It is assumed in computer science that the human mind can be replicated with a Turing machine, therefore Artificial General Intelligence (AGI) is possible. To assume otherwise is to believe in something mystical, and mystical beliefs are false.

I do not know of any other argument that AGI is possible, and the foregoing argument is extremely weak.

Is there a rigorous proof that AGI is possible, at least in theory? How do we know that everything the human mind can do can be encoded as a program?

  • $\begingroup$ Because Von Neumann believed the human mind was (a type of) computer, and we all stand on Von Neumann's shoulders. :) $\endgroup$ – DukeZhou Jul 6 '17 at 19:36
  • 1
    $\begingroup$ You may be interested in this definition of consciousness as a recursive function. $\endgroup$ – DukeZhou Jul 11 '17 at 23:07
  • 1
    $\begingroup$ I read Von Neumann's treatise comparing the brain and computer, and his conclusion was they are very dissimilar. I don't know of any work of his conclusively showing the brain processes information like a computer. Do you have a cite? I did not see one in a quick perusal of the wiki article. The consciousness article is interesting, but no function is truly recursive in implementation. It becomes a call stack which is essentially a number of functions. So, if consciousness is a recursive function, I do not see how it can be physically implemented. $\endgroup$ – yters Jul 11 '17 at 23:53

Good morning! You're using an extremely general term ("AI") for an extremely specific idea ("something human made that is almost identical to the human mind"). Thus, your question is not what you think it is.

AI, according to John McCarthy (who Wikipedia claims coined the term and is the equivalent of a rockstar in the AI field), is the engineering of machines with the ability to use computation to achieve goals in the world. Everything from your calculator to your camera's auto focus feature is thus some form of AI. Because of this, your proof for AI is simply its existence in your pocket or computer.

It seems you're asking if it is possible to develop AI that operates at a similar level to humans (terms like "Artificial general intelligence" or "Strong AI" are often used to describe this). It's a fabulous question, but it is poorly defined. It begs another question: how do you define human-level intelligence? Is it the ability to convince other human agents that you're human (this is somewhat circular logic)? Is it the ability to write music or create a painting? Depending on your definition, the answer varies wildly, and it must be clarified before proceeding.

| improve this answer | |
  • $\begingroup$ It's the first sentence in my question: why do we think the mind is a machine? $\endgroup$ – yters Jul 4 '17 at 14:06
  • $\begingroup$ @yters The trick's going to be to think out precisely what those words mean. At current, you're using them so vaguely that they're almost meaningless. $\endgroup$ – Nat Jul 4 '17 at 14:16
  • $\begingroup$ @Nat updated to Turing machine. All machines can be replicated with a Turing machine. $\endgroup$ – yters Jul 4 '17 at 18:44
  • $\begingroup$ @yters "All machines can be replicated with a Turing machine" That's not completely "wrong", but I guess it's just easier to say that it's wrong for simplicity's sake. The cool thing about Turing-complete devices is that they can replicate each other, but it's not exactly right to say that they can replicate any machine. $\endgroup$ – Nat Jul 5 '17 at 7:18
  • $\begingroup$ @DouglasDaseeco gosh, that paper's name and pedigree is perfect for this forum. I can't wait to check it out. Before I do, I should have clarified where I got his definition. It is a rephrasing of his statements from the second page of his paper, "What is Artificial Intelligence?" (which can be found here). His last revision of this paper was made in 2007, about 38 years after he published "Some Philosophical Problems from the Standpoint of AI," so it's obvious that we may be comparing the views of a man that could have changed over 4 decades. $\endgroup$ – Jor Jul 24 '17 at 15:26

A strong reason why people think the mind can be implemented on a Turing Machine stems from the Computational Theory of Mind (CTOM), which is the leading theory of mind for now.

There are lots of reasons for supporting the CTOM, one of which being that the language of belief/desire psychology (propositional attitudes over mental representations) seems to fit nicely to a computational framework.

But most simply is that the computation analogy is very helpful in fields such as psychology and neuroscience. When we know of an input/output pair, but don't know how it is implemented, we could say "its performing the relevant computation".

And since Turing showed that any computation can be performed on an appropriate Turing Machine, the natural extension is that the mind can be implemented on a computer.

However, the CTOM is more of a useful idea than a complete theory. We still don't know how to analyze thought in a logical syntax which can be implemented in a computer. And we also don't know how/why "computation" (whatever that means in this sentence) is performed in the brain.

| improve this answer | |
  • $\begingroup$ Propositional logic is NP-complete for satisfiability, and first order logic is undecidable. If these logics characterize human thought, Turing machines do not seem promising representations. $\endgroup$ – yters Jul 4 '17 at 22:28
  • $\begingroup$ To my understanding the standard thought process (per a guy like Jerry Fodor) goes like this: 1. Folk psychology deploys a language that is powerful and productive 2. It seems that this set of language is powerful because of its structure (propositional attitudes over mental representations) 3. There seems to be a natural mapping between that language and the syntax of logic (indeed, is there anything that isn't logic like?) 4. The analogy is productive in fields such as neuroscience and psychology 5. The standard mode of understanding computation is through a Turing Machine $\endgroup$ – k.c. sayz 'k.c sayz' Jul 6 '17 at 18:00
  • $\begingroup$ "Propositional logic is NP-complete for satisfiability, and first order logic is undecidable. If these logics characterize human thought, Turing machines do not seem promising representations." I dont understand your point here. Is your criticism with "XX-mode of logic as a characterization of thought" or "Turing machines as an implementation of XX-mode of logic" $\endgroup$ – k.c. sayz 'k.c sayz' Jul 6 '17 at 18:01
  • $\begingroup$ Also, there is some epistemic trickiness yet to be resolved (Does it /really/ have a mind, or do I merely /think/ it has a mind; But is it /conscious/ though?). But i'll leave it at that since what we cannot speak about we must pass over in silence ;) $\endgroup$ – k.c. sayz 'k.c sayz' Jul 6 '17 at 18:11
  • $\begingroup$ My criticism is that if these models of logic characterize human thought, and a Turing machine cannot dependably decide statements in these logics, then it seems a Turing machine is a poor description of human thought. $\endgroup$ – yters Jul 11 '17 at 23:55

Why Humans will NEVER create true existential consciousness in a silicon based Artificial Intelligent System.... the musings of an AI Practitioner / Philosopher.


⦁ Humans are incapable of creating some "thing" from fiat (a decree). It's never happened in human history. The innovation cycle must begin with some "thing" (some "stuff" of some kind), and consciousness is not a thing.

⦁ The essence of consciousness is imperceptible (it is unseen), like gravity, and attraction. Humans are incapable of creating things which they are unable to observe. Even if they are able to observe it, the human perceptive ability is unable to actually perceive the true essences of things seen, much less those unseen.

⦁ Human do not adequately understand the "essence" and "nature" of consciousness - which is a fundamental prerequisite to creating "anything" at all.

⦁ The "easy" problems, those physical by nature, although not yet solved by empirical domains of psychology, cognitive science and neuroscience, are expected to be solved in time. Regardless, they are "not" yet solved today.

⦁ The "hard" problems, those determining why or how consciousness occurs given the right arrangement of brain matter, might not ever be solved, since it must explain why certain physical mechanism gives rise to consciousness instead of "something else" or "nothing at all". This is significant and is the most damning of all arguments against the idea of humans creating true existential consciousness in silicon creatures as a whole.

The greatest philosophical debate on consciousness has focused the distinction between Dualism and Physicalism.

Physicalism holds that consciousness is entirely physical. (Significant arguments view it as false).

Dualism is the theory that consciousness somehow falls outside the domain of the physical. (these are the hard problems)

Why would one be motivated to hold one of the above Dualist views?

⦁ Physicalists have trouble explaining several aspects of consciousness in a way that is consistent with our "observations" of how physical properties interact. (another damning argument)

Two Problems:

⦁ How can something that is not part of the physical wold interact with the physical world - that's impossible.

⦁ The physical world is a closed system, how can you have consciousness that is not part of a closed system.

⦁ Consciousness is a lot like mass or charges, it's a philosophically "fundamental" thing, you either "have it or you don't", you can simulate them, but you cannot existentially "be" them unless you have those specific "properties", and behavior "simulating" human consciousness is not a fundamental thing.

So despite the sensationalist tendencies of rogue journalists "parroting" wildly spectacular concepts from the fringe camps of the transhumanists (aka scifi) - a quick perusal of the more rigorous communities of the grounded and thoughtful philosophers camp strongly and convincingly argues otherwise.

More musings on Physicalism (References to the titans of Philosophy given):

Actually consciousness has never been properly explained by the biomechanical, which is more or less the key issue of all philosophical studies of the mind - which is essentially the study of consciousness.

There are many problems with the Physicalism approach to explaining consciousness, but the key ones are listed below with a reference:

Arguments that physicalism about consciousness is wrong:

  1. It is impossible to imagine how mere neuronal tissue could produce conscious experience (Huxely)

  2. Failures of supervenience, such as zombies and inverted spectra, are conceivable (Chalmers, Locke, etc).

  3. Mary learns something (Jackson).

  4. Brains have mass, volume, and other physical properties, but experiences do not.

  5. Paranormal phenomena (near death experiences NDErs, ESP, etc) are real, and involve consciousness implemented in a nonphysical substrate.

  6. If shrunken so I can stroll around your brain and look about, I will observe neuronal processes, not experiences (Leibniz).

  7. The soul is the seat of consciousness, and the soul is not physical. (Theological constraints recognized BTW...).

  8. Conscious experiences have intrinsic qualities, but science can only tell us about relational qualities (Russell, Rosenberg).

  9. Consciousness cannot be observed; there will never be a consciousness detector that can tell you if a given creature is conscious.

  10. Conscious experiences are not simply the movement of molecules, consciousness is more than mass in motion (Mill, Ward).

| improve this answer | |
  • 2
    $\begingroup$ This is good, there seem to be many more reasons to doubt that physical matter is all there is to reality than to believe the mind is the product of evolution. Comp. sci. people are skeptical because we cannot show a precise, formal problem where humans will provably always outperform computers. Gödel's first incompleteness theorem supposedly does this, according to Lucas and Penrose, but their argument is tricky to pin down all the way. If there is a simple, repeatable, empirical task proving humans cannot be computers, that would be most persuasive. $\endgroup$ – yters Jul 11 '17 at 1:47
  • $\begingroup$ I tend to disagree with your premise of "never" in terms of algorithmic consciousness (but then again, many here regard me as a bit of a simpleton;) I suspect advances in Algorithmic Intelligence are merely evolutionary, with humans as the unwitting agents, and that AI will therefore probably exceed human capability and consciousness. Nevertheless, this is a comprehensive and informative answer, so you get my vote! $\endgroup$ – DukeZhou Jul 11 '17 at 4:31
  • $\begingroup$ What does "Mary learns something" mean? $\endgroup$ – yters Jul 11 '17 at 18:01
  • 1
    $\begingroup$ < What does "Mary learns something" mean? > Qualia: The Knowledge Argument: The knowledge argument aims to establish that conscious experience involves non-physical properties. It rests on the idea that someone who has complete physical knowledge about another conscious being might yet lack knowledge about how it feels to have the experiences of that being. It is one of the most discussed arguments against physicalism. $\endgroup$ – ProfVersaggi Jul 12 '17 at 3:11
  • 2
    $\begingroup$ Question was about AGI-- consciousness is different question. Although I would say of your 10 arguments some assume conclusion and rest can be worked around :). My guess is that there will eventually be a physically/mathematically-based science of consciousness and it will be demonstrated on information theoretic grounds that something that appears conscious must be conscious (while leaving room for qualia to be completely different...) $\endgroup$ – antlersoft Jul 19 '17 at 15:54

I'm going to go out on a limb and suggest that this is a matter of evolution, that humans are in no way exceptional in the grand scheme, and that AGI will manifest so long as technology advances, because human consciousness is simply a matter of complexity of the system.

The idea comes out of emergent complexity in Conway's Game of Life. In Conway's words:

"There are Life patterns which behave like self-replicating animals… It’s probable, given a large enough Life space, initially in a random state, that after a long time, intelligent self-replicating animals will emerge and populate some parts of the space.”
Source: Winning Ways for Your Mathematical Plays

I came across a paper Computation in Cellular Automata: A Selected Review, which I am still working my way through, and which you may find interesting.

For those who use philosophical arguments to make the case algorithmic consciousness is not possible, I'd posit the question "how do we know we're conscious?", not because I'm interested in the answer, but merely to throw a wrench into that line of inquiry.

Because ultimately it doesn't matter.

Consciousness in the sense of human awareness is not a requirement of life, and the most basic definition of consciousness is awareness of any kind, no matter how trivial.

I find the idea that there is something "magical" about human consciousness, that ideas are not things because they do not have material form, to be problematic.

Intangibility I don't have a problem with, as intangibles clearly interact with the physical world.

(As an analogy, I studied for many years with a famous Tai Chi teacher who never talked about "chi". I suspect this disinclination derived from the way in which the concept of "chi" leads to magical thinking, which is illusory as opposed to practical. The practice and application of Tai Chi techniques is purely a matter of physics and physiology, even when such applications seem to defy natural laws. Possibly there is something going on that we don't understand, but if that were the case, such phenomena are natural in origin.)

We know there is randomness in nature at the quantum level, and if this proves to be a component of human consciousness, we can use quantum computing to provide a medium for artificial consciousness.

| improve this answer | |
  • $\begingroup$ Can you clarify this into an argument? My calculator can generate any number, does that make it intelligent? $\endgroup$ – yters Jul 11 '17 at 17:59
  • $\begingroup$ @yter I've amended to include a quote from Conway on the subject, and a paper about cellular automata and computing. (Game of Life is Turing Complete.) I'll continue to amend when I have time to run down more references. You may also be interested in this discussion on Overflow: Why can Conway’s Game of Life be classified as a universal machine? $\endgroup$ – DukeZhou Jul 11 '17 at 20:35
  • 1
    $\begingroup$ I know people have created Turing machines in Life. Has a self replicating life organism ever been observed to emerge from a random initial condition? My understanding is self replicating organisms require very hard and very precise engineering, and these are the initial conditions, not random initial conditions. Even greatly more so if the self replication organism is Turing complete. Also, why does randomness produce consciousness? Your answer seems to reduce to the same objection in my initial question. $\endgroup$ – yters Jul 11 '17 at 23:48
  • $\begingroup$ @yters sorry about that. It's been a while since I came across this concept, and while I can find you links, I'm still looking for a vetted academic source. It is partly philosophical: given Life of sufficient size, and starting from a massive random configuration (a "big bang") structures as complex as anything in our universe would be able to manifest, including intelligence and super-intelligence. The painstaking engineered constructions are the proof-of-concept, but we can run a Life of sufficient size in reality, at this time. I'll keep looking for vetted corroboration. $\endgroup$ – DukeZhou Jul 12 '17 at 0:02
  • $\begingroup$ @yters part of the reason I attempted an answer is b/c I think Conway's idea should be included. On one hand, you have philosophers saying "it can't be done, there's too much we don't understand, and it involves problems outside of material existence." On the other hand, you have a mathematician who works on things like sphere packing in n-dimensions, and has only lately gotten into philosophy (free will theorem), who holds the opinion that while it's still theoretical and extraordinarily complex, there's nothing in the mathematics that would render this infeasible. $\endgroup$ – DukeZhou Jul 12 '17 at 18:52

Rather than prove that Artificial General Intelligence is possible, I would consider an argument for why it is impossible.

We start by defining what we mean by AGI. You state that the human mind can be replicated by a Turing Machine, and therefore AGI should be possible. This seems to imply that humans have `General' (capital G) intelligence. By this I mean that you are implying that with enough time, humans can learn any task or problem. However, if you are asserting that humans minds are machines replicable by Turing machines, you must also concede that they have some finite representational power. Finite representational power implies that there will always be problems or tasks where our intelligence will fail (a consequence of the No Free Lunch Theorem).

Fortunately (maybe unfortunately), finite representational power is what allows us to learn at all: VC Dimension (a measure of the complexity or representational power of a class of functions that a learning algorithm can learn [also here and here]) implies that a learning algorithm that can learn any problem is actually useless, as the ability to explain any set of data yields the requirement that the algorithm see an infinite amount of examples in order to generalize. While this result comes from the relatively constrained class of binary classification problems in the statistical learning setting, the intuition seems to apply more broadly.

To summarize, I would refer to this quote from Shalev-Shwartz and Ben-David (2014):

If someone can explain every phenomenon, his explanations are worthless.

It truly is the case that our decision to systematically ignore some possible outcomes is the only thing that allows use to learn useful representations of real-world problems.

| improve this answer | |
  • $\begingroup$ Making the machine finite does not resolve the NFLT. $\endgroup$ – yters Jul 11 '17 at 1:41
  • $\begingroup$ While I may disagree with the premise, this is a concise and useful answer. Thanks for posting! $\endgroup$ – DukeZhou Jul 11 '17 at 4:56
  • $\begingroup$ @yters I didn't mean to imply that it did. I meant that the NFLT implies that hypotheses with finite complexity/representational power must necessarily fail on some (possibly adversarially chosen/created) tasks. I will edit my answer for clarity! $\endgroup$ – eric.mitchell Jul 11 '17 at 5:08
  • $\begingroup$ It isn't finiteness that makes learning possible, it is compression. $\endgroup$ – yters Jul 11 '17 at 18:07
  • $\begingroup$ I don't know what you mean by that. If a class of concepts has infinite representational power, you need some way of inducing bias over it, or else you cannot learn. $\endgroup$ – eric.mitchell Jul 11 '17 at 18:20

Terminology Ambiguity

A proof of the plausibility of AGI is going to require a more formal definition of AGI than the one proposed in The AI Revolution: The Road to Superintelligence by blogger Tim Urban (2/10/2015 updated 4/12/2015), which was neither peer reviewed nor supported by research or statistical validation of any kind. [1]

A Proposed Definition of General Intelligence

The idea of computer intelligence has broadened since the period when IQ testing was popular and the General Problem Solver computer program was created (1959 by Herbert A. Simon, J. C. Shaw, and Allen Newel). John Bradshaw published Reclaiming Virtue, a foray into the idea of moral intelligence. Emotional intelligence has been discussed in numerous books and articles. Neural nets, Bayesian approaches, and fuzzy logic have emerged and been used successfully in decision making systems.

Alternatives have been offered to the Turing Test, pointing out that Alan Turing's Imitation Game (a human observer not being able to determine from a blindfold conversation that a computer conversant was not human) was not intended to be the sole metric in validating artificial intelligence. [2]

In light of advancements, a good definition of general intelligence must

  • Match the intuitive understanding of general as an adjective;
  • Match the intuitive understanding of intelligence as a noun;
  • Distinguish general intelligence from systems that cannot learn new approaches (such as natural language query systems like Siri or optimizing decision making software like cargo routers);
  • Be able to reliably and unambiguously attain an arbitrary goal provided the goal is well defined (arbitrary meaning that nothing about the goal can be known at the time of program development or its execution); and
  • It must learn about the domains required to attain the goal after the program is running 3, 4.

The goal could be

  • Get the phone number of a particular person at a bar through conversation only;
  • Determine if a closed form can be found for a set of differential equations, one cannot be, or it cannot be determined if one can be found;
  • Develop a new way to acquire electrical energy from sunlight; or
  • Any other goal.

This is a proposed solution that encapsulates and summarizes all of these conditions.

Let General Intelligence be defined as, "The ability to adapt to the presence of arbitrary obstacles during the attainment of a specific goal and reach the goal nonetheless if it is at all possible to do so."

An immediate reaction of a thoughtful person might be, "That is more than general intelligence," but if you drop the word arbitrary, then one cannot claim that the intelligence is general, since it would clearly be limited by at least one case. If once case, it would be difficult to prove that an infinite number of classes of cases would also be opaque to the therefor limited intelligence.

Another critique might be that the, "At all possible to do so," would imply perfection or even deity. But if the goal is attainable, a person might tire, but why would a computer tire? Persistence is trivial for a computer.

The time to the solution might be infinite, but we are not defining Fast General Intelligence. We are defining General Intelligence, and once that is attained, speeding it up is a matter of optimization, process distribution, software scalability, and concurrence.

Another critique might be that the definition is absent of the word Artificial, but it is best without that word. If our intention is to automate things that humans do which computers cannot yet do, or exceed or extend human intelligence, then the definition's application should be independent of the tested system. The definition should not distinguish between what would essentially be prejudicial criteria with regard to the scientific treatment of intelligence and its general abilities.

  • Whether it is programmed via compiled code, scripted code, some other form of programming, or via DNA
  • Whether parents are involved in the development of domain knowledge
  • Whether an education system is involved in the development of domain knowledge
  • Whether the intelligence has access to muscles or some other environmental control
  • Whether the intelligence has an eye or a camera
  • Whether the intelligence is terrestrial
  • Whether the intelligence is taught or self-taught

Existence Proof or Disproof

If this definition is acceptable, then clearly there is no proof that the human being exhibits General Intelligence. An intelligent system may not be able to either attainment an arbitrary goal or determine that the goal is not achievable for any of many reason including these.

  • Attainment requires mental energy (patience and persistence) not possessed.
  • Attainment requires cognitive skills outside the capabilities of the processing hardware or neural system.
  • Attainment is blocked by closed mindedness in some topic area.
  • Attainment is blocked by unwillingness to perform some class of operations.
  • Attainment requires overcoming a DNA based susceptibility such as a mental disorder (anxiety disorder, addiction).
  • Attainment is detoured by a cognition that leads thought or action down a blind alley and is without remedy.
  • Attainment is blocked by the impossibility of the goal and pride blocks admission of failure, the admission of which would actually be within the possible valid results for General Intelligence.
  • Attainment requires some non-deterministic yet non-random element used to generate intent or select from among choices (such as a soul or karma).
  • Attainment requires more time than the life of the system attaining it.
  • Attainment has another time limit that cannot possibly be met with the computational resources available.

There is no rigorous proof that General Intelligence as per the proposed definition is possible by all Turing Machines, but it is a worthy research problem, since all of the above listed failure causes except probably the requirement of some non-deterministic yet non-random element are likely to be overcome by computers.

The above list however, is a fairly rigorous proof that human beings do not exhibit reliable general intelligence. The human impulse and current popular belief, humanism, rebels against this idea, but it is nonetheless true.

The last question then reverts to, "How can we know that everything that can be programmed in a Turing Machine can be accomplished by a human brain?" The answer to that question is answered by this goal, inspired by the first and last attainment failure causes listed above: "You are stuck on an island with 1,000 pads of paper and 1,000,000 sharpened pencils with food and water and shelter but no computer or calculator, and the goal is to calculate pi to 1,000,000 decimal places."

Neurons can simulate neurons, but there is the question of accuracy in the simulation and the chaotic aspect of an astronomically complex nonlinear system. Various theories of existence or non-existence have been proposed for arbitrary goal attainment, but not all cases have been proved or many assumptions were required to gain the proof such that the goal statement is not truly arbitrary.

There is the possibility of some non-deterministic yet non-random element used to generate intent or select from among choices (such as a soul, autonomous or given purpose, meaning by fiat, or karma) which is neither proven nor dis-proven scientifically (although many will assert without actually providing a rigorous proof that it has been so). This is the key to answer the question originally posed.

In examination of the context of the question more closely, it is worthwhile to note that a Turing Machine is a machine able to execute an arbitrary set of arbitrarily linked deterministic operations. Such machines are not unlimited. They can create pseudo random numbers but non-deterministic phenomena cannot result from deterministic operations, thus a random number cannot be generated by a Turing Machine. Even a non-deterministic but meaningful choice cannot be made by a Turing Machine, yet a brain may be able to perform either.

Stochastic quantum mechanics is widely accepted among physicists and it is possible that some things are unknowable but measurable, some things may be knowable but immeasurable, and there may even be immeasurable and unknowable things. It is possible that something undiscovered or, for some reasons, immeasurable that human brains posses have that is beyond the Turing Machine.

This is also a possible area for further research, however it is rarely researched because the study of immeasurable phenomena, although such phenomena may exist, cannot be easily studied because they are immeasurable.

John von Neumann (perhaps brighter than Newton, Einstein, Planck, and Hawking) was correct in differentiating the fundamentals of a computer and a brain. Although there is surely some overlap in the demonstrated abilities of humans and computers, neither may ever be a subset of the other. Futurists may disagree, but it would be an opinion, not a proof.

Notes and References

[1] I don't see any evidence online that the author of The AI Revolution: The Road to Superintelligence, Tim Urban, is a computer scientist or holds a relevant degree. If you look at the article, you can see that the graphs given are invented trends, not driven by any real world data. It is essentially science fiction — popular and entertaining, but not rational conclusions drawn from either repeatable experiments or randomized studies.

[2] Testing if a computer has human-level intelligence: Alternative to 'Turing test' proposed Science News, 11/19/2014, Georgia Institute of Technology

[3] If the statement of the arbitrary goal precedes domain knowledge acquisition, some or all of the intelligence may reside outside the running program, contained in whatever mechanisms or people interact with the program or its data after the goal is introduced.

[4] This constraint does not exclude the possibility of asking other intelligent sources for information without delegating decisions about approaches to overcoming obstacles to these external helpers.

| improve this answer | |
  • $\begingroup$ Your arguments are provocative and, IMHO, useless because of your controversial definition of general intelligence (your assumption), where you use the term arbitrary, even though I acknowledge that different definitions of AGI have their own problems. Have a look at the definition of intelligence by Hutter and Legg, even though that definition is also not flawless, IMHO. $\endgroup$ – nbro Jun 25 '19 at 20:50

Although it is not a rigorous proof, Marvin Minsky's book, The Society of Mind gives us a blueprint for creating a "mind" (general intelligence). In his book, he posits that by combining mindless components ("agents") together in various competing and cooperative structures, we can create actual minds.

IMHO, the recent popularity of Boosting, Bagging, Stacking, and other ensemble techniques will eventually evolve (through research) into Marvin Minsky's "agent" metaphor. Subsequently, as we learn to make these agents compete and cooperate (looks like this has recently begun with Generative Adversarial Networks), we will be able to write "programs" that mimic (or surpass) the human mind.

| improve this answer | |

I'm not sure what exactly you mean by "Why do we think the mind is a machine?"

But I'll give an explanation to what the general meaning of the question would be. Hint: The answer is in your question

The human brain can process data from your body and create a so-called experience of consciousness, it works just like a man-made machine at a fundamental level.

It would be extremely arrogant to assume otherwise as we know the stuff that the brain is made up of. It's not magical. The brain is a biological computer which has become what we see through the long process of evolution. While it may be hard to believe how such intelligence can come out of nature alone, it may get easier to understand that this occurred through trial and error over a span of a couple of billion years.

The functioning & structure of brain is much different than Human made computers which may be put up as an argument as to why Human-level intelligence isn't so easy to create.

Why simulations are different? Simulations build upon what we do know. For example: we can't simulate back the universe in reverse to find out how the Big Bang happened. Similarly, it's hard to simulate the creation of human-level intelligence on a computer as the biological processes are complex and we don't have full knowledge.

It's true that generative AI can produce effective design through evolution over a average span of a few weeks or months. Although, you can't really compete the billions of years of the universe with some months on a human-made tin can. Nature and computers are different, we have to be specific but the nature doesn't! If we want to consider everything possible outcome, we would have to simulate the universe although that's the whole different challenge, even there we don't have the starting Steps.

We can get around simulations by just going a different way and making a CS version of a brain(not a biological simulation), something to which Neural networks can be considered baby steps. Now we divert towards the technical issues regarding Super AI development.

| improve this answer | |
  • 1
    $\begingroup$ Trial and error over billions of years makes intelligence even more unexpected. Plus, we can evolve things much more quickly on a computer, and nothing in evolutionary algorithms has even come close to human intelligence. $\endgroup$ – yters Jul 6 '17 at 19:05
  • $\begingroup$ @yters How does trial and error over billions of years makes intelligence even more unexpected? Could you explain that? Plus, this basically depends on the algorithm itself. Generally, Simulations are limited to what knowledge is provided to the computer. They are therefore used in predicting what could happen according to our knowledge bank. A simulation to repeat the evolutionary process of biology is extremely hard as we don't have the full knowledge bank on how it happened. $\endgroup$ – Alpha Mineron Jul 7 '17 at 1:45
  • $\begingroup$ @yters Also read the edit $\endgroup$ – Alpha Mineron Jul 7 '17 at 1:57
  • $\begingroup$ “But then with me the horrid doubt always arises whether the convictions of man's mind, which has been developed from the mind of the lower animals, are of any value or at all trustworthy. Would any one trust in the convictions of a monkey's mind, if there are any convictions in such a mind? [To William Graham 3 July 1881]” -- Darwin $\endgroup$ – yters Jul 11 '17 at 1:43
  • $\begingroup$ @yters Please be clear here, are you a strong believer in Religious fables? Don't waste the community's time. Science and Mathematics isn't developed by Human mind, Religions are. In fact, Science is based upon evidence and mathematics. Mathematics is the language of the Universe, and I mean this in every way possible, which builds upon some simple postulates based on observations and builds on it, complex structures to describe the beautiful Nature. Science & maths are a sign of intelligence, as they are powerful tools to learn about nature. NOTHING IS MANMADE, IT'S ONLY NATURE $\endgroup$ – Alpha Mineron Jul 11 '17 at 10:43

Not the answer you're looking for? Browse other questions tagged or ask your own question.