This looks like a fascinating article that I will have to come back to. I'm not sure how well it fits the blog's tagline, "Making deep learning accessible." ...
Maybe if you read the title of his blog as "Making deep learning understandable" and this blog post as "If you understand how deep learning really works (and not just treat it as a magic black box) and understand how the brain works you will realize we are nowhere near the singularity" it makes perfect sense.
The math this guy performs is so questionable and his idea that Deep Learning is the pinnacle of biologically plausible cognitive modelling is incorrect. I write about this [in a blog post](https://medium.com/@seanaubin/deep-learning-is-almost-the-br...).
tl;dr doing pure Deep Learning (also known as connectionist) models of the brain limits your tools in a bad way. Using tools from Dynamicism and Symbolicism is better. As proof, check out Spaun, the world's largest functioning brain model.
Note: I mostly just disagree with him philosophically, in terms of his reasoning because it's overlooking some evidence. Don't really have an option on his conclusion. Probably agree with him more than I disagree with him.
Long, and fascinating. I can see why this made the front page. I oscillated between 20% strident disagreement and 60% strident agreement. (The rest -- no strong feelings, or I feel that the questions are too ill-formed to answer.)
To pick one point of disagreement,
“We do not need as much computational power as the brain has, because our algorithms are (will be) better than that of the brain.”
I hope you can see after the descriptions in this blog post that this statement is rather arrogant.
Machines are already better than brains at many cognitive tasks of practical interest. Believing that we'll continue to find "tricks" to allow computers to outperform brains on useful cognitive tasks, despite the brains' much greater complexity, seems like a perfectly sober and conservative prediction.
If I had to advance my own pet reasons for discounting the likelihood of a technological singularity, here are my top two:
1) It's a more challenging case of the general Fermi paradox. Show me the Hubble images of the computronium Dyson swarms. If it takes less than a century to go from the first transistorized computers to superintelligence, and superintelligence is as prone to run amok as Bostrom/Yudkowsky think, signs should already be visible from Earth.
2) You need experiments to validate scientific models. Even if a machine-intelligence could think a billion times faster than a biological intelligence, it couldn't complete experiments a billion times faster. Technologies that act on the material world will improve sublinearly with respect to thinking/computing power, for at least this reason and probably others as well.
I agree with you. Our current results are impressive enough for me to believe we're not that far from AGI (of the kind that would satisfy the author here). This article is already largely outdated because of advances in Reinforcement Learning and other Deep Learning applications that show very strong primordial features of what he claims as epitomes of human capability. For example, trained RL agents will recognize objects without being able to label them, having internal representations for position, object "type", functionality and relevance. You just have to train the agent with a problem that requires this kind of comprehension for efficient performance (kind of comprehension which isn't needed in simple image labeling).
AlphaGo and even labeling are indeed milestones in superhuman performance, and I believe Terence Tao's view that AI is a moving target is relevant here.
---
1) Regarding the Fermi paradox, I haven't read too much technical or scientific work on it (it's highly speculative anyway), but I find Isaac Arthur's videos lay the basic ideas pretty well: https://www.youtube.com/watch?v=oIva_60l3ww&t=1016s
It becomes pretty convincing that technological intelligent life is an incredible coincidence. There might not be any huge "filters", as they are called (stages which reduce the probability of observing TIL), such as a superintelligence extinction event, but an enormous sequence of minor filters, ranging from low probability of a planet with adequate conditions, to low probability of actual technological development, to low probability of occurrence in our light cone (we can only see fairly young distant galaxies), and limitations to the visibility and spreading velocity of such civilizations.
There is a wide range of parameters such that this does not contradict the generalized Mediocrity principle, such that there is probably more than one TIL in the universe, but they're few and far from each other in space and time.
2) I don't follow. This objection is only valid for discovering new laws of physics -- technological development can happen even with fixed knowledge of basic physical laws. At this stage it's not clear that even continued refinement of physical models. For example, the discovery of quarks certainly helped understanding nature, but it hasn't led to any direct technological applications due to quark confinement, and the fact that the particles are just too small. Neutrinos are another conceptually important discovery that doesn't really have applications due to low interactivity with matter.
This objection is only valid for discovering new laws of physics -- technological development can happen even with fixed knowledge of basic physical laws.
No, experimentation is also important for translating fundamental scientific knowledge into working technologies. Consider the problem of building a fusion reactor that generates net electricity. You don't need new laws of physics, but you can't make progress with simulations alone. You have to experiment to find out if devices behave in reality the way they do in your model.
Well as simulations improve that tends to be less of a problem. Engineering by now is highly reliant on simulations, building actual prototypes is only done as a last validating step as you mentioned (specially for something as expensive and complicated as a fusion reactor). So validation devices are a O(1) step in the innovation chain, I don't see a scaling issue here. You can also simply make arbitrarily many experimentation/testing devices.
I think a greater issue is the general misconception of intelligence as something of a magical attribute, and even more so "superintelligence", and even more so "recursive-self-improving superintelligence" (RSIS for short?). There are not only various limits to hardware, some of which discussed in the article (size of atoms, capacity of communication between parts), but also there are limits to software. For example, while it seems that an RSIS could really solve any problem it wanted, Turing's almost century-old Halting Problem solution already proved that no such computer exists. There's no algorithm that takes an arbitrary conjecture, e.g. the Riemann hypothesis, and outputs either {yes,no,malformed_problem}. In fact the general expectation of what AGIs can do (think creatively, solve arbitrary problems, replace mathematicians, replace programmers) are things that we ourselves can't really do. We solve problems using a small set of heuristics and lots of trial and error, with no success guarantee.
Another fact is that there are many asymptotically optimal algorithms that are already know, and the trivial fact that almost all tasks performed by this super-intelligence (many of which are ordinary tasks we already do, analogous to maybe sorting, database queries, and search engines) will just improve constants on near-optimal algorithms.
If you want to make something intelligent, then you got to give it free hand to run experiments, to check if the correlations it detects are causal or not. I think that is the link between pattern recognition models from current day and AI. So the system must be made of an experimenter (implemented as a RL agent) and a lab (implemented as a simulator) - basically the same setup we use as humans to advance science.
Current day philosophy is like current day ML - detects correlations but can't run experiments to filter out the bad ones, so they (philosophers) are stuck with a combinatorial explosion of theories. They key component missing is a good enough simulator, both at physical and abstract level.
>> “We do not need as much computational power as the brain has, because our algorithms are (will be) better than that of the brain.”
> I hope you can see after the descriptions in this blog post that this statement is rather arrogant.
The brain does more than computers: the brain builds its internal structure all by itself. When was a computer able to evolve from a bunch of transistors laying on the table? Also, it uses very little energy and is resilient for 80-100 years, compared to the computers of today, its apples and oranges.
> Show me the Hubble images of the computronium Dyson swarms. If it takes less than a century to go from the first transistorized computers to superintelligence, and superintelligence is as prone to run amok as Bostrom/Yudkowsky think, signs should already be visible from Earth.
What if the AGI will prefer to build virtual worlds and societies of virtual agents instead of grand space domination ? Even humans prefer games to reality nowadays (or a large percentage of us do). If we're part of such a sim, then it would explain the lack of external signals from extraterrestrial aliens or AGIs.
The probability that AGI will appear and create amazing simulations is much larger than that of picking up signs of life in the vastness of space. Also, by replacing physical with sim we can do all sorts of things such as use less energy, recover from any accident, hack our own brains/minds, even immortality.
>the brain builds its internal structure all by itself
>when was a computer able to evolve from a bunch of transistors laying on the table
These two statements are not really connected. Brain builds its internal structure in a sense, it changes with learning for example. But brain did not evolve by itself. It's all thanks to the force of evolution.
So it boils down to the question: "Can we do better than evolution?", so we need to understand what evolution is. It's just random mutations that make organism better than it was before. If these random mutations make it worse, organism gets lower chances of survival and usually just dies. Essentially, evolution is a method of design by brute force. And brute force works, if you have a lot of time, but it's common sense that intelligently designed algorithm can do a lot better. So we can answer the question now with "Yes, most likely we can do better."
If intelligence is not computable then by definition deep learning is not capable of human intelligence. There are any reasons to think intelligence is beyond computation: Gödel's incompleteness theorems, no free lunch theorem, data processing inequality, Solomonoff induction and Kolmorov complexity is uncomputable, the halting problem, if the mind is code which program are you, all programs are finite yet we can think about infinity, split brain but unified consciousness, the inherent difference between third person and first person descriptions, reasoning about paradoxes, the ability to know we are wrong, the ability to write AI programs, the whole connectionist vs. modules problem Fodor points out, and probably many more.
I think the major problem is to think of "intelligence" as a problem in "computation".
We are so used to this framing, that it may seem foreign. Others will argue, that in principle it is a problem of computability but in extreme almost every problem is, but that is not necessarily how we frame other problems. I'd say "intelligence" is a control problem (as in controlling a robot). This framing, though subtle, makes the entire problem quite different. You no longer talk about computability but you talk about survival in "high temperature thermal bath" (or otherwise called "physical reality"), full of unpredictability and dangerous stuff.
When you frame it like this, it is clear we have not even began to address the problem properly, not to even mention solving it.
Aren't other problems getting framed that way, too? Car driving is not inherently a problem in computation, but is becoming one. Robotic control (or other control problems) are certainly being solved by computation.
Perhaps treating AI as a derivative of the control branch of computation could practically help speed up progress in some areas, but it shouldn't fundamentally change its nature.
Well this is exactly my point. A lot of problems are currently framed as computation problem. As much as it might be useful for some, I'd argue that we should be careful with this.
We get all the great marvels of deep learning, but robots remain stupid as bricks over 30 years of Moore's law. To me this (Moravec's paradox) is a signal that we are doing something wrong, and typically we do things wrong when they are not framed properly.
Ironically, the frame problem is one of those things AI cannot solve. The myopic belief in computational intelligence is a result of our materialism, which in turn is a result of our unwillingness to acknowledge God and live according to His will, which is ultimately best for us.
This is Penrose's argument. One issue with this line of thinking is that it confuses a system which produces infallible proofs and the human brain. None of those theorems actually apply to people; we may gain quite a bit by being wrong sometimes.
We aren't infallible, but that is different than the inconsistency of a formal system. Insistency means the very concept of truth is nonsensical, since an inconsistent system can prove everything. It is fallacious to equivocate the two.
Nobody has ever established that human intelligence is not computable. Human intelligence can't solve any of the well-defined problems you mention, or transcend the limitations of any of those theorems. Don't mistake the idea of an idealized perfect intelligence with the very limited and finite ones that we actually possess. Asking what God could do always leads to madness.
Some of what you mention are what I consider non-well-defined philosophical problems, which really have no bearing on the ability to create algorithms that could pass the Turing test and are outside of the scope of a discussion on the practicality of creating AGI.
No one has ever established human intelligence is computable. A number of the problems I mention just have no computable corollary. E.g. As far as computers are concerned there is no such thing as truth. For us humans to make that claim is inherently nonsensical (I.e. Is it true there is no truth?)
The default position here is that human intelligence is computable, since we have some 7.5 billion biological computers clearly implementing human intelligence. Arguments to the contrary typically rest on a belief in mind-body dualism (popularly disguised as "quantum effects" these days).
I see from another post of yours that you are religious, which is a common personal reason to assume a form of dualism as at least central tenets in the Abrahamic religions seem to require it. Do you see dualism necessary for your position of human intelligence being incomputable, and if so could you reference an argument you would find compelling also for readers whose religion (or lack thereof) does not necessitate dualism?
If you don't see dualism as necessary for the possibility of human intelligence being incomputable, could you open up your thinking on this?
I apologize for being so forward, but I worry some might read your comments without realizing that they might rely on unshared assumptions.
Yes, of course dualism is necessary. You have the causality wrong. I used to be a computational mind person an believe in monism. But the listed reasons have convinced me otherwise. Essentially computation cannot create information. So, if information exists then so does at least one uncomputable process, which obviously implies dualism. Same direction of causality for my religiosity. Since computationalism is untenable, then religion must be true, not visa versa for me.
Please at least read the Wikipedia articles on computability (https://en.wikipedia.org/wiki/Computability) and hypercomputation (https://en.wikipedia.org/wiki/Hypercomputation) before making arguments like this. There is literally zero evidence that anything in our known universe can perform hypercomputation - and that includes all quantum mechanical effects that we've been able to measure. So the Bayesian prior for human intelligence being computable is somewhere near 100%, and it would take very, very extreme evidence to contradict that.
"Is it true there is no truth?" is not even close to a compelling argument, there are rigorously defined notions of truth within any of the eminently computable logical frameworks that we use.
Humans seem to perform hyper computation all the time. Eg in programming we can decide a whole lot of programs halt. We seem to do something close to Solomonoff induction. Etc. computational presuppositions are the only reason to not find this compelling, and materialism and hard atheism are the only reasons to be committed to computationalism.
No, that's a misunderstanding of what hypercomputation is.
If you can write down or talk yourself through a proof of why a program will halt, you have solved a computable problem. By definition. When you say "we can decide a whole lot of programs halt", you're talking about plain Jane computation.
If you have knowledge that any program that I give you will or will not halt, then you're a hypercomputing oracle for the halting problem, and you can do magic, basically.
There's a world of difference between the two situations. And I'm fairly sure (but not positive) that there's no way for you to prove in finite space and time that you are an oracle for the halting problem.
I don't want to get into a philosophical debate here, but please don't overstate the meaning of mathematical theorems.
For example, Gödel's incompleteness theorem is a technical result stating that certain definitions of "model theoretic truth" in classical set theory are incompatible with certain other notions of "truth as provability". Both the statement and its consequences have been so massively oversold for most of the past century that you should never use it in a discussion - think of it as a logical version of Godwin's law.
Similarly, no free lunch type theorems are formally the same as the statement that you cannot compress all n byte sequences into less than n bytes, which is really obvious for cardinality reasons. Again, there is no magic, just clever reductions.
The argument behind the halting problem also applies to your brain and anything that is somehow an abstract model of computation. More fundamentally, the fact that the self halting problem is undecidable is simply an instance of Cantor's theorem, it's not something that can be avoided.
Mathematical logic and therefore computers can be used to talk about infinity and more. Logical "paradoxes" are not a problem either. Some may be genuine proofs of inconsistency of some logical theories and others are simply theorems. The usage of the word "paradox" in natural language is simply imprecise.
---
I could go on, but really I don't take offense at any particular point. What bothers me is that you seem to be overselling mathematical results to argue a non-mathematical point... If you really want to apply, e.g., Gödel type theorems to discussions about your brain, you would first have to argue that the assumptions of Gödel's theorems apply. For instance, you could argue that the definition of model based truth in Peano arithmetic is something that your brain can decide. Then it would follow that what your brain does is uncomputable.
tldr: 2037-2080 for "brainlike computers". sounds like a reasonable estimate to me. but who would say that is "nowhere near"? A few decades compared to billions of years of evolution to create life here on earth? a few decades compared to millions of years to go from monkeys to humanity? compared to thousands of years to go from prehistory to contemplating this question? a date that many of us will live to see? IMO this means we are on the verge.
Since we're already at the level of coding genes, we are short-circuiting the normal process of evolution, which was a trial-and-error process. We are living in interesting times.
Slightly more in detail was his approximation that brain could perform 10^21 FLOPS vs 10^15 currently.
But yeah 2080 is within my daughter's lifetime. My 1950s house I bought will probably still be standing. I think it also coincides with Elon Musk's guess for when there'd be 1M people living on Mars.
The key to the claim that other computational estimates are way off is essentially that there's a lot of data crunching happening within a single neuron, rather than it being something we can model as collecting a bunch of inputs and either firing or not. He's arguing that each neuron does a ton of internal computation that can itself be modeled as a (sometimes very large) convolutional network.
I think in a sense this is well-established when looking at real neurons, but several times in the article he uses phrases like "shown to be important for information processing", and that's where I get off the boat a bit. When you're saying that it's so important for information processing that it warrants a 1000x or more increase in the computational power necessary to implement an algorithm, I think it's necessary to dig into what the actual work being done there is, not just that there's some non-trivial transformation. A lot of interesting and extremely tough to model fluid and chemical dynamics are in play when I drink too much water and have to pee, but that doesn't mean that we need to understand them to build a waste disposal system using pipes.
In particular, does the within-neuron processing actively tune itself based on the data it processes to an extent on-par with inter-neuron connections (in which case the argument that it's fundamental to the learning process would hold a lot more weight), or is it mostly static? I think a lot of us consider "important for information processing" to mean "is a meaningfully dynamic parameter involved in a learning algorithm", rather than an accidental shmearing of data.
I'd really love more info on what the actual processing that's happening is.
That's basically what the entire field of computational/theoretical neuroscience is trying to figure out right now. What's being computed? What's the right level of abstraction? Basically, it depends on what you're modelling: https://www.ncbi.nlm.nih.gov/pubmed/24709593
I worked in this field a bit at Baylor. There are those who want to model the 'hardware' down to every complexity and those who want to figure out the 'software' that emerges from the physical hardware. The groups communicate but are fairly divided.
Does it tune itself or is static? It definitely tunes, but it can be meaningful to the algorithm or ignored.
Inside any cell there is a system that works like a neural network - the gene regulatory network. Each gene acts like a neuron, with chemical inputs and outputs. That would make the processing power of any cell on par with that of a small neural net.
Gene regulation is very slow, on the order of minutes to hours. It's involved in our brains, for sure, but it's just way too slow to be a dynamic part of our brains' processing of information.
We should distinguish between the information processing of the machinery of the brain and the information processing of the mind, even though the issue of what makes up the latter is an open question. It would not be a useful measure of a program's complexity to simulate the hardware running the program, and calculate the complexity of that simulation (even if we didn't follow the implicit infinite recursion in that approach...)
My non-specialist take on it is that, even if what neurons really do is hugely complicated, we should be able to approximate it with something pretty simple and anything sorta close should work. I believe this because real brains are incredibly fault tolerant and can take an absolute beating while still continuing to function 'well enough'.
I'd be willing to bet that once we understand the basic functioning, we'll be able to build a working brain using a few different types of neuron which operate on relatively simple rules-of-thumb, arranged in a few different basic structures which are internally relatively homogeneous.
I think I tend to agree (and actually, the author does, as well), the interesting question is whether we've been underestimating the number of artificial "neurons" we'll need. If each biological neuron is actually essentially implementing an entire neural net internally, then we might be a lot further from achieving that amount of computation than we expected.
We absolutely need to start doing what you're suggesting, and figuring out how to derive useful basic structures, represent them, and construct networks using those as the building blocks. My prediction is that's where the field will go over the next 5-10 years, a much deeper dive into how to specify connectivity and usefully control it than has been done so far (we're still using fully connected layers, for the most part, which we know is not a scalable approach as we go from thousands to millions of nodes).
>Quantum tunneling will become relevant in 2016-2017 and has to be taken into account from there on. New materials and "insulated" circuits are required to make everything work from here on.
He wrote this in 2015. Was he correct? I don't really know where to start researching something like this; is there anyone here familiar with the field who could comment on it?
I think he's referring to the general idea that transistor fabrication size shrinking cannot last forever. Since atoms have a fixed physical size, if the shrinking continues indefinitely, at some point the "wires" in the transistors will become zero-atoms thick (actually a few atoms thick). See this for explanations better than I could https://en.wikipedia.org/wiki/5_nanometer
Note the "quantum" here is not a good thing (like the theoretical "quantum speedup" possible for certain computations on quantum computers), but a bad thing: imagine you want to send current down this wire, but the current is jumpy and often leaks out of the wire to a neighbouring wire thus causing errors in computations.
An interesting read, but the conclusion that "the singularity is nowhere near" was reached by assuming that only neural modeling could get us there, and that assumption wasn't defended well. (In fact it looks rather dubious, given all the quasi-intelligent things computers have achieved without copying neural dynamics.)
48 comments
[ 2.3 ms ] story [ 104 ms ] threadtl;dr doing pure Deep Learning (also known as connectionist) models of the brain limits your tools in a bad way. Using tools from Dynamicism and Symbolicism is better. As proof, check out Spaun, the world's largest functioning brain model.
Note: I mostly just disagree with him philosophically, in terms of his reasoning because it's overlooking some evidence. Don't really have an option on his conclusion. Probably agree with him more than I disagree with him.
To pick one point of disagreement,
“We do not need as much computational power as the brain has, because our algorithms are (will be) better than that of the brain.”
I hope you can see after the descriptions in this blog post that this statement is rather arrogant.
Machines are already better than brains at many cognitive tasks of practical interest. Believing that we'll continue to find "tricks" to allow computers to outperform brains on useful cognitive tasks, despite the brains' much greater complexity, seems like a perfectly sober and conservative prediction.
If I had to advance my own pet reasons for discounting the likelihood of a technological singularity, here are my top two:
1) It's a more challenging case of the general Fermi paradox. Show me the Hubble images of the computronium Dyson swarms. If it takes less than a century to go from the first transistorized computers to superintelligence, and superintelligence is as prone to run amok as Bostrom/Yudkowsky think, signs should already be visible from Earth.
2) You need experiments to validate scientific models. Even if a machine-intelligence could think a billion times faster than a biological intelligence, it couldn't complete experiments a billion times faster. Technologies that act on the material world will improve sublinearly with respect to thinking/computing power, for at least this reason and probably others as well.
AlphaGo and even labeling are indeed milestones in superhuman performance, and I believe Terence Tao's view that AI is a moving target is relevant here.
---
1) Regarding the Fermi paradox, I haven't read too much technical or scientific work on it (it's highly speculative anyway), but I find Isaac Arthur's videos lay the basic ideas pretty well: https://www.youtube.com/watch?v=oIva_60l3ww&t=1016s
It becomes pretty convincing that technological intelligent life is an incredible coincidence. There might not be any huge "filters", as they are called (stages which reduce the probability of observing TIL), such as a superintelligence extinction event, but an enormous sequence of minor filters, ranging from low probability of a planet with adequate conditions, to low probability of actual technological development, to low probability of occurrence in our light cone (we can only see fairly young distant galaxies), and limitations to the visibility and spreading velocity of such civilizations.
There is a wide range of parameters such that this does not contradict the generalized Mediocrity principle, such that there is probably more than one TIL in the universe, but they're few and far from each other in space and time.
2) I don't follow. This objection is only valid for discovering new laws of physics -- technological development can happen even with fixed knowledge of basic physical laws. At this stage it's not clear that even continued refinement of physical models. For example, the discovery of quarks certainly helped understanding nature, but it hasn't led to any direct technological applications due to quark confinement, and the fact that the particles are just too small. Neutrinos are another conceptually important discovery that doesn't really have applications due to low interactivity with matter.
No, experimentation is also important for translating fundamental scientific knowledge into working technologies. Consider the problem of building a fusion reactor that generates net electricity. You don't need new laws of physics, but you can't make progress with simulations alone. You have to experiment to find out if devices behave in reality the way they do in your model.
I think a greater issue is the general misconception of intelligence as something of a magical attribute, and even more so "superintelligence", and even more so "recursive-self-improving superintelligence" (RSIS for short?). There are not only various limits to hardware, some of which discussed in the article (size of atoms, capacity of communication between parts), but also there are limits to software. For example, while it seems that an RSIS could really solve any problem it wanted, Turing's almost century-old Halting Problem solution already proved that no such computer exists. There's no algorithm that takes an arbitrary conjecture, e.g. the Riemann hypothesis, and outputs either {yes,no,malformed_problem}. In fact the general expectation of what AGIs can do (think creatively, solve arbitrary problems, replace mathematicians, replace programmers) are things that we ourselves can't really do. We solve problems using a small set of heuristics and lots of trial and error, with no success guarantee.
Another fact is that there are many asymptotically optimal algorithms that are already know, and the trivial fact that almost all tasks performed by this super-intelligence (many of which are ordinary tasks we already do, analogous to maybe sorting, database queries, and search engines) will just improve constants on near-optimal algorithms.
A god-like creature those AIs will be not.
Current day philosophy is like current day ML - detects correlations but can't run experiments to filter out the bad ones, so they (philosophers) are stuck with a combinatorial explosion of theories. They key component missing is a good enough simulator, both at physical and abstract level.
Yann LeCun saying the main problem in AI is how to make a predictive model of the world (a simulator): https://youtu.be/cWzi38-vDbE?t=1933
> I hope you can see after the descriptions in this blog post that this statement is rather arrogant.
The brain does more than computers: the brain builds its internal structure all by itself. When was a computer able to evolve from a bunch of transistors laying on the table? Also, it uses very little energy and is resilient for 80-100 years, compared to the computers of today, its apples and oranges.
> Show me the Hubble images of the computronium Dyson swarms. If it takes less than a century to go from the first transistorized computers to superintelligence, and superintelligence is as prone to run amok as Bostrom/Yudkowsky think, signs should already be visible from Earth.
What if the AGI will prefer to build virtual worlds and societies of virtual agents instead of grand space domination ? Even humans prefer games to reality nowadays (or a large percentage of us do). If we're part of such a sim, then it would explain the lack of external signals from extraterrestrial aliens or AGIs.
The probability that AGI will appear and create amazing simulations is much larger than that of picking up signs of life in the vastness of space. Also, by replacing physical with sim we can do all sorts of things such as use less energy, recover from any accident, hack our own brains/minds, even immortality.
This is against the 13W that the brain uses: https://www.scientificamerican.com/article/thinking-hard-cal...
So the brain has a million times more the processing power and uses a million times less power.
A commodity smartphone is millions of times more powerful than all of NASA's combined computing in 1970 and more energy efficient.
A factor of 10^6 is maybe achievable with a specialized quantum computer in the mid-term future ?
>when was a computer able to evolve from a bunch of transistors laying on the table
These two statements are not really connected. Brain builds its internal structure in a sense, it changes with learning for example. But brain did not evolve by itself. It's all thanks to the force of evolution.
So it boils down to the question: "Can we do better than evolution?", so we need to understand what evolution is. It's just random mutations that make organism better than it was before. If these random mutations make it worse, organism gets lower chances of survival and usually just dies. Essentially, evolution is a method of design by brute force. And brute force works, if you have a lot of time, but it's common sense that intelligently designed algorithm can do a lot better. So we can answer the question now with "Yes, most likely we can do better."
We are so used to this framing, that it may seem foreign. Others will argue, that in principle it is a problem of computability but in extreme almost every problem is, but that is not necessarily how we frame other problems. I'd say "intelligence" is a control problem (as in controlling a robot). This framing, though subtle, makes the entire problem quite different. You no longer talk about computability but you talk about survival in "high temperature thermal bath" (or otherwise called "physical reality"), full of unpredictability and dangerous stuff.
When you frame it like this, it is clear we have not even began to address the problem properly, not to even mention solving it.
Perhaps treating AI as a derivative of the control branch of computation could practically help speed up progress in some areas, but it shouldn't fundamentally change its nature.
We get all the great marvels of deep learning, but robots remain stupid as bricks over 30 years of Moore's law. To me this (Moravec's paradox) is a signal that we are doing something wrong, and typically we do things wrong when they are not framed properly.
Some of what you mention are what I consider non-well-defined philosophical problems, which really have no bearing on the ability to create algorithms that could pass the Turing test and are outside of the scope of a discussion on the practicality of creating AGI.
I see from another post of yours that you are religious, which is a common personal reason to assume a form of dualism as at least central tenets in the Abrahamic religions seem to require it. Do you see dualism necessary for your position of human intelligence being incomputable, and if so could you reference an argument you would find compelling also for readers whose religion (or lack thereof) does not necessitate dualism?
If you don't see dualism as necessary for the possibility of human intelligence being incomputable, could you open up your thinking on this?
I apologize for being so forward, but I worry some might read your comments without realizing that they might rely on unshared assumptions.
"Is it true there is no truth?" is not even close to a compelling argument, there are rigorously defined notions of truth within any of the eminently computable logical frameworks that we use.
If you can write down or talk yourself through a proof of why a program will halt, you have solved a computable problem. By definition. When you say "we can decide a whole lot of programs halt", you're talking about plain Jane computation.
If you have knowledge that any program that I give you will or will not halt, then you're a hypercomputing oracle for the halting problem, and you can do magic, basically.
There's a world of difference between the two situations. And I'm fairly sure (but not positive) that there's no way for you to prove in finite space and time that you are an oracle for the halting problem.
Edit: for more on that last bit, see https://pdfs.semanticscholar.org/19f9/0c34cdda43efdcf0831b2f...
For example, Gödel's incompleteness theorem is a technical result stating that certain definitions of "model theoretic truth" in classical set theory are incompatible with certain other notions of "truth as provability". Both the statement and its consequences have been so massively oversold for most of the past century that you should never use it in a discussion - think of it as a logical version of Godwin's law.
Similarly, no free lunch type theorems are formally the same as the statement that you cannot compress all n byte sequences into less than n bytes, which is really obvious for cardinality reasons. Again, there is no magic, just clever reductions.
The argument behind the halting problem also applies to your brain and anything that is somehow an abstract model of computation. More fundamentally, the fact that the self halting problem is undecidable is simply an instance of Cantor's theorem, it's not something that can be avoided.
Mathematical logic and therefore computers can be used to talk about infinity and more. Logical "paradoxes" are not a problem either. Some may be genuine proofs of inconsistency of some logical theories and others are simply theorems. The usage of the word "paradox" in natural language is simply imprecise.
---
I could go on, but really I don't take offense at any particular point. What bothers me is that you seem to be overselling mathematical results to argue a non-mathematical point... If you really want to apply, e.g., Gödel type theorems to discussions about your brain, you would first have to argue that the assumptions of Gödel's theorems apply. For instance, you could argue that the definition of model based truth in Peano arithmetic is something that your brain can decide. Then it would follow that what your brain does is uncomputable.
But yeah 2080 is within my daughter's lifetime. My 1950s house I bought will probably still be standing. I think it also coincides with Elon Musk's guess for when there'd be 1M people living on Mars.
I think in a sense this is well-established when looking at real neurons, but several times in the article he uses phrases like "shown to be important for information processing", and that's where I get off the boat a bit. When you're saying that it's so important for information processing that it warrants a 1000x or more increase in the computational power necessary to implement an algorithm, I think it's necessary to dig into what the actual work being done there is, not just that there's some non-trivial transformation. A lot of interesting and extremely tough to model fluid and chemical dynamics are in play when I drink too much water and have to pee, but that doesn't mean that we need to understand them to build a waste disposal system using pipes.
In particular, does the within-neuron processing actively tune itself based on the data it processes to an extent on-par with inter-neuron connections (in which case the argument that it's fundamental to the learning process would hold a lot more weight), or is it mostly static? I think a lot of us consider "important for information processing" to mean "is a meaningfully dynamic parameter involved in a learning algorithm", rather than an accidental shmearing of data.
I'd really love more info on what the actual processing that's happening is.
Does it tune itself or is static? It definitely tunes, but it can be meaningful to the algorithm or ignored.
I'll just leave this here: https://en.wikipedia.org/wiki/Gene_regulatory_network
Inside any cell there is a system that works like a neural network - the gene regulatory network. Each gene acts like a neuron, with chemical inputs and outputs. That would make the processing power of any cell on par with that of a small neural net.
I'd be willing to bet that once we understand the basic functioning, we'll be able to build a working brain using a few different types of neuron which operate on relatively simple rules-of-thumb, arranged in a few different basic structures which are internally relatively homogeneous.
We absolutely need to start doing what you're suggesting, and figuring out how to derive useful basic structures, represent them, and construct networks using those as the building blocks. My prediction is that's where the field will go over the next 5-10 years, a much deeper dive into how to specify connectivity and usefully control it than has been done so far (we're still using fully connected layers, for the most part, which we know is not a scalable approach as we go from thousands to millions of nodes).
He wrote this in 2015. Was he correct? I don't really know where to start researching something like this; is there anyone here familiar with the field who could comment on it?
Note the "quantum" here is not a good thing (like the theoretical "quantum speedup" possible for certain computations on quantum computers), but a bad thing: imagine you want to send current down this wire, but the current is jumpy and often leaks out of the wire to a neighbouring wire thus causing errors in computations.