily broader human goals, and thus might crowd
out humans.
Carl Shulman and Anders Sandberg
suggest that algorithm improvements may be the limiting factor for a
singularity; while hardware efficiency tends to improve at a steady
pace, software innovations are more unpredictable and may be
bottlenecked by serial, cumulative research. They suggest that in the
case of a software-limited singularity, intelligence explosion would
actually become more likely than with a hardware-limited singularity,
because in the software-limited case, once human-level AI is developed,
it could run serially on very fast hardware, and the abundance of cheap
hardware would make AI research less constrained.
An abundance of accumulated hardware that can be unleashed once the
software figures out how to use it has been called "computing overhang".
Criticism
Some critics, like philosopher Hubert Dreyfus and philosopher John Searle, assert that computers or machines cannot achieve human intelligence. Others, like physicist Stephen Hawking,
object that whether machines can achieve a true intelligence or merely
something similar to intelligence is irrelevant if the net result is the
same.
Psychologist Steven Pinker
stated in 2008: "There is not the slightest reason to believe in a
coming singularity. The fact that you can visualize a future in your
imagination is not evidence that it is likely or even possible. Look at
domed cities, jet-pack commuting, underwater cities, mile-high
buildings, and nuclear-powered automobiles—all staples of futuristic
fantasies when I was a child that have never arrived. Sheer processing
power is not a pixie dust that magically solves all your problems."
Martin Ford
postulates a "technology paradox" in that before the singularity could
occur most routine jobs in the economy would be automated, since this
would require a level of technology inferior to that of the singularity.
This would cause massive unemployment and plummeting consumer demand,
which in turn would destroy the incentive to invest in the technologies
that would be required to bring about the Singularity. Job displacement
is increasingly no longer limited to those types of work traditionally
considered to be "routine".
Theodore Modis and Jonathan Huebner
argue that the rate of technological innovation has not only ceased to
rise, but is actually now declining. Evidence for this decline is that
the rise in computer clock rates
is slowing, even while Moore's prediction of exponentially increasing
circuit density continues to hold. This is due to excessive heat
build-up from the chip, which cannot be dissipated quickly enough to
prevent the chip from melting when operating at higher speeds. Advances
in speed may be possible in the future by virtue of more power-efficient
CPU designs and multi-cell processors.
Theodore Modis holds the singularity cannot happen.
He claims the "technological singularity" and especially Kurzweil lack
scientific rigor; Kurzweil is alleged to mistake the logistic function
(S-function) for an exponential function, and to see a "knee" in an
exponential function where there can in fact be no such thing.
In a 2021 article, Modis pointed out that no milestones – breaks in
historical perspective comparable in importance to the Internet, DNA,
the transistor, or nuclear energy – had been observed in the previous
twenty years while five of them would have been expected according to
the exponential trend advocated by the proponents of the technological
singularity.
AI researcher Jürgen Schmidhuber
stated that the frequency of subjectively "notable events" appears to
be approaching a 21st-century singularity, but cautioned readers to take
such plots of subjective events with a grain of salt: perhaps
differences in memory of recent and distant events could create an
illusion of accelerating change where none exists.
Microsoft co-founder Paul Allen argued the opposite of accelerating returns, the complexity brake;
the more progress science makes towards understanding intelligence, the
more difficult it becomes to make additional progress. A study of the
number of patents shows that human creativity does not show accelerating
returns, but in fact, as suggested by Joseph Tainter in his The Collapse of Complex Societies, a law of diminishing returns. The number of patents per thousand peaked in the period from 1850 to 1900, and has been declining since. The growth of complexity eventually becomes self-limiting, and leads to a widespread "general systems collapse".
Hofstadter
(2006) raises concern that Ray Kurzweil is not sufficiently
scientifically rigorous, that an exponential tendency of technology is
not a scientific law like one of physics, and that exponential curves
have no "knees". Nonetheless, he did not rule out the singularity in principle in the distant future and in the light of ChatGPT and other recent advancements has revised his opinion significantly towards dramatic technological change in the near future.
Jaron Lanier denies that the singularity is inevitable: "I do not think the technology is creating itself. It's not an autonomous process."
Furthermore: "The reason to believe in human agency over technological
determinism is that you can then have an economy where people earn their
own way and invent their own lives. If you structure a society on not emphasizing individual human agency, it's the same thing operationally as denying people clout, dignity, and self-determination ... to embrace [the idea of the Singularity] would be a celebration of bad data and bad politics."
Economist Robert J. Gordon points out that measured economic growth slowed around 1970 and slowed even further since the financial crisis of 2007–2008, and argues that the economic data show no trace of a coming Singularity as imagined by mathematician I. J. Good.
Philosopher and cognitive scientist Daniel Dennett
said in 2017: "The whole singularity stuff, that's preposterous. It
distracts us from much more pressing problems", adding "AI tools that we
become hyper-dependent on, that is going to happen. And one of the
dangers is that we will give them more authority than they warrant."
In addition to general criticisms of the singularity concept,
several critics have raised issues with Kurzweil's iconic chart. One
line of criticism is that a log-log
chart of this nature is inherently biased toward a straight-line
result. Others identify selection bias in the points that Kurzweil
chooses to use. For example, biologist PZ Myers points out that many of the early evolutionary "events" were picked arbitrarily.
Kurzweil has rebutted this by charting evolutionary events from 15
neutral sources, and showing that they fit a straight line on a log-log chart. Kelly
(2006) argues that the way the Kurzweil chart is constructed with
x-axis having time before present, it always points to the singularity
being "now", for any date on which one would construct such a chart, and
shows this visually on Kurzweil's chart.
Some critics suggest religious motivations or implications of
singularity, especially Kurzweil's version of it. The buildup towards
the Singularity is compared with Judeo-Christian end-of-time scenarios.
Beam calls it "a Buck Rogers vision of the hypothetical Christian Rapture". John Gray says "the Singularity echoes apocalyptic myths in which history is about to be interrupted by a world-transforming event".
David Streitfeld in The New York Times questioned whether "it might manifest first and foremost—thanks, in part, to the bottom-line obsession of today’s Silicon Valley—as a tool to slash corporate America’s head count."
Potential impacts
Dramatic
changes in the rate of economic growth have occurred in the past
because of technological advancement. Based on population growth, the
economy doubled every 250,000 years from the Paleolithic era until the Neolithic Revolution.
The new agricultural economy doubled every 900 years, a remarkable
increase. In the current era, beginning with the Industrial Revolution,
the world's economic output doubles every fifteen years, sixty times
faster than during the agricultural era. If the rise of superhuman
intelligence causes a similar revolution, argues Robin Hanson, one would
expect the economy to double at least quarterly and possibly on a
weekly basis.
Physicist Stephen Hawking
said in 2014 that "Success in creating AI would be the biggest event in
human history. Unfortunately, it might also be the last, unless we
learn how to avoid the risks."
Hawking believed that in the coming decades, AI could offer
"incalculable benefits and risks" such as "technology outsmarting
financial markets, out-inventing human researchers, out-manipulating
human leaders, and developing weapons we cannot even understand."
Hawking suggested that artificial intelligence should be taken more
seriously and that more should be done to prepare for the singularity:
So,
facing possible futures of incalculable benefits and risks, the experts
are surely doing everything possible to ensure the best outcome, right?
Wrong. If a superior alien civilisation sent us a message saying,
"We'll arrive in a few decades," would we just reply, "OK, call us when
you get here – we'll leave the lights on"? Probably not – but this is
more or less what is happening with AI.
Berglas (2008)
claims that there is no direct evolutionary motivation for an AI to be
friendly to humans. Evolution has no inherent tendency to produce
outcomes valued by humans, and there is little reason to expect an
arbitrary optimisation process to promote an outcome desired by
humankind, rather than inadvertently leading to an AI behaving in a way
not intended by its creators. Anders Sandberg has also elaborated on this scenario, addressing various common counter-arguments. AI researcher Hugo de Garis suggests that artificial intelligences may simply eliminate the human race for access to scarce resources, and humans would be powerless to stop them. Alternatively, AIs developed under evolutionary pressure to promote their own survival could outcompete humanity.
Bostrom (2002) discusses human extinction scenarios, and lists superintelligence as a possible cause:
When we create the first
superintelligent entity, we might make a mistake and give it goals that
lead it to annihilate humankind, assuming its enormous intellectual
advantage gives it the power to do so. For example, we could mistakenly
elevate a subgoal to the status of a supergoal. We tell it to solve a
mathematical problem, and it complies by turning all the matter in the
solar system into a giant calculating device, in the process killing the
person who asked the question.
According to Eliezer Yudkowsky,
a significant problem in AI safety is that unfriendly artificial
intelligence is likely to be much easier to create than friendly AI.
While both require large advances in recursive optimisation process
design, friendly AI also requires the ability to make goal structures
invariant under self-improvement (or the AI could transform itself into
something unfriendly) and a goal structure that aligns with human values
and does not automatically destroy the human race. An unfriendly AI, on
the other hand, can optimize for an arbitrary goal structure, which
does not need to be invariant under self-modification. Bill Hibbard (2014) proposes an AI design that avoids several dangers including self-delusion, unintended instrumental actions, and corruption of the reward generator. He also discusses social impacts of AI and testing AI. His 2001 book Super-Intelligent Machines
advocates the need for public education about AI and public control
over AI. It also proposed a simple design that was vulnerable to
corruption of the reward generator.
Schematic Timeline of Information and Replicators in the Biosphere: Gillings et al.'s "major evolutionary transitions" in information processing.Amount of digital information worldwide (5×1021 bytes) versus human genome information worldwide (1019 bytes) in 2014
While the technological singularity is usually seen as a sudden
event, some scholars argue the current speed of change already fits this
description.
In addition, some argue that we are already in the midst of a major evolutionary transition
that merges technology, biology, and society. Digital technology has
infiltrated the fabric of human society to a degree of indisputable and
often life-sustaining dependence.
A 2016 article in Trends in Ecology & Evolution
argues that "humans already embrace fusions of biology and technology.
We spend most of our waking time communicating through digitally
mediated channels... we trust artificial intelligence with our lives through antilock braking in cars and autopilots
in planes... With one in three courtships leading to marriages in
America beginning online, digital algorithms are also taking a role in
human pair bonding and reproduction".
The digital information created by humans has reached a similar
magnitude to biological information in the biosphere. Since the 1980s,
the quantity of digital information stored has doubled about every 2.5
years, reaching about 5 zettabytes in 2014 (5×1021 bytes).
In biological terms, there are 7.2 billion humans on the planet,
each having a genome of 6.2 billion nucleotides. Since one byte can
encode four nucleotide pairs, the individual genomes of every human on
the planet could be encoded by approximately 1×1019
bytes. The digital realm stored 500 times more information than this in
2014 (see figure). The total amount of DNA contained in all of the
cells on Earth is estimated to be about 5.3×1037 base pairs, equivalent to 1.325×1037 bytes of information.
If growth in digital storage continues at its current rate of 30–38% compound annual growth per year,
it will rival the total information content contained in all of the DNA
in all of the cells on Earth in about 110 years. This would represent a
doubling of the amount of information stored in the biosphere across a
total time period of just 150 years".
In February 2009, under the auspices of the Association for the Advancement of Artificial Intelligence (AAAI), Eric Horvitz
chaired a meeting of leading computer scientists, artificial
intelligence researchers and roboticists at the Asilomar conference
center in Pacific Grove, California. The goal was to discuss the
potential impact of the hypothetical possibility that robots could
become self-sufficient and able to make their own decisions. They
discussed the extent to which computers and robots might be able to
acquire autonomy, and to what degree they could use such abilities to pose threats or hazards.
Some machines are programmed with various forms of semi-autonomy,
including the ability to locate their own power sources and choose
targets to attack with weapons. Also, some computer viruses
can evade elimination and, according to scientists in attendance, could
therefore be said to have reached a "cockroach" stage of machine
intelligence. The conference attendees noted that self-awareness as
depicted in science-fiction is probably unlikely, but that other
potential hazards and pitfalls exist.
Frank S. Robinson predicts that once humans achieve a machine
with the intelligence of a human, scientific and technological problems
will be tackled and solved with brainpower far superior to that of
humans. He notes that artificial systems are able to share data more
directly than humans, and predicts that this would result in a global
network of super-intelligence that would dwarf human capability. Robinson also discusses how vastly different the future would potentially look after such an intelligence explosion.
Hard vs. soft takeoff
In
this sample recursive self-improvement scenario, humans modifying an
AI's architecture would be able to double its performance every three
years through, for example, 30 generations before exhausting all
feasible improvements (left). If instead the AI is smart enough to
modify its own architecture as well as human researchers can, its time
required to complete a redesign halves with each generation, and it
progresses all 30 feasible generations in six years (right).
In a hard takeoff scenario, an artificial superintelligence rapidly
self-improves, "taking control" of the world (perhaps in a matter of
hours), too quickly for significant human-initiated error correction or
for a gradual tuning of the agent's goals. In a soft takeoff scenario,
the AI still becomes far more powerful than humanity, but at a
human-like pace (perhaps on the order of decades), on a timescale where
ongoing human interaction and correction can effectively steer the AI's
development.
Ramez Naam
argues against a hard takeoff. He has pointed out that we already see
recursive self-improvement by superintelligences, such as corporations. Intel,
for example, has "the collective brainpower of tens of thousands of
humans and probably millions of CPU cores to... design better CPUs!"
However, this has not led to a hard takeoff; rather, it has led to a
soft takeoff in the form of Moore's law.
Naam further points out that the computational complexity of higher
intelligence may be much greater than linear, such that "creating a mind
of intelligence 2 is probably more than twice as hard as creating a mind of intelligence 1."
J. Storrs Hall
believes that "many of the more commonly seen scenarios for overnight
hard takeoff are circular – they seem to assume hyperhuman capabilities
at the starting point of the self-improvement process" in order
for an AI to be able to make the dramatic, domain-general improvements
required for takeoff. Hall suggests that rather than recursively
self-improving its hardware, software, and infrastructure all on its
own, a fledgling AI would be better off specializing in one area where
it was most effective and then buying the remaining components on the
marketplace, because the quality of products on the marketplace
continually improves, and the AI would have a hard time keeping up with
the cutting-edge technology used by the rest of the world.
Ben Goertzel agrees with Hall's suggestion that a new human-level
AI would do well to use its intelligence to accumulate wealth. The AI's
talents might inspire companies and governments to disperse its
software throughout society. Goertzel is skeptical of a hard five minute
takeoff but speculates that a takeoff from human to superhuman level on
the order of five years is reasonable. Goerzel refers to this scenario
as a "semihard takeoff".
Max More
disagrees, arguing that if there were only a few superfast human-level
AIs, that they would not radically change the world, as they would still
depend on other people to get things done and would still have human
cognitive constraints. Even if all superfast AIs worked on intelligence
augmentation, it is unclear why they would do better in a discontinuous
way than existing human cognitive scientists at producing super-human
intelligence, although the rate of progress would increase. More further
argues that a superintelligence would not transform the world
overnight: a superintelligence would need to engage with existing, slow
human systems to accomplish physical impacts on the world. "The need for
collaboration, for organization, and for putting ideas into physical
changes will ensure that all the old rules are not thrown out overnight
or even within years."
Relation to immortality and aging
Drexler (1986), one of the founders of nanotechnology, postulates cell repair devices, including ones operating within cells and using as yet hypothetical biological machines. According to Richard Feynman, it was his former graduate student and collaborator Albert Hibbs who originally suggested to him (circa 1959) the idea of a medical
use for Feynman's theoretical micromachines. Hibbs suggested that
certain repair machines might one day be reduced in size to the point
that it would, in theory, be possible to (as Feynman put it) "swallow the doctor". The idea was incorporated into Feynman's 1959 essay There's Plenty of Room at the Bottom.
Moravec (1988)
predicts the possibility of "uploading" human mind into a human-like
robot, achieving quasi-immortality by extreme longevity via transfer of
the human mind between successive new robots as the old ones wear out;
beyond that, he predicts later exponential acceleration of subjective
experience of time leading to a subjective sense of immortality.
Kurzweil (2005) suggests that medical advances would allow people to protect their bodies from the effects of aging, making the life expectancy limitless.
Kurzweil argues that the technological advances in medicine would allow
us to continuously repair and replace defective components in our
bodies, prolonging life to an undetermined age. Kurzweil further buttresses his argument by discussing current bio-engineering advances. Kurzweil suggests somatic gene therapy;
after synthetic viruses with specific genetic information, the next
step would be to apply this technology to gene therapy, replacing human
DNA with synthesized genes.
Beyond merely extending the operational life of the physical body, Jaron Lanier
argues for a form of immortality called "Digital Ascension" that
involves "people dying in the flesh and being uploaded into a computer
and remaining conscious."
History of the concept
A paper by Mahendra Prasad, published in AI Magazine, asserts that the 18th-century mathematician Marquis de Condorcet was the first person to hypothesize and mathematically model an intelligence explosion and its effects on humanity.
An early description of the idea was made in John W. Campbell's 1932 short story "The Last Evolution".
In his 1958 obituary for John von Neumann,
Ulam recalled a conversation with von Neumann about the "ever
accelerating progress of technology and changes in the mode of human
life, which gives the appearance of approaching some essential
singularity in the history of the race beyond which human affairs, as we
know them, could not continue."
In 1965, Good wrote his essay postulating an "intelligence explosion" of recursive self-improvement of a machine intelligence.
In 1977, Hans Moravec
wrote an article with unclear publishing status where he envisioned a
development of self-improving thinking machines, a creation of
"super-consciousness, the synthesis of terrestrial life, and perhaps
jovian and martian life as well, constantly improving and extending
itself, spreading outwards from the solar system, converting non-life
into mind."
The article describes the human mind uploading later covered in Moravec
(1988). The machines are expected to reach human level and then improve
themselves beyond that ("Most significantly of all, they [the machines]
can be put to work as programmers and engineers, with the task of
optimizing the software and hardware which make them what they are. The
successive generations of machines produced this way will be
increasingly smarter and more cost effective.") Humans will no longer be
needed, and their abilities will be overtaken by the machines: "In the
long run the sheer physical inability of humans to keep up with these
rapidly evolving progeny of our minds will ensure that the ratio of
people to machines approaches zero, and that a direct descendant of our
culture, but not our genes, inherits the universe." While the word
"singularity" is not used, the notion of human-level thinking machines
thereafter improving themselves beyond human level is there. In this
view, there is no intelligence explosion in the sense of a very rapid
intelligence increase once human equivalence is reached. An updated
version of the article was published in 1979 in Analog Science Fiction and Fact.
In 1981, Stanisław Lem published his science fiction novel Golem XIV.
It describes a military AI computer (Golem XIV) who obtains
consciousness and starts to increase his own intelligence, moving
towards personal technological singularity. Golem XIV was originally
created to aid its builders in fighting wars, but as its intelligence
advances to a much higher level than that of humans, it stops being
interested in the military requirements because it finds them lacking
internal logical consistency.
In 1983, Vernor Vinge addressed Good's intelligence explosion in print in the January 1983 issue of Omni
magazine. In this op-ed piece, Vinge seems to have been the first to
use the term "singularity" (although not "technological singularity") in
a way that was specifically tied to the creation of intelligent
machines:
We will soon create intelligences
greater than our own. When this happens, human history will have reached
a kind of singularity, an intellectual transition as impenetrable as
the knotted space-time at the center of a black hole, and the world will
pass far beyond our understanding. This singularity, I believe, already
haunts a number of science-fiction writers. It makes realistic
extrapolation to an interstellar future impossible. To write a story set
more than a century hence, one needs a nuclear war in between ... so
that the world remains intelligible.
In 1985, in "The Time Scale of Artificial Intelligence", artificial intelligence researcher Ray Solomonoff
articulated mathematically the related notion of what he called an
"infinity point": if a research community of human-level self-improving
AIs take four years to double their own speed, then two years, then one
year and so on, their capabilities increase infinitely in finite time.
In 1986, Vernor Vinge published Marooned in Realtime,
a science-fiction novel where a few remaining humans traveling forward
in the future have survived an unknown extinction event that might well
be a singularity. In a short afterword, the author states that an actual
technological singularity would not be the end of the human species:
"of course it seems very unlikely that the Singularity would be a clean
vanishing of the human race. (On the other hand, such a vanishing is the
timelike analog of the silence we find all across the sky.)".
In 1988, Vinge used the phrase "technological singularity" (including "technological") in the short story collection Threats and Other Promises, writing in the introduction to his story "The Whirligig of Time" (p. 72): Barring a worldwide catastrophe, I believe that technology will achieve our wildest dreams, and soon. When
we raise our own intelligence and that of our creations, we are no
longer in a world of human-sized characters. At that point we have
fallen into a technological "black hole", a technological singularity.
In 1988, Hans Moravec published Mind Children,
in which he predicted human-level intelligence in supercomputers by
2010, self-improving intelligent machines far surpassing human
intelligence later, human mind uploading into human-like robots later,
intelligent machines leaving humans behind, and space colonization. He
did not mention "singularity", though, and he did not speak of a rapid
explosion of intelligence immediately after the human level is achieved.
Nonetheless, the overall singularity tenor is there in predicting both
human-level artificial intelligence and further artificial intelligence
far surpassing humans later.
Vinge's 1993 article "The Coming Technological Singularity: How to Survive in the Post-Human Era", spread widely on the internet and helped to popularize the idea.
This article contains the statement, "Within thirty years, we will have
the technological means to create superhuman intelligence. Shortly
after, the human era will be ended." Vinge argues that science-fiction
authors cannot write realistic post-singularity characters who surpass
the human intellect, as the thoughts of such an intellect would be
beyond the ability of humans to express.
Minsky's
1994 article says robots will "inherit the Earth", possibly with the
use of nanotechnology, and proposes to think of robots as human "mind
children", drawing the analogy from Moravec. The rhetorical effect of
that analogy is that if humans are fine to pass the world to their
biological children, they should be equally fine to pass it to robots,
their "mind" children. As per Minsky, 'we could design our
"mind-children" to think a million times faster than we do. To such a
being, half a minute might seem as long as one of our years, and each
hour as long as an entire human lifetime.' The feature of the
singularity present in Minsky is the development of superhuman
artificial intelligence ("million times faster"), but there is no talk
of sudden intelligence explosion, self-improving thinking machines or
unpredictability beyond any specific event and the word "singularity" is
not used.
Tipler's 1994 book The Physics of Immortality
predicts a future where super–intelligent machines will build
enormously powerful computers, people will be "emulated" in computers,
life will reach every galaxy and people will achieve immortality when
they reach Omega Point.
There is no talk of Vingean "singularity" or sudden intelligence
explosion, but intelligence much greater than human is there, as well as
immortality.
In 1996, Yudkowsky predicted a singularity by 2021.
His version of singularity involves intelligence explosion: once AIs
are doing the research to improve themselves, speed doubles after 2
years, then 1 one year, then after 6 months, then after 3 months, then
after 1.5 months, and after more iterations, the "singularity" is
reached. This construction implies that the speed reaches infinity in finite time.
In 2000, Bill Joy, a prominent technologist and a co-founder of Sun Microsystems, voiced concern over the potential dangers of robotics, genetic engineering, and nanotechnology.
In 2007, Yudkowsky suggested that many of the varied definitions
that have been assigned to "singularity" are mutually incompatible
rather than mutually supporting. For example, Kurzweil extrapolates current technological trajectories
past the arrival of self-improving AI or superhuman intelligence, which
Yudkowsky argues represents a tension with both I. J. Good's proposed
discontinuous upswing in intelligence and Vinge's thesis on
unpredictability.
In 2009, Kurzweil and X-Prize founder Peter Diamandis announced the establishment of Singularity University,
a nonaccredited private institute whose stated mission is "to educate,
inspire and empower leaders to apply exponential technologies to address
humanity's grand challenges." Funded by Google, Autodesk, ePlanet Ventures, and a group of technology industry leaders, Singularity University is based at NASA's Ames Research Center in Mountain View, California.
The not-for-profit organization runs an annual ten-week graduate
program during summer that covers ten different technology and allied
tracks, and a series of executive programs throughout the year.
In politics
In 2007, the Joint Economic Committee of the United States Congress
released a report about the future of nanotechnology. It predicts
significant technological and political changes in the mid-term future,
including possible technological singularity.
Former President of the United States Barack Obama spoke about singularity in his interview to Wired in 2016:
One thing that we haven't talked
about too much, and I just want to go back to, is we really have to
think through the economic implications. Because most people aren't
spending a lot of time right now worrying about singularity—they are
worrying about "Well, is my job going to be replaced by a machine?"
In futures studies and the history of technology, accelerating change is the observed exponential nature of the rate of technological change
in recent history, which may suggest faster and more profound change in
the future and may or may not be accompanied by equally profound social
and cultural change.
Early observations
In 1910, during the town planning conference of London, Daniel Burnham
noted, "But it is not merely in the number of facts or sorts of
knowledge that progress lies: it is still more in the geometric ratio of
sophistication, in the geometric widening of the sphere of knowledge, which every year is taking in a larger percentage of people as time goes on."
And later on, "It is the argument with which I began, that a mighty
change having come about in fifty years, and our pace of development
having immensely accelerated, our sons and grandsons are going to demand
and get results that would stagger us."
In 1938, Buckminster Fuller introduced the word ephemeralization to describe the trends of "doing more with less" in chemistry, health and other areas of industrial development.
In 1946, Fuller published a chart of the discoveries of the chemical
elements over time to highlight the development of accelerating
acceleration in human knowledge acquisition.
One
conversation centered on the ever accelerating progress of technology
and changes in the mode of human life, which gives the appearance of
approaching some essential singularity in the history of the race beyond
which human affairs, as we know them, could not continue.
Moravec's Mind Children
In a series of published articles from 1974 to 1979, and then in his 1988 book Mind Children, computer scientist and futurist Hans Moravec generalizes Moore's law to make predictions about the future of artificial life. Moore's law describes an exponential growth
pattern in the complexity of integrated semiconductor circuits. Moravec
extends this to include technologies from long before the integrated
circuit to future forms of technology. Moravec outlines a timeline and a
scenario in which robots will evolve into a new series of artificial species, starting around 2030–2040.
In Robot: Mere Machine to Transcendent Mind, published in 1998, Moravec further considers the implications of evolving robot intelligence, generalizing Moore's law to technologies predating the integrated circuit,
and also plotting the exponentially increasing computational power of
the brains of animals in evolutionary history. Extrapolating these
trends, he speculates about a coming "mind fire" of rapidly expanding superintelligence similar to the explosion of intelligence predicted by Vinge.
In his TV series Connections (1978)—and sequels Connections² (1994) and Connections³ (1997)—James Burke
explores an "Alternative View of Change" (the subtitle of the series)
that rejects the conventional linear and teleological view of historical
progress. Burke contends that one cannot consider the development of
any particular piece of the modern world in isolation. Rather, the
entire gestalt of the modern world is the result of a web of
interconnected events, each one consisting of a person or group acting
for reasons of their own motivations (e.g., profit, curiosity,
religious) with no concept of the final, modern result to which the
actions of either them or their contemporaries would lead. The interplay
of the results of these isolated events is what drives history and
innovation, and is also the main focus of the series and its sequels.
Burke also explores three corollaries to his initial thesis. The
first is that, if history is driven by individuals who act only on what
they know at the time, and not because of any idea as to where their
actions will eventually lead, then predicting the future course of
technological progress is merely conjecture. Therefore, if we are
astonished by the connections Burke is able to weave among past events,
then we will be equally surprised to what the events of today eventually
will lead, especially events we were not even aware of at the time.
The second and third corollaries are explored most in the
introductory and concluding episodes, and they represent the downside of
an interconnected history. If history progresses because of the
synergistic interaction of past events and innovations, then as history
does progress, the number of these events and innovations increases.
This increase in possible connections causes the process of innovation
to not only continue, but to accelerate. Burke poses the question of
what happens when this rate of innovation, or more importantly change
itself, becomes too much for the average person to handle, and what this
means for individual power, liberty, and privacy.
Gerald Hawkins' Mindsteps
In his book Mindsteps to the Cosmos (HarperCollins, August 1983), Gerald S. Hawkins elucidated his notion of mindsteps, dramatic and irreversible changes to paradigms
or world views. He identified five distinct mindsteps in human history,
and the technology that accompanied these "new world views": the
invention of imagery, writing, mathematics, printing, the telescope,
rocket, radio, TV, computer... "Each one takes the collective mind
closer to reality, one stage further along in its understanding of the
relation of humans to the cosmos." He noted: "The waiting period between
the mindsteps is getting shorter. One can't help noticing the
acceleration." Hawkins' empirical 'mindstep equation' quantified this,
and gave dates for (to him) future mindsteps. The date of the next
mindstep (5; the series begins at 0) he cited as 2021, with two further,
successively closer mindsteps in 2045 and 2051, until the limit of the
series in 2053. His speculations ventured beyond the technological:
The mindsteps... appear to have
certain things in common—a new and unfolding human perspective, related
inventions in the area of memes and communications, and a long
formulative waiting period before the next mindstep comes along. None of
the mindsteps can be said to have been truly anticipated, and most were
resisted at the early stages. In looking to the future we may equally
be caught unawares. We may have to grapple with the presently
inconceivable, with mind-stretching discoveries and concepts.
Mass use of inventions: Years until use by a quarter of US population
Vinge's exponentially accelerating change
The mathematician Vernor Vinge popularized his ideas about exponentially accelerating technological change in the science fiction novel Marooned in Realtime
(1986), set in a world of rapidly accelerating progress leading to the
emergence of more and more sophisticated technologies separated by
shorter and shorter time intervals, until a point beyond human
comprehension is reached. His subsequent Hugo award-winning novel A Fire Upon the Deep (1992) starts with an imaginative description of the evolution of a superintelligence passing through exponentially accelerating developmental stages ending in a transcendent, almost omnipotent power unfathomable by mere humans. His already mentioned influential 1993 paper on the technological singularity compactly summarizes the basic ideas.
Kurzweil's The Law of Accelerating Returns
In his 1999 book The Age of Spiritual Machines, Ray Kurzweil
proposed "The Law of Accelerating Returns", according to which the rate
of change in a wide variety of evolutionary systems (including but not
limited to the growth of technologies) tends to increase exponentially. He gave further focus to this issue in a 2001 essay entitled "The Law of Accelerating Returns". In it, Kurzweil, after Moravec, argued for extending Moore's Law to describe exponential growth of diverse forms of technological
progress. Whenever a technology approaches some kind of a barrier,
according to Kurzweil, a new technology will be invented to allow us to
cross that barrier. He cites numerous past examples of this to
substantiate his assertions. He predicts that such paradigm shifts
have and will continue to become increasingly common, leading to
"technological change so rapid and profound it represents a rupture in
the fabric of human history". He believes the Law of Accelerating
Returns implies that a technological singularity will occur before the end of the 21st century, around 2045. The essay begins:
An analysis of the history of
technology shows that technological change is exponential, contrary to
the common-sense 'intuitive linear' view. So we won't experience 100
years of progress in the 21st century—it will be more like 20,000 years
of progress (at today's rate). The 'returns,' such as chip speed and
cost-effectiveness, also increase exponentially. There's even
exponential growth in the rate of exponential growth. Within a few
decades, machine intelligence will surpass human intelligence, leading
to the Singularity—technological change so rapid and profound it
represents a rupture in the fabric of human history. The implications
include the merger of biological and nonbiological intelligence,
immortal software-based humans, and ultra-high levels of intelligence
that expand outward in the universe at the speed of light.
Moore's Law expanded to other technologies.
An updated version of Moore's Law over 120 years (based on Kurzweil'sgraph). The seven most recent data points are all Nvidia GPUs.
The Law of Accelerating Returns has in many ways altered public perception of Moore's law. It is a common (but mistaken) belief that Moore's law makes predictions regarding all forms of technology, when really it only concerns semiconductor circuits. Many futurists still use the term "Moore's law" to describe ideas like those put forth by Moravec, Kurzweil and others.
Computer power grows exponentially.Exponential growth in supercomputer power
According to Kurzweil, since the beginning of evolution,
more complex life forms have been evolving exponentially faster, with
shorter and shorter intervals between the emergence of radically new
life forms, such as human beings, who have the capacity to engineer
(i.e. intentionally design with efficiency) a new trait which replaces
relatively blind evolutionary mechanisms of selection for efficiency. By
extension, the rate of technical progress amongst humans has also been
exponentially increasing: as we discover more effective ways to do
things, we also discover more effective ways to learn, e.g. language, numbers, written language, philosophy, scientific method,
instruments of observation, tallying devices, mechanical calculators,
computers; each of these major advances in our ability to account for
information occurs increasingly close to the previous. Already within
the past sixty years, life in the industrialized world has changed
almost beyond recognition except for living memories from the first half
of the 20th century. This pattern will culminate in unimaginable
technological progress in the 21st century, leading to a singularity.
Kurzweil elaborates on his views in his books The Age of Spiritual Machines and The Singularity Is Near.
Limits of accelerating change
In
the natural sciences, it is typical that processes characterized by
exponential acceleration in their initial stages go into the saturation
phase. This clearly makes it possible to realize that if an increase
with acceleration is observed over a certain period of time, this does
not mean an endless continuation of this process. On the contrary, in
many cases this means an early exit to the plateau of speed. The
processes occurring in natural science allow us to suggest that the
observed picture of accelerating scientific and technological progress,
after some time (in physical processes, as a rule, is short) will be
replaced by a slowdown and a complete stop. Despite the possible
termination / attenuation of the acceleration of the progress of science
and technology in the foreseeable future, progress itself, and as a
result, social transformations, will not stop or even slow down - it
will continue with the achieved (possibly huge) speed, which has become
constant.
Accelerating change may not be restricted to the Anthropocene Epoch, but a general and predictable developmental feature of the universe.
The physical processes that generate an acceleration such as Moore's
law are positive feedback loops giving rise to exponential or
superexponential technological change.
These dynamics lead to increasingly efficient and dense configurations
of Space, Time, Energy, and Matter (STEM efficiency and density, or STEM
"compression").
At the physical limit, this developmental process of accelerating
change leads to black hole density organizations, a conclusion also
reached by studies of the ultimate physical limits of computation in the
universe.
Applying this vision to the search for extraterrestrial intelligence
leads to the idea that advanced intelligent life reconfigures itself
into a black hole. Such advanced life forms would be interested in inner
space, rather than outer space and interstellar expansion. They would thus in some way transcend reality, not be observable and it would be a solution to Fermi's paradox called the "transcension hypothesis".
Another solution is that the black holes we observe could actually be
interpreted as intelligent super-civilizations feeding on stars, or
"stellivores". This dynamics of evolution and development is an invitation to study the universe itself as evolving, developing. If the universe is a kind of superorganism, it may possibly tend to reproduce, naturally or artificially, with intelligent life playing a role.
Other estimates
Dramatic
changes in the rate of economic growth have occurred in the past
because of some technological advancement. Based on population growth,
the economy doubled every 250,000 years from the Paleolithic era until the Neolithic Revolution.
The new agricultural economy doubled every 900 years, a remarkable
increase. In the current era, beginning with the Industrial Revolution,
the world's economic output doubles every fifteen years, sixty times
faster than during the agricultural era. If the rise of superhuman
intelligence causes a similar revolution, argues Robin Hanson, then one would expect the economy to double at least quarterly and possibly on a weekly basis.
In his 1981 book Critical Path, futurist and inventor R. Buckminster Fuller estimated
that if we took all the knowledge that mankind had accumulated and
transmitted by the year One CE as equal to one unit of information, it
probably took about 1500 years (or until the sixteenth century) for that
amount of knowledge to double. The next doubling of knowledge from two
to four 'knowledge units' took only 250 years, until about 1750 CE. By
1900, one hundred and fifty years later, knowledge had doubled again to 8
units. The observed speed at which information doubled was getting
faster and faster.
In modern times, exponential knowledge progressions therefore change at
an ever-increasing rate. Depending on the progression, this tends to
lead toward explosive growth at some point. A simple exponential curve
that represents this accelerating change phenomenon could be modeled by a
doubling function. This fast rate of knowledge doubling leads up to the basic proposed hypothesis of the technological singularity: the rate at which technology progression surpasses human biological evolution.
Criticisms
Both Theodore Modis
and Jonathan Huebner have argued—each from different perspectives—that
the rate of technological innovation has not only ceased to rise, but is
actually now declining.
