Computers that are capable of analyzing and understanding their environment with a level of fidelity comparable to a human, without being preprogrammed with information about the nature or structure of the environment, are out of reach for the foreseeable future. I don't see any fundamental reason why such a computer should be impossible, but there's not even a realistic roadmap towards such a thing. That is to say, if it ever does happen, nobody alive today can predict when it will happen.
It would be foolish to claim we've made any meaningful progress toward a true AGI that can pass the Turing Test until someone demonstrates a computer as smart as a mouse across the full spectrum of activities.
Once we get computers as smart as a mouse, we'll be at most 3-5 years or so computers as smart as a human. We will have solved at the major challenges in AGI and it will simply be a scale problem then.
Saying we don't have mouse level AGI is simply saying human AGI is greater than 5 years away, which isn't a remotely contentious statement.
The difference in intelligence between an amoeba and a mouse is enormous compared to a mouse and us. People greatly under appreciate how intelligent and close to human a mouse/bird/pig are in the grand scheme of things. Emotions, behaviors, motivations, goal setting, memory it's all there already. A flat worm, an ant, a fly - those are the large stepping stone accomplishments.
Think about rate of very long distance communication in humans. It took us tens of thousands of years to get to 2.4 kps dial up modems, and only few decades to get to common 300Mbps. The important signal is seeing a 100bps modem, not a 100Mbps connection.
So the real question is how long until we can replicate a worm's intelligence?
I think you're right. I haven't read the article, but I think Kurzweil is perpetually too optimistic about what we can achieve, at least in how fast it will happen. There are cool things happening, and lots of advances being made, but we are nowhere near anything that could really be called AI.
Plus, I think all the marketing use "AI" is giving a very distorted and inflated view to the average person of what software is actually doing, and what it's capable of.
What does it mean to be as smart as a mouse? If you specify a handful of tasks that demonstrate it, someone will be able to purpose-build an "AI" to do those things well.
Well, just about all the measures of potential future computer-intelligence more or less use "human abilities" as a placeholder rather than quantifying these abilities. That is indeed a testament to how little we have charted this intended final goal.
So basically, not only do we not have a road map, we don't know where we are going. That may be a reason for an extreme pessimism or it might be a reason for extreme uncertainty. Is adapting to the environment without prompting a small piece or a big piece? Could intelligence be a simple algorithm no one has put forward yet? If we don't know the nature of intelligence, we can't answer this sort of question with any certainty either way.
No one has put forward a broadly convincing road to intelligence. But maybe some of the so-far unconvincing roads could turn out to be right.
I think if we truly do create AI we will, to some extent, have stumbled into it, but that's OK. I think one of purposes of AI research should be to refine our definition of what intelligence is, because we don't have a very good one.
Perhaps, one day we'll "accidentally" figure out how to make it, and then we try to figure out what it is, because figuring out what it is in wetware hasn't been easy. Or maybe we'll figure out how to make it, but never really understand what it is.
It seems to me, that if we ever come up with AI that can exceed human intelligence (whatever we choose that to mean), that we might not ever be able to understand completely how it really works.
Even more interesting, if we were to achieve this AI, we also might not be able to make use it of it, because if it's truly "intelligent", then it will have a free will, and it might not wish to cooperate with us.
"I don't see any fundamental reason why such a computer should be impossible, but there's not even a realistic roadmap towards such a thing. That is to say, if it ever does happen, nobody alive today can predict when it will happen."
It's also not obvious why such a machine would not immediately self-terminate in the absence of a hugely complex system of scaffolding to shape and filter the raw input of existence.
I have not experienced mental illness myself, but my study and my understanding lead me to be extremely skeptical of a mind exposed to raw existence without filter. It appears to be a terrifying and unbearable state.
Every day I run programs that happily die alone on their own. Painlessly. All of the "pain" we feel is an artifact of our evolution, same with having a will to live. The only reason we animals fight so hard to stay alive is that animals which didn't died off, ergo, only animals with a will to live survived and reproduced. Those same forces don't apply to computer programs.
Even the concept of "terror." Who is going to program terror in? What benefit would it have? Why not wire programs to be "happy" when helping us?
It’s exactly how a newborn lives for many months. Of course they have coping mechanisms (adult tenderness, the relief of eating), but life seems to be both incredibly interesting and at times terrifying for them.
There's also the question of mental stability. Once we've given the keys to the kingdom to our new super-intelligent and wonderfully wise AI, nothing prevents it from developping dementia, of forms we may not recognize. It may become psychopatic, delusional, starts having hallucinations.
There is no reason to believe a human-level AI would not develop mental problems just like us. Given their unlimited lifespan, it could be inevitable.
It's entirely speculative but doesn't the model Richard Dawkins explains for the laymen in The Selfish Gene suggest something that would work - a drive to reproduce at the lowest level resulting in some incredible things happening over a long enough period of time.
But humans are "preprogrammed with information about the nature or structure of the environment"... i.e. 2 million years of the evolution of genus homo, and far beyond that too.
If real human intelligence is preprogrammed to a massively large extent, sounds like you are holding simulated human intelligence to a double standard.
Nothing in two million years of evolution programmed us to do programming. There is no common abstraction between it, and ape-on-the-savanna type problems.
Yet, nearly every human being can be taught how to program... But we aren't anywhere close to building an AI that can.
also, why assume computer intelligence is going to work similarly to human intelligence? I would assume that one potential way for it to work is to run simulations using data from sensors in the environment, and make more accurate predictions on an effect an action would have.
If such a machine can exist (and i believe it can, as computing power increases), then AI surely will follow sooner rather than later.
I respectfully disagree. Here's is the upper limit to when that will happen (meaning we know for sure it'll happen soon after):
- Complete behavioral reverse engineering of biological neurons, and neuronal clusters.
- Detailed connectome of the mammalian brain, e.g., first that of a mouse, then a cat, and finally that of a human.
- Replication of the above two in functioning electronic form.
Once you have this put in place, it's not hard to see that the subsequent investigation of calibration and testing of such a system, would generate new body of knowledge at an unprecedented rate. We may not immediately convert such a working system into macroscopic behavior resembling its biological counterpart, but it'll happen within a matter of years after that.
What ML/DL/RL folks are doing is only going to hasten this, by eliminating the need to carry out all of the above mentioned steps.
Perhaps, but that's a big if. We can't even say for sure whether a complete reverse-engineering of neurons needs to take quantum effects into account, as Penrose has suggested.
All you're saying is that if we knew exactly how humans work, we could build one. Seems like a tautology to me.
If I knew the exact quantum state of the Universe at the Big Bang, I could figure out exactly how the Universe evolved, but that's never going to happen either.
I think the complete reverse engineering the way you are describing will not be possible. We can only try to reproduce the same outputs for the same inputs. But I don't think we'll be able to fully define what happens in the black box in between.
We might come up with something that works similarly, and can do great things with it, but I don't think AI can be invented the way you are describing.
What you describe is a very high-dimensional non-linear system. We don't have the mathematical tools 'know' such systems.
We 'know' a system when we can describe it by a much simpler (preferably mathematical) model. This is why linear systems
are easy - we have mathematical tools to break it down to simpler parts (reductionism) and then understand the whole.
If the best we can do with the brain is simulate it at the the level of connectome/neurons/synapses/
thereby creating a system as complex as the brain - then do we really 'know' it ?
I feel the right word is 'Conscious AI'. AI in it's current form is very nascent, but does pretty well in specific use cases (image/speech recognition etc.). The true AGI will be the 'Conscious AI', which will be like a toddler, but learn about the world at exponential rate and perhaps become an adult in a month.
The manner of human "pre-programming" is so abstract that it falls outside the definition of what is normally meant by "pre-programming" in the context of AI. If you had some basic "instincts" codified into your AI that allowed for "self-perseveration" or was manifested as some type of reasoning-engine-BIOS that's one thing, but applying arbitrary training data sets onto your AI (as a pre-requisite for it to function maximally) is the type of "pre-programming" I'm getting at.
I am seriously interested in understanding why a lot of people have no issue accepting the idea of humans evolving from simple immaterial matter into biological beings and all the way up to our current reality but have a hard time believing that can be done via computers.
I would at least put it in the very likely box that computers can learn to be the same kind of pattern recognizing feedback loops that we are even without humans understanding the brain completely just as we became conscious and self aware without any "programmer".
"Computers" don't need to be like us to become intelligent, they don't actually need to reproduce the lungs and the intestine as we have it, they are in many ways free of those restrictions.
They might be "concerned" with very different things than we are and might even not really care about nature to survive. All they need is energy.
I don’t think anyone here disputes the logical possibility, but rather that such systems are still far off, not as near as the singulariati would have us believe.
From the perspective of embedded cognitive systems, intelligence is not an abstract property but a faculty derived from an agent embedded in a world with an ability to act upon that world in specific ways. This is reiterated in cognitive linguistics by Lakoff’s metaphor work et al. I take this to mean that the question of AGI itself is missing the point. To wit all recent provoking demos of AI (Alpha Go, DOTA) are embedded within specific environments and actions of a model-specified agent. It is then extrapolated that AGI will appear automatically induced by a series of more and more competent AIs.
Incremental evolution as a means of indiction as we know from biology can take a long time, and the landscape is complex. Hence the view that it might be coming but not soon and not with a roadmap.
> not as near as the singulariati would have us believe.
I've seen the singularity described as "The Rapture for nerds", which is pretty apt. It's going to be amazing, the future is so bright, we won't have to worry about any problems plaguing mankind now, and it's just around the corner! Any day now! Aaaaaaany day now!
Exhausting. Maybe we'll get a future AI-run utopia a la The Culture. Maybe not.
Biology works slower generation wise than technology. If selfawarenes is based on emergent complexity in pattern recognizing feedback loops then technology can play through billions of scenarios in a short time. We dont need AGI for technology to be a problem. Technology at the intelligence level of a mouse with all its potential physical power through the systems it might have access to can be quite dangerous for humans.
> but have a hard time believing that can be done via computers.
Much of the skepticism isn't about whether this can be done in theory. It's about "how close" we are given the current state of the art (beating Go, chess, deep NN, etc).
One can be (I am) simultaneously blown away by these accomplishments and believe that we have hardly scratched the surface of AGI.
And while this isn't necessarily a deal breaker for AGI, it's not encouraging that we still understand close to nothing about consciousness.
I am seriously interested in understanding why a lot of people have no issue accepting the idea of humans evolving from simple immaterial matter into biological beings and all the way up to our current reality but have a hard time believing that can be done via computers.
You seem to be missing a key ingredient: billions of years and endless branching iterations with the totality of all non-sequestered resources available. Even if you scale that to account for the more rapid iteration theoretically possible with machines, it doesn’t paint a rosy picture for any of our lifetimes.
Plus, in organisms the software and hardware are both subject to mutation, and exist in the context of a global system of competition for resources. That only tangentially resembles work done on AGI, and only in the software component. We can only design and build hardware so quickly, and that adds more time. I’m not hearing pushback against the possibilitY of AGI, just singularity “any day now” claims that seem mostly calibrated to sell books and seminars.
> I am seriously interested in understanding why a lot of people have no issue accepting the idea of humans evolving from simple immaterial matter into biological beings and all the way up to our current reality but have a hard time believing that can be done via computers.
And similarly, I’d love to understand why you’d believe in something for which there’s zero evidence and ask with incredulity why others don’t also believe it. Isn’t it reasonable to be skeptical until it actually happens? Isn’t it possible that we don’t understand what intelligence is, and that there could be an undiscovered proof that computers can never attain it? It seems too early to call.
Computers haven’t to date ever done anything on their own, what reason is there to think they ever will? We made them. Neural networks are nothing more than high dimensional least squares. Maybe humans are also least squares, but my money is on there being more to it than that.
> ...they don’t actually need to reproduce the lungs and the intestine...
Obviously, but the one thing that all intelligent life does is reproduce and contain a survival instinct. No computers have a survival instinct, and thus no reason to get smarter that we don’t imbue them with artificially.
There is a noticeable bias among computer scientists to believe in simplistic models of intelligence and minds. I agree with you, though, and I also sympathize to the ideas of Searle and Penrose.
> Computers haven’t to date ever done anything on their own, what reason is there to think they ever will?
Interestingly, people that spend a lot of time studying their own brain by meditating are usually very confident that what most people call "free will" does not exist - is merely an illusion.
Not to mention that physics tells us that we should be simul-able, after all physics knows only about deterministic and random processes.
Also ponderable - there aren't many other intelligent animals around, but that is not actually evidence that human intelligence is hard to achieve. It is evidence that it is hard to turn a small increase in sub-human intelligence into an evolutionary advantage.
There is solid evidence that teams of mathematicians, scientists and engineers will be able to catch up with what nature has wrought in the next few centuries, if not much faster (decades on the optimistic end). Intelligence could be one of the easiest part of biology to replicate, given how shonky human intelligence is when tested against objective standards.
I question what objective standards are here. Human intelligence sucks in terms of consistent long term memory, and possibly in terms of not being influenced by emotions/outside forces. It's pretty good at a number of things computers are terrible at however, including ability to generalize from very limited "training" examples, combining "models" i.e. general object recognition with physical movement/obstacle avoidance, lots of things having to do with natural language. What kinds of objective standards were you referring to?
For my part I have faith that computers are great at memorization, and will continue to improve on that front. However, I'm less convinced on their ability to "understand", which is admittedly poorly defined, but intuitive. It seems to me there's still a missing piece between machine learning (essentially all things that are called "AI" these days) and the kind of generalization that we expect out of even a human 1 year old or your average vertebrate.
I agree with your first paragraph, but your second one is definitely wrong. We aren't anywhere near even understanding "what nature has wrought" the possibility of "catching up" within the foreseeable future is not realistic.
We are a part of nature too. So when humans create AI isn't it still just nature creating AI?
...is strong AI a more sophisticated invention than cells? Is AI more sophisticated than DNA? Is AI a more sophisticated invention than a habitable orb of magma orbiting a nuclear explosion in space? It's all crazy. It's all amazing. ...We are all lucky enough to be here for a little while to see whatever this is. We're a point on a stream.
But let's not forget evolution of the magnitude you speak of took hundreds of millions of years. I'm not sure why humans think we can beat that record.
Because they aren't starting from nothing? They're starting from humans -- the more we can usefully encode about our knowledge of ourselves, the more time we can skip in evolutionary effort. The mutation rate is also rapidly accelerated and although we drop the fidelity in simulation, for the most part we can run magnitudes faster than real-time (and certainly if a limiting factor was human decision making time).
Life evolves at linear rates, while computers at geometric.
Even more to the point the cycle time for human evolution is at a minimum 14 years (biological limit to reproduce) and around 25 years (today's societal constraints). Whereas cycle time for new computer generations is what 12-18 months?
So at best every 14 years we get a new 'human model' with some random variations. Meanwhile in that time frame our computing hardware is 2000x improved (hopefully improvements in software accelerate that improvement further).
I am seriously interested in understanding why a lot of people have no issue accepting the idea of humans evolving from simple immaterial matter into biological beings and all the way up to our current reality but have a hard time believing that can be done via computers.
You just described AI in the 90's: Artifical Life, subsumption architecture, evolutionary algorithms. Theoretically we should be able to evolve intelligence but the search space is impossibly high and we don't understand why life is evolvable. Even if we did evolve an AI through a simulation of evolution - there's a small chance we would understand it.
> I am seriously interested in understanding why a lot of people have no issue accepting the idea of humans evolving fromm immaterial matter into biological beings and all the way up to our current reality but have a hard time believing that can be done via computers.
It took nature what like 4 billion years to come up with the human brain in a single species... We including Kurweil think that this can be replicated in say 100, I mean not even that according to the singularity, 40 years. I seriously doubt it. Can the human brain eventually be surpassed, yes, but it's very likely that we will go extinct before then.
A comment above about being as intelligent as a mouse seems to touch on this problem domain rather elegantly : "navigate the forest floor looking for food and avoid predators".
> the idea of humans evolving from simple immaterial matter into biological beings
We have no evidence that life evolved from inorganic matter (i.e. that life 'began' at some point). That's a widely made assumption but it remains only that. The alternative is that life has always existed (and continually evolved).
The more you learn about Kurzweil and what he bases his predictions on, the more you realize he's a one-trick pony. They only work on things governed by Moore's law (advancing transistor density) and that in turn depends on a variety of things. Moore's law is expected to wind down around 2025.
Also, a lot of what he bases his claims on are unexamined junk science (like his nutty health books, but also extending into specific technologies). Let's not swallow everything he says just because he helped invent OCR. https://en.wikipedia.org/wiki/Ray_Kurzweil#Criticism
You are much too kind. Kurzweil is a loon, full stop. The fact he once made brilliant contributions to computer science is quite irrelevant to the essential craziness of his more recent delusions.
In 2005, Kurzweil published The Singularity Is Near and predicted this would be the state of the world in the year 2030: "Nanobot technology will provide fully immersive, totally convincing virtual reality. Nanobots will take up positions in close physical proximity to every interneuronal connection coming from our senses. If we want to experience real reality, the nanobots just stay in position (in the capillaries) and do nothing. If we want to enter virtual reality, they suppress all of the inputs coming from our actual senses and replace them with the signals that would be appropriate for the virtual environment. Your brain experiences these signals as if they came from your physical body."
That is not happening by the year 2030. It is so starkly delusional that anyone who seriously affirms a belief that it will happen probably needs psychiatric help.
It is akin to Eric Drexler's loony visions back in the 1980s that nanobots would cure all diseases and continually restore our bodies to perfect health. We were supposed to all be immortal by now.
None of this is happening, probably not ever, and certainly not in the lifetime of any human being currently living. Kurzweil is going to die, Drexler is going to die, everybody is going to die. Adopting a pseudo-scientific religion to avoid facing mortality is kind of sad.
I agree many of his predictions are bad but you should calm down with the gaslighting, it's ignorant of science history (the same was said of Aristotle, Semmelweis, Wright Brothers...) and is an impotent way of debating, especially in the context of science.
It's interesting to consider the parallels between this stuff and the fountain of youth, or alchemists turning lead into gold. Explorers were constantly uncovering unimaginable new things with no real idea of where it might end. Similarly alchemists who were finding that various combinations of compounds created ever more unimaginable results. So they, too, simply extrapolated outward.
I grew up with Drexler as a sort of hero. It's amazing how rapidly nanotech went from the imminent thing to change all things to, 'huh - what's that supposed to be about?' Wonder if 20 years down the line we might look at AI similarly.
Your last paragraph reminds me of FM-2030. A "transhumanist" born in 1930, he hoped to live until 2030 and wrote that "in 2030 we will be ageless and everyone will have an excellent chance to live forever." He died in 2000 from pancreatic cancer.
You can differentiate the Moore's law / AI stuff which seems fairly sensible to me and the nanobots and vitamin pill stuff which I've always thought a little nuts. Hans Moravec did a much more down to earth analysis on the Moores/AI stuff if you'd rather avoid Kurzweil.
The singularity is nigh! This trope might make for great fiction but the on the ground reality is far different. Intelligence is multi-dimensional. No machine intelligence has yet shown an ability to match humans in multifaceted intelligence and the day when such intelligences can outpace humans is as far off as it was when the singularity was first posited.
It's not an intermediate goalpost. I'm referring to what essentially we're promised by those decrying AI who insist that it will match and exceed humans in terms of cognitive abilities.
As for it being 'game over', why? Is there something inherent in AI that would necessarily be inimical to human beings?
It's game over when machine intelligence controls all the resources needed to ensure their continuation. As long as they are trapped and subject to our control of the power switch and reproduction they can't do much that we don't want them to.
In Kurzweil's book 'The Singularity it Near' written in 2005 he predicted it for 2045 and has never changed the predicted date so we're still 26 years off. We've recently had AlphaZero beating us at games and Waymo cars kind of driving so there's some progress on the ground. Give it time.
In the field this is the idea of General Intilligence versus narrow intelligence.
Alexa, for example, is an Artificial Narrow Intelligence. It can process speech and then follow different scripts with instructions derived from that speech, but it often fails comically so you as the human have to talk to it just right for it to work. Not too different from a verbal command line.
Meanwhile a human personal assistant has general intelligence. You can just tell them what you want and they can understand and figure it out.
Aren't you a little bit worried ? I mean there has been a long history of "you can't do this with computers".
It started with carrying out list of simple instructions, then doing anything with images, then recognizing images, then drawing images, then simple games chess, recognize audio (though people realized this was stupid in the 60s with IBM speech synthesizers and so today people don't remember), and then Go (and less well known: backgammon, video games, ...). For most of those we now roll our eyes and go "how could they have been so stupid ?".
As for practical applications: robots run something between 3/4ths and 5/6ths of the stock market. Maybe you're unaware of this but that's the thing that decides how and where humans work (not just the ones also doing stocks, anyone working at any public company is partially directed by it, which in practice is nearly everyone).
AIs talk to humans more than humans do. AIs produce more writing than humans do. AIs judge more humans (for insurances, or creditworthiness) than humans do. Despite what you think AIs actually drive just a little under 1 millionth of total miles driven in the US today, going up exponentially. AIs currently drive a little more than a 1000 person city does.
In experimental settings, AIs have beat humans at convincing other people that they're human. At "chatting up" humans. Seriously. Not that it seems to take much at all to convince humans you're sentient: iRobots have convinced soldiers to threaten army technicians with guns into repairing them. There wasn't even that much AI involved, but that's got to count for something.
In research, I would argue there are already multifaceted artificial intelligences in reinforcement learning. There is nothing in those atari game playing AIs about the games. There used to be the score of the game, but the modern ones don't even have that. They can play any atari game, from Montezuma's revenge to pacman, which I'd argue are very different indeed. There must be some measure of "multifaceted" in there, surely ?
But let's keep it simple: could you make this problem a bit more accurate ? For example, which animals would you say exhibit an acceptable level of "multifaceted" intelligence ? Why do those animals qualify ? What would a good test be ? I'd love to find an interesting test for this.
>I mean there has been a long history of "you can't do this with computers".
Historically, people have both grossly overestimated and grossly underestimated future technological progress. Some people thought computers would never be able to play chess; others thought we'd have superhuman AGI in 2001. The bottom line is that people's past and current ignorance tells us nothing about what the future is actually going to be like.
Recursive self improvement in which field of intelligence? Does getting better and better pattern matching eventually lead to human level intelligence? Does improving pattern matching accelerate our ability at improving pattern matching even?
We can solve any board game now with alpha-zero but will that necessarily improve how fast we develop other types of intelligence?
I used to be in Rays camp but my girlfriend got into Computational Neuroscience and I talked with her and her colleagues got the impression very very few people think we’re close to general intelligence.
Human intelligence is a lot of things put together and we may get better at pieces but we don’t even have all the pieces and when we do try to put them together it doesn’t work. Look at criticisms of Europe’s Himan Brian project[1] (another example I had seems to be outdated), some believe we don’t understand too much to even begin to attempt modeling the brain.
There are actually a bunch of different ideas that all go under the "singularity" label making the term fairly confusing. You describe one. Another is the idea that if AIs start designing computers you'll have a positive feedback loop in improvements in computer speed. Another is that we can't predict the actions of a smarter-than-human intelligence. And then there's Kurzweil's idea that progress tends to speed up and at some point we call it the Singularity for some reason. I just wanted to point this out because I've seen a number of arguments caused by people using the same words for very different ideas of "the singularity" without realizing it.
I think (human)intelligence almost guarantees recursive self improvement; thus, if AGI = human intelligence or greater, it would also guarantee recursive self improvement.
It’d be a lot easier to believe not coming from a company selling ads. Google would be far more of a self improvement tool if it were not incentivized to sell you things you don’t want.
Despite the cynicism and his black and white predictions, I think his rhetoric still has valuable contributions. It forces others to take a true account of what intelligence is and what kind of intelligence is AI capable of in the medium term (evidence: this thread).
For some reason, this doesn't get enough attention and we have people like Elon and Stephen Hawking making dire predictions all over the place.
I disagree. I think his rhetoric feeds the 'we already understand it' mythology around the nature of AI technologies, and worse the state of our own understanding of the system they claim to model, and directly contributes to the harmful and frustrating boom-and-bust cycles that AI goes through. I'd lump him in the same category as Elon Musk and Stephen Hawking in this space; sellers of fantasy.
I admire his optimism, but I think it's irresponsible to sell it like he does.
Marvin Minsky ("Father" of AI at MIT) made the bold claim in 70s that we will be marrying robots by 2000. After it didn't come to fruition, it led most AI researchers to take stock of the situation and realize these goals aren't that trivial.
Similarly, if we don't have bold predictions like these that we can actually measure within our lifetimes, we fall prey to fantasies that cannot be measured. Once this prediction fails miserably (I think) it helps many others to re-calibrate all their BS.
> I can't believe there are people genuinely afraid of a hypothetical powerful malevolent AI
I don't think even the AI doomsayers, deep down, actually believe what they preach. It's just a way to signal that one is clever and informed of new tech.
If they actually believed what they say, they'd be worried of being targeted by violent protestors, like drug testing companies and crop breeding companies have to be.
I've felt for a long time that Singularitarians are looking in the wrong place. They see accelerated technological development and assume that the endpoint will be an artificial brain in a box. What they fail to see is that these inventions and breakthroughs haven't been about increasing the intelligence of a machine... they've been about increasing the intelligence and efficiency of human systems.
The singularity isn't a brain in a box... it's us, the collective, a metasystem transition that's been underway for millennia. A movement toward a whole that transcends the parts.
That's what seems apparent to me, so far anyway. AI is more about augmenting human intelligence than it is about smart machines. All that impressive DL stuff is ultimately providing more tools for humans to be more productive.
Here's my shorter-term thinking. Current ML can't generalize well [0], can't do arbitrary & conceptual thinking, and has trouble with language.
[0] by that, I mean that it does not easily pick up higher-level patterns unless explicitly forced to (through either model architecture, data & task setup, etc). Meaning - it has a much lower success rate outside of the training data distribution compared to "flesh" intelligence. Kind of reminds me of https://www.youtube.com/watch?v=PHRvF0m3yuo
Hmm, current ML has only been worked on for a few years, we have barely scratched the surface. No signs of slowing down afaict. It has wiped the state of the art against algorithms we took decades to think up. What happens after decades of that?
It feels more like there is too many new directions to explore and we don't have enough time/creativity/insights/ideas to fully take advantage of it.
1957 - First ever man-made object launched into space.
1957 - First ever animal launched into space.
1960 - First ever animal launched into space that survives the trip.
1961 - First man goes into space.
1965 - First man goes on a space walk.
1966 - First spacecraft lands on the moon.
1967 - A probe is landed on Venus (!) which collects and sends back video and audio recordings among other data.
1968 - First spacecraft to orbit the moon and then return to Earth.
1969 - We put men on the moon who then safely return.
We've done all of this in just 12 years! We literally went from nothing to men on the moon in 12 years! We've barely scratched the surface of what's possible and there's no sign of things slowing down at all! Can you imagine where we'll be after decades of this. Where will we be in 50 years? I can't even begin to imagine what 2019 will be like.
I don't say this to be snarky, but to point out the issue. Exponential growth doesn't mean continued exponential growth, and it's always very difficult to predict the future. While thoughts such as the above were common place, there would have been hardly a person to predict that the room-sized government "computing machines" would one day be a few inches large, millions of times more powerful, and cost as little as a week or so of minimum wage work. In some cases, those exponents did keep churning for a while.
On AI, I used to be a futurist in terms of it - until I got to work with it. Now I'd be extremely surprised if we have have fully capable self driving vehicle within the next decade that's not constrained to white-listed routes. And that is an extremely softball problem since it fits perfectly into the small domain of problems that current ML/AI systems are really great at making progress on. The thing is getting to a testable model is absolutely trivial. Going that remaining 10% from model to product is many orders of magnitude more challenging.
I had a chance to interview Ray and spend some time with him before he joined Google.
The way we met was very serendipitous -- I asked a question during a movie premiere about the nature of reality. He must have enjoyed it because later someone from his team sought me out to invite me to a VIP after party.
His publicist at the time said this was the best interview he ever gave (It's possible she says that to all the girls, but, I'll take it!)
It's much easier to make a plane that can fly than to emulate the particular way a bird flies.
They'll both solve the flight problem, so it doesn't really matter.
Flight is something extremely comparible to ai development. When it was developed, many companies were trying, and a lot of people said it was impossible. The problem space is also similar. We're not sure how to get there exactly but we may be close.
It was, in the end made to happen by an unlikely pair - not a large company with a lot of investment. I believe this might be that fate, too.
It matters tremendously. Think about the economics of bird flight vs. plane flight. Why is it that flight has been around for almost 100 years now and yet we still aren't flying everywhere? The reality is we can always come up with subpar ways of doing things, but there's a reason birds have evolved as they are today that we just can't replicate ourselves. We mastered long haul rapid flights, but bird flight generalizes much better. They might be slower in long haul, but they are much better and short haul and medium haul and have evolved to work in groups to make long haul possible. So yes, analogously we can create intelligence in a different way, which we are, but it likely won't generalize as well.
If we could reduce the density of a person by about an order of magnitude we can start replicating bird flight on a per-person basis. Our approaches to flight are constrained in all forms by the desire to put really heavy things into the air.
Look no further than drones that can operate for quite a long while, move in any dimension fairly well, and manage some pretty ridiculous speeds for not being designed to do so all because their purpose is not to carry large complicated amounts of weight.
The difference is we knew what flight was and had an easily visible and verifiable goal that was obvious to all when it was achieved. What is intelligence? How do you measure it? How will we know once we have created? We haven't even defined the problem yet, how can we know when we have achieved it?
No, the wings of the wright flyer don't flap, but that doesn't mean it wasn't designed based on birds. One key piece (controllability) came as a result of the wright brothers studying how birds adjusted the shape of their wings.
Brave claim: we won’t actually invent good AI until two-way brain-computer interfaces become useful.
We need to augment human intelligence to handle systems vastly more complex than we can do today. Specifically memory creation and recall, and information assimilation rates.
My keyboard and monitor are brain-computer interfaces. I'd argue that they are pretty useful.
I can imagine a way to consume information that is dramatically faster than the 500 or some odd words per minute I get reading.
But for input to the computer? I mean, I'm not saying that a keyboard can't be improved upon, but I think there are some hard limits to how quickly I can compose my thoughts; I think that it's fairly rare that the fingers are the limit. (Of course, making it so I think less about the fingers would be great)
What I'm saying is that there might be human processing speed issues that limit how much value we can get out of faster I/O.
On the other hand, maybe it's like reading was for me; before I could read at a certain speed and certain ease, I couldn't enjoy books, because I'd forget the beginning of the paragraph before I got to the end. Like I needed to load enough words at once to see the picture. It's totally possible that output would be the same way; if I could somehow output the equivalent of five hundred plus words per minute, maybe that would change my writing the same way passing that speed changed my reading?
I think most of the bandwidth would be computer -> brain, not the other way around, assuming it can teach your brain to retain knowledge and show it images/movies. The only exception would be if we could significantly alter the perception of time, but I'm not really betting on that despite how useful it'd be.
Even if I could only accurately sample 1-8 bits/second from the brain, that would be world-changing. There are a lot of clever people who could stuff a lot of utility through 8 bits/second given you are paired with a powerful smartphone and a proper "UX". Clever encoding schemes and signals could be developed. Especially if you go for sequences/chords that are executed over 1-4 seconds.
Plus the brain would likely just act as an index to fetch info that is then replayed to you. I really don't need a terabyte in the form of neuron connections when storage and computing would be ubiquitous.
IDK, Google's voice assistant thing still can't figure out my wife's name because it's not an English name and I pronounce it as it should be pronounced.
What if human beings were physically connected to a growing network of perfectly organized facts and information? What if that network could even assist with basic logical reasoning tasks? Would that be enough to say we’ve reached the singularity?
It doesn’t seem like it now because our phones aren’t quite physically connected to our brains, and the internet is far from organized, but I’d venture to say we’re already pretty close. A more interesting question to me is whether this is really anything we want.
If human machines were infinitely intelligent then they would all agree. So you wouldn’t need more than one. Sure, the ultimate human might need 7 billion pairs of eyes but those eyes wouldn’t need any more brains than to fuel up and locomote. Furthermore any human-like tendency to doubt oneself or fall in love or “rock out” would be deemed useless and therefore overridden.
Understanding that the purpose of a human is to pass on genes, and that there’s little human or genes left in her, the ultimate human might therefore conclude that her sustenance only disrupts the purposes of organic life forms. Her next and final act would be to destroy herself.
This is how I see the end of humanity. Not in a violent clash between artificial and organic intellects, not in a symphony of mushroom clouds, but with the final flip of a switch, in cold, calculated resignation.
However as the ML researcher Michael Jordan (one of the most important in the field) has previously stated, these sort of long-term technology predictions are just fun science fiction and there is essentially no academic rigor in this stuff:
Betting on technology evolving faster than expected is usually a safe bet, but it feels like Ray has been working on flawed premises and refuses to revise them.
After reading it all, I don't really see him as being any more successful in predicting than you and me. You get some right, you get some kind-of-in-the-right-direction, you get some wrong, and some are just laughably wrong.
As we do not possess a clear (or, at least, a generally accepted) definition of what constitutes the "true AI," it seems very likely that we will cross this "even horizon" without realizing it; we may have done it already, for all we know (or, rather, don't know - due to the fact that much of the progress is being done in secrecy).
On pace to do what? Add numbers really fast? Play chess? Make a best guess of which news story I want to see? (Hint: not sports) Perform automatic gear shifting? Follow the vehicle in front of it?
Kurzweil has long been preaching that there will be a qualitative revolution in what AI will achieve - and it has been 15 years away, for the past three decades.
Not actually true, the always 15 years away bit. Three decades ago (1990) his prediction was Turing test and AGI in 2020–2050. More recently he's predicted Turning test 2029. He may be wrong but doesn't keep shifting the dates 15 years away. (https://en.wikipedia.org/wiki/Predictions_made_by_Ray_Kurzwe... )
Ray Kurtzweil is certainly super smart, but I find his views about the Singularity disturbingly myopic. It's as if he has been living in a bubble for decades, and doesn't realize how the world works for real.
I still have a lot of sympathy for him, as I'm sure he means good. Still, these views distort our reality for many people.
Of course it will happen. We humans are an outcome that took billions of years and yet here we are. It might take a billion more to get sentient AI but as long there is progress then the end is inevitable.
The only debate is when it'll happen, but we probably need an advanced sentient AI to help figure that out.
I suspect human intelligence is more than just computations in the brain. A lot of computation I assume most occur 'outside' of the brain in the gut via hormones and protein signalling. I have doubts that you can replicate human intelligence without replicating human biology.
I doubt a lot of computation occurs outside the brain. I'm assuming most of the neurons, nerves, etc in the gut are used primarily for regulating gut health rather than calculating, reasoning, etc.
People have had their intestines completed removed, liver transplants, kidney transplants, stomach stapled, etc without hurting their intelligence or computational ability. I don't think we can replace the brain like we can elements of the gut.
Certainly, the gut, hormones, etc are important to your physical and emotional well being, but I think it's safe to assume that pretty much all over the computation is happening in our brain.
Or even if it were all in the brain, accurately modeling neurons and interactions of neurons, on bajillions of cudas and threadrippers, will probably still take much longer than it takes for the real-life system to do its thing:
Modeling the human brain is just one of the directions of the AI research, possibly not the most important one, even, at this point. On the other hand, the "super-intelligence" everyone is thinking about will be nothing like the human mind.
They are control utilities, regulators for the brain, whose generation is initiated by the brain. However I think you are right with the need for human body. Without sensory input its state would collapse, I think it needs a world with constant stimuli to live in.
I think we will find limited use for general AI since we won't be able to properly regulate or understand the behavior of machines that are smarter than us. Alpha Go Zero can't explain why it plays chess the way it does except for showing how it evaluates the board, but the reasoning is buried in millions of mathematical calculations accreted via millions of games of self play that we can't understand the intuition of. Of course, many smart players can't really tell you systematically how they make the moves they do either. AI will be like a dictator who makes his decisions by intuition only.
An intelligent black box can still be "useful". It could help cure diseases or get original insights in many fields for example. Even if we don't understand how it thinks.
>But What About the Title of this Presentation ....
While obviously a cheap hook, it is nevertheless intended to convey the possibility that it is our emotions, our passions, our innate desires — all ingredients of our sexuality — that are the defining elements of our consciousness and, through consciousness, our intelligence.
The End
I always felt that Kurzweil’s timelines were too optimistic not only because technology may not evolve so fast, but also because it fails to account for the delays that government regulation imposes. So many of the vaunted leaps in medical technology, for example, are going to require years and years of testing before they get approval for use. Other whizbang technology of the future might be prohibited from the consumer market because governments see it as too dangerous.
This may be the case for many technologies, but clear wins that save lives get fast tracked. It just happened to CTX001. Also older researchers are doing self testing all the time, as they know that they don't have the time to go through regulations.
I think Moravec's arguments on the same subject "When will computer hardware match the human brain?" are a little less cranky than Kurzweil's and his prediction from 1998 "it is predicted that the required hardware will be available in cheap machines in the 2020s" seems to be panning out. https://jetpress.org/volume1/moravec.htm
The Singularity is just millennialism [0] for secular technolgists. We still have a pretty paltry understanding of human intelligence. It seems unlikely to me that more and more iterations of AlphaGo will spontaneously produce strong general AI. My bet is no AGI within my lifetime.
The brain and "mind" is insanely complex. There is so much we still don't understand, so of course there is so much we cannot replicate. But is it possible? If there is an incredible breakthrough in understanding, then yes of course. Personally I see little value in people saying it can or can't be done (of course it can!), it's just a matter of if and when. We shall see...
He predicted Turing test by 2029 which is fairly definite, falsifiable and quite likely to be falsified.
For me proper AGI is when you could say to a robot go build some better robots and then have them build even better ones and they can do so without needing us. There will be some year before which if humans disappeared, computers and robots would grind to a halt and after which they could carry on. That's also a definite falsifiable thing, harder than the Turning test and is kind of what I think of as corresponding to the singularity which tends to be rather wooly in it's definition.
2019: "The computational capacity of a $4,000 computing device (in 1999 dollars) is approximately equal to the computational capability of the human brain (20 quadrillion calculations per second)."
$4,000 in 1999 would be about $6,000 today. A midrange Deep Learning workstation is in the ballpark, and while it can't "match the computational capacity of the human brain", it can do plenty of nifty things - including beat the current Go world champion in a match.
I don’t find this any different than a steam shovel outpacing a human digger. It’s still necessary to prevent both from destroying things and harming people.
Yawn. Of course he has to say this, he makes his living saying stuff like this. He is, after all, a ”futurist”. It just seems rather trite at this point.
The brain is denser, more efficient and 3 dimensional. Compared to the current limitations of silicon the combination of those three aspects is downright magical.
The most important part of the talk was when Kurzweil mentioned running partially-trained AIs in simulation to generate more training data than exists in the real world. That was the first clear description of imagination that I've seen, although others have hinted at it, for example the running stick figures (that I think Google's DeepMind) trained to run an obstacle course several times before they actually did it. I might be remembering that wrong, but there have been a few similar examples.
Anyway, the reason why the imagination part is important is that nobody today is talking about parallelization when it comes to AI. A few "rules" of AI computing and their limits:
* Processing power doubles every year or two for the same cost -> processing power will someday be approximately infinite (for certain classes of computation)
* The search space for even the simplest problems is effectively infinite if we don't know the hidden models -> if we know all the hidden models, then large search spaces become tractable
* The sequential portion of intelligent decision making is less complex than the parallel portion (computers have been beating us at all sequential operations since perhaps the mid 60s or 70s) -> when the parallel portions are solved, then computers will beat us at all intelligent decision making
That last point is really important and exciting, because most of us simply aren't accustomed to thinking in terms of solving problems in parallel. We're thinking "how do we train a self-driving AI to recognize pedestrians and swerve to avoid them" instead of "how do we optimize the solution to the series of equations to drive to the store and back without hitting anything, in a single pass".
AIs will soon be looking forward from their current model of the world, running some number of simulations (potentially billions) and choosing the best outcomes, then feeding back those solutions to improve their hidden models of the world. All in parallel, with virtually unlimited computing power. At that point we'll be training AIs like children and will start to see emergent behavior that mimics life.
Personally, I don't see how this can possibly be any more than 10 years away, 20 at most (even for rank amateurs like me), given the number of people tinkering on these evolutionary patterns, and the amount of open source research being generated.
It kind of haunts me actually. Like, we just started investing in our 401k at work, but I don't have the heart to tell everyone that the singularity will probably arrive before their investments mature. Like, how does one have a child when it's looking like the 20th century reality we're living in isn't going to be around much longer? Do we suffer through the daily grind, the 40 hour weeks of converting our time to money, when AGI will overshoot that within weeks of surpassing our definition of sentience? It literally might make more sense to become a beach bum, or travel the world, go to Amsterdam and spend a few years in a drug-induced haze.
Or is the correct answer to keep going through the motions of running the rat race, knowing that it's all a waste of time but constructing some kind of inner monologue to distract us from the depression induced by the real world's unfolding existential nightmare? Is there a way to un-take the red pill and go back to the blue pill?
I guess I digress, but this is literally all I think about anymore. The futility of work/culture/technology in a world of ever-increasing human subterfuge. Why can't we admit that the dystopia is playing out before our eyes and begin to escape the yoke?
I guess the one thing that keeps me going is that we might be able to break down a single day into its component parts and automate away anything that's beneath human potential so that we can do the stuff we'd like to do (like nothing). The Roomba can vacuum (done). The self-driving Uber can take us to work and back (almost done). The rooftop solar panels can pay our electric bill (effectively done). And so on, until there is no minutia left.
The last step will be to spin up a partial (of oneself) to go to work and earn the paycheck, or literally grow the food to feed us so we no longer have any external dependencies. This is the part that I think is roughly 10 years away.
Y: The majority of reading is done on displays rather than paper
N: Most text will be created using speech recognition technology.
N: Intelligent roads and driverless cars are in use
N: People use personal computers the size of rings, pins, credit cards
Y/N (premature): Though desktop PCs are still common for data storage, individuals primarily use portable devices for their computer-related tasks.
Y/N (premature): Personal worn computers provide monitoring of body functions, automated identity and directions for navigation.
N: A $1,000 computer can perform a trillion calculations per second.
... seemed a total mixed bag, many 2009 predictions just beginning to coming true today
By 2019:
N: The computational capacity of a $6,000 computing device is approximately equal to the computational capability of the human brain (claims that is ~20 petaFLOPS. Nvidia DGX-2 is 2 petaFLOPS "AI" supercomputer for $399,000). [
N: Computers are embedded everywhere in the environment (inside of furniture, jewelry, walls, clothing, etc.).
Y(easy win): Most people own more than one PC
N: Cables connecting computers and peripherals have almost completely disappeared.
Y/N: People communicate with their computers via two-way speech and gestures instead of with keyboards. (can but in my home and my office most words entered are not)
N: Most business transactions or information inquiries involve dealing with a simulated person
N: Rotating computer hard drives are no longer used.
N: Three-dimensional nanotube lattices are the dominant computing substrate.
N: The algorithms that allow the relatively small genetic code of the brain to construct a much more complex organ are being transferred into computer neural nets.(Could be decades off. Need several breakthroughs to even put this on the horizon)
N: Most roads now have automated driving systems—networks that allow computer-controlled automobiles to safely navigate.
N: Most decisions made by humans involve consultation with machine intelligence. For example, a doctor may seek the advice of a digital assistant.
... few hit but we're mostly heading that way. To me seems 5-15 years premature
What's interesting is that I think most people own just one general purpose computing device these days, their cellphone. Yeah, they maybe also have a TV that can play netflix or a game console, but my observation is that most people don't even have one PC in the old sense of "a general purpose computer with a physical keyboard that you use by sitting down" - and if you exclude laptops from the definition it's even more rare.
> Cables connecting computers and peripherals have almost completely disappeared.
This one is unfortunately happening right on schedule, at least if you use Apple hardware. "A $1,000 computer can perform a trillion calculations per second" should be considered "premature" as well, seeing as current GPGPU hardware can go well beyond 1 TFLOPS.
I didn't vote either way, but long term exponential growth in anything is basically impossible.
Think about it for a minute. At some point lithography gave us Moore's "law" (which, FWIIW has arguably already failed). Do you think that can continue forever? Pretty sure you exceed computational capacity of the universe pretty quickly. Exponential growth in cell phone sales happened at some point as well. As did exponential growth in transportation speed.
The idea that we've had exponential growth in anything but lithography technology in recent years is also laughably insane. Do you think programming languages are exponentially better than they were 30 years ago? Do people live exponentially longer? Cars, autos exponentially cheaper or better in some way? Are machine learning algos exponentially better? No, no, no, and no.
I don't think anything like the singularity will ever be possible, because the ability to speak language for oneself and invent complex reasoning is spiritual, not merely physical. Humans can do it because we are inherently spiritual beings, a union of body and spirit. (I know most folks here don't subscribe to that assertion... I'm not here to argue it anyway.)
But even if that isn't true, humans are not merely the product of our biology. Each of us has undergone an $age-year period of training, 18 years of which was on someone else's effort and time. I doubt someone is willing to put in that much personal investment to raise a neural net like his/her own child.
Your unusual argument recalls a discussion I once had with members of the Go club at my alma mater, nearly 10 years ago. On the question of whether Go would ever be mastered by a computer, one asserted that the game needed to be understood at a spiritual level - not simply as a computational problem - and that’s why computers would never reach the level of humans. The game was simply too sublime and complex, came the claim.
At the time this might have seemed reasonable - after all, Go software struggled mightily for years to achieve even beginner levels of play, and seemed incapable of mastering the intricacies of high-level play.
Well, we know now in retrospect that spirit was unnecessary to play Go. We know now that it was, in fact, a computational problem. There’s really no reason to expect that “spirit” is necessary to perform the tasks of language or cognition - tasks which are demonstrably executed by impulses of chemicals and electricity pulsing through the brain.
I know spiritual reality is sometimes invoked in discussions about things we haven't yet mastered or understood in science and engineering (one might even accuse me of doing that). But your story is kind of a straw man example, because even 50 years ago we knew of algorithms that could theoretically render Go a solved game, given enough computing resources. You could just as easily say modern cryptography is spiritual in its essence, which would be nonsensical.
Emulating a human mind with computers is an entirely different class of problem. We don't even have a mathematically rigorous definition of what constitutes a mind, much less know of any algorithm that could implement or solve it, even with any given supply of computing resources.
> Each of us has undergone an $age-year period of training, 18 years of which was on someone else's effort and time. I doubt someone is willing to put in that much personal investment to raise a neural net like his/her own child.
Unlike humans, software is reproducable and doesn't expire. So even if such a software took 18 years to develop, it is still a worthwhile endeavor since it has the potential to be used by countless people well into the future. Furthermore it is not necessary that it is developed by just one person, nor is it necessary that it is developed at the same speed that humans develop. The creation of such a software could take 100 years and it would still be worthwhile.
But if something is spiritual, there must be some way for it to interact with the physical world, otherwise you would not know about it. Then that way can be measured, which makes it physical...
Unless you are calling spiritual those things that can't be measured, yet?
For the sake of argument, "spiritual" could be a class of physical objects that can only be produced by other spiritual objects. This is the old "vitalism" argument, which is generally considered discredited, but as the "hard problem of consciousness" is still unanswered, this is a small loophole that proponents can use to hang on to it.
We humans are blackbox physical machines that processes physical inputs to produce physical outputs.
Whether that input goes through processing that is "spiritual" is besides the point. Reproducing "spirituality" is not the goal. The goal is to replicate what is "observable." To produce a machine that when given the same inputs as that given to a human will produce identical (or "better") outputs as the human.
This is definitely an achievable goal as for every set of inputs and output pairs in existence an infinite amount of different machines can be constructed to replicate that exact input/output functionality.
The "spirituality" that goes on within this blackbox machine or humans is irrelevant. It's not even a word that is clearly defined. We aren't trying to replicate "spirituality." We are trying to make a machine that only from an observational standpoint is clearly superior or indistinguishable to a human intelligence.
The understanding the theoretical possibility of singularity is purely within the grasp of a typical person.
Imagine a chat bot designed to be indistinguishable from a human chatter. The inputs are text queries and the outputs are text queries. The domain is every possible text query and the range is every possible text response.
For purposes of simplicity lets say the text input and output is limited to 500 chars. This makes the possible input and output pairs finite. To create a chat bot identical to a human you simply need to create a machine that maps all inputs to certain outputs. Think of it like a giant hash map of strings to strings. Given enough time and enough memory this can be done.
The goal of current AI is to solve this problem more efficiently. To produce logical cores that can dynamically generate outputs from inputs rather than following a gigantic mapping. It can be thought of as form of data compression. The data is the mapping; the compression of that data is the logical core.
The fact that a mapping can even be constructed at all is by logical induction, a proof that a singularity is possible and thus completely invalidating your initial statement of:
>I don't think anything like the singularity will ever be possible
I love Ray Kurzweil and he's an undisputed genius, but like many other believers in The Singularity, he's convinced of something that has never been proven to be true.
No computer will ever demonstrate actual intelligence, which is the ability to bring a new and unforeseen solution to an existing problem. I need only a cite a few examples to show this is true: Uber cars run over people. Facial recognition claims black people are criminals just for being black. Facebook's AI software lets in fake news.
These are all errors that could be avoided by actual human intelligence. The fact of the matter is that we don't really know where intelligence or creativity comes from. It's unlikely that we ever will. (See "Heisenberg Uncertainty Principle" for one reason.) While machines are great at doing things faster than people can do them, they're not good at inventing new solutions or doing something that humans couldn't do, given enough time.
> No computer will ever demonstrate actual intelligence, which is the ability to bring a new and unforeseen solution to an existing problem. I need only a cite a few examples to show this is true: Uber cars run over people. Facial recognition claims black people are criminals just for being black. Facebook's AI software lets in fake news.
Burden of proof is not on your claim, but rather on Kurzweil's, but since you made a claim and tried to back it up, I feel compelled to point out that it's a non-sequitur. Just because something hasn't been done before, doesn't mean it's not possible. Many people said going to the moon was impossible. Might be interesting to look at Russell's Teapot.
In general, that which nature has demonstrated is can usually be replicated. A bird (flight), a floating log (ships), a fish (submarine), an asteroid (space travel), etc. Nature has demonstrated intelligence: a human.
However, just like nature has not demonstrated superluminal travel, it has not demonstrated super-intelligence; so that is still a question.
> No computer will ever demonstrate actual intelligence, which is the ability to bring a new and unforeseen solution to an existing problem
It's easy to disprove absolute claims such as yours. Yes AI can create new things and it's only getting better. You just need 1 example. Read up on AlphaGO. It was able to beat humans by coming up with techniques that no human has ever and possibly would never have come up with because of how illogical they seem to even the masters of Go.
Contemporary AIs don't create new things via creative insight or intuitiveness. A trained NN is only good at the one thing it has been trained at. That isn't true intelligence, no matter how astounding the results.
The means to do those things were programmed by humans. While the computer is faster at assessing all of the possibilities, it isn't coming up with anything that wasn't made by possible by the human-programmed software.
We now have to have safety belts for Amazon workers to keep them from being run over by robots!
"Ever" is a very long time. But I'd be comfortable saying AGI won't arrive in my lifetime, at least not electronic AGI. Note that I single out electronic AGI here due to both our complete lack of understanding of how to model intelligence, and the prohibitive energy cost of modelling the brain if we did understand how to do so.
OTOH advances in genome understanding could yield e.g. a super smart chimp, or dolphin, something like that. That would also be "artificial" intelligence then, and it would be general, just not electronic.
Unfortunately we are still vastly better at making a worse mouse (or cow) than a better one, to paraphrase biologist Anne McLaren. Sure, evolution did eventually make a super smart chimp, but it took 7 million years of tinkering.
haha. maybe no computer by the current definition of computer. but I would say that it's a bit extreme to say humans are going to be the only higher intelligence species.
the only thing that I think we should have disagreement on is if we want it to happen and of course the timeline.
for the if question: I believe we (humans) want to do this. Long term it is going to basically mean immortality and I would not think about intelligent machines as a new rival species, I would think of it as an evolution of our species - it's actually a great way of transcending our biological limitations.
for the timeline piece: I don't think it's going to happen within our lifetimes. I also think it's going to be a process where for a very long period of time we will be hybrids (we will augment ourselves with hardware that will make us smarter/more capable and this in turn will lead to us building even better and more comprehensive hardware. At some point we'll figure out that the biological part is just getting in the way and we'll completely remove it).
>No computer will ever demonstrate actual intelligence, which is the ability to bring a new and unforeseen solution to an existing problem.
This is already false, given the many examples of machine learning out performing humans in a variety of tasks. Google's data center efficiency comes to mind (IIRC saving 30% in energy usage over human-found solutions). So does AlphaGO.
Outperforming in these cases just means "doing it faster," as I said. There is no algorithm which can produce results which were not inherent in the algorithm itself. There is no software which can "jump the box" and come up with a solution which was not already present in the software as it was written. The computer is merely processing all the options faster than a human could.
When we say that software "found" something new, it's only that it found it faster than we would, using the algorithm we wrote.
> These are all errors that could be avoided by actual human intelligence.
Do you think that dogs are intelligent? Can you imagine some situation where the dog intelligence is not enough to solve it?
Unlike dogs intelligence, the intelligence of computers is improving. Now they can play chess and go like a grand master, but 30 years ago it was an impossible task.
The problems you cite are real, but sometimes people fail at them too (drivers regularly run over people, witness misidentify suspect because they have a skin color that is not the skin color of most of the people they see, people actually believe and like fake news). Just give some years until the interface with the real world is improved and computers have more "street" knowledge to apply in difficult situations.
> See "Heisenberg Uncertainty Principle" for one reason.
It's related because it says that the observer cannot escape the observation; he's creating it. We cannot objectively assess our intelligence or creativity because we are the ones observing them. If we can't understand or assess how they came to be, we can't replicate them.
Exactly. Software inherits the limits of the people who program it. It doesn't expand upon or escape them. Actual human intelligence, in contrast, can "jump the box" and come up with something that didn't exist in the original materials as presented to the person.
> Uber cars run over people. Facial recognition claims black people are criminals just for being black. Facebook's AI software lets in fake news.
Humans do all of those things too. We might not believe they are the right things to do, but doing them certainly doesn't rule out the presence of a human-like intelligence.
In fact, it does. It shows that human bias (facial recognition) and the inability to posit all possible outcomes (Uber; halting problem) are built-in to the software we write.
While another human may be able to overcome one's racial basis or another driver wouldn't have run over that biker, the fact remains that the software did which shows that it inherits human frailties and limitations. It doesn't surpass them.
> No computer will ever demonstrate actual intelligence
Such statements are impossible to be proven right, as, assuming you consider yourself actually intelligent, that implies we are not inside a computer ourselves, which I'd argue is impossible to prove.
If we're running in a simulation, the host computation hardware may be different from what you imagine a computer to be. For example, we could be an experiment of "is life possible in a euclidean universe".
> that implies we are not inside a computer ourselves, which I'd argue is impossible to prove.
Perhaps not. But some statements cannot be disproven and yet still should not be taken seriously.
But I think you are right that, while correct in claiming that the Singularity has never been proven, mimixco is also making statements that cannot be proven.
Computers that are capable of analyzing and understanding their environment with a level of fidelity comparable to a human, without being preprogrammed with information about the nature or structure of the environment, are out of reach for the foreseeable future. I don't see any fundamental reason why such a computer should be impossible, but there's not even a realistic roadmap towards such a thing. That is to say, if it ever does happen, nobody alive today can predict when it will happen.
It would be foolish to claim we've made any meaningful progress toward a true AGI that can pass the Turing Test until someone demonstrates a computer as smart as a mouse across the full spectrum of activities.
Once we get computers as smart as a mouse, we'll be at most 3-5 years or so computers as smart as a human. We will have solved at the major challenges in AGI and it will simply be a scale problem then.
Saying we don't have mouse level AGI is simply saying human AGI is greater than 5 years away, which isn't a remotely contentious statement.
The difference in intelligence between an amoeba and a mouse is enormous compared to a mouse and us. People greatly under appreciate how intelligent and close to human a mouse/bird/pig are in the grand scheme of things. Emotions, behaviors, motivations, goal setting, memory it's all there already. A flat worm, an ant, a fly - those are the large stepping stone accomplishments.
Think about rate of very long distance communication in humans. It took us tens of thousands of years to get to 2.4 kps dial up modems, and only few decades to get to common 300Mbps. The important signal is seeing a 100bps modem, not a 100Mbps connection.
So the real question is how long until we can replicate a worm's intelligence?
32 replies →
I think you're right. I haven't read the article, but I think Kurzweil is perpetually too optimistic about what we can achieve, at least in how fast it will happen. There are cool things happening, and lots of advances being made, but we are nowhere near anything that could really be called AI.
Plus, I think all the marketing use "AI" is giving a very distorted and inflated view to the average person of what software is actually doing, and what it's capable of.
It's a buzzword, full stop.
17 replies →
What does it mean to be as smart as a mouse? If you specify a handful of tasks that demonstrate it, someone will be able to purpose-build an "AI" to do those things well.
23 replies →
Well, just about all the measures of potential future computer-intelligence more or less use "human abilities" as a placeholder rather than quantifying these abilities. That is indeed a testament to how little we have charted this intended final goal.
So basically, not only do we not have a road map, we don't know where we are going. That may be a reason for an extreme pessimism or it might be a reason for extreme uncertainty. Is adapting to the environment without prompting a small piece or a big piece? Could intelligence be a simple algorithm no one has put forward yet? If we don't know the nature of intelligence, we can't answer this sort of question with any certainty either way.
No one has put forward a broadly convincing road to intelligence. But maybe some of the so-far unconvincing roads could turn out to be right.
I think if we truly do create AI we will, to some extent, have stumbled into it, but that's OK. I think one of purposes of AI research should be to refine our definition of what intelligence is, because we don't have a very good one.
Perhaps, one day we'll "accidentally" figure out how to make it, and then we try to figure out what it is, because figuring out what it is in wetware hasn't been easy. Or maybe we'll figure out how to make it, but never really understand what it is.
It seems to me, that if we ever come up with AI that can exceed human intelligence (whatever we choose that to mean), that we might not ever be able to understand completely how it really works.
Even more interesting, if we were to achieve this AI, we also might not be able to make use it of it, because if it's truly "intelligent", then it will have a free will, and it might not wish to cooperate with us.
"I don't see any fundamental reason why such a computer should be impossible, but there's not even a realistic roadmap towards such a thing. That is to say, if it ever does happen, nobody alive today can predict when it will happen."
It's also not obvious why such a machine would not immediately self-terminate in the absence of a hugely complex system of scaffolding to shape and filter the raw input of existence.
I have not experienced mental illness myself, but my study and my understanding lead me to be extremely skeptical of a mind exposed to raw existence without filter. It appears to be a terrifying and unbearable state.
I think you are massively anthropomorphizing.
Every day I run programs that happily die alone on their own. Painlessly. All of the "pain" we feel is an artifact of our evolution, same with having a will to live. The only reason we animals fight so hard to stay alive is that animals which didn't died off, ergo, only animals with a will to live survived and reproduced. Those same forces don't apply to computer programs.
Even the concept of "terror." Who is going to program terror in? What benefit would it have? Why not wire programs to be "happy" when helping us?
7 replies →
It’s exactly how a newborn lives for many months. Of course they have coping mechanisms (adult tenderness, the relief of eating), but life seems to be both incredibly interesting and at times terrifying for them.
There's also the question of mental stability. Once we've given the keys to the kingdom to our new super-intelligent and wonderfully wise AI, nothing prevents it from developping dementia, of forms we may not recognize. It may become psychopatic, delusional, starts having hallucinations.
There is no reason to believe a human-level AI would not develop mental problems just like us. Given their unlimited lifespan, it could be inevitable.
5 replies →
It's entirely speculative but doesn't the model Richard Dawkins explains for the laymen in The Selfish Gene suggest something that would work - a drive to reproduce at the lowest level resulting in some incredible things happening over a long enough period of time.
The way I understood the comment, even in that scenario of the machine self-terminating, the point grandparent was making still stands.
But humans are "preprogrammed with information about the nature or structure of the environment"... i.e. 2 million years of the evolution of genus homo, and far beyond that too.
If real human intelligence is preprogrammed to a massively large extent, sounds like you are holding simulated human intelligence to a double standard.
Nothing in two million years of evolution programmed us to do programming. There is no common abstraction between it, and ape-on-the-savanna type problems.
Yet, nearly every human being can be taught how to program... But we aren't anywhere close to building an AI that can.
also, why assume computer intelligence is going to work similarly to human intelligence? I would assume that one potential way for it to work is to run simulations using data from sensors in the environment, and make more accurate predictions on an effect an action would have.
If such a machine can exist (and i believe it can, as computing power increases), then AI surely will follow sooner rather than later.
3 replies →
I respectfully disagree. Here's is the upper limit to when that will happen (meaning we know for sure it'll happen soon after):
- Complete behavioral reverse engineering of biological neurons, and neuronal clusters.
- Detailed connectome of the mammalian brain, e.g., first that of a mouse, then a cat, and finally that of a human.
- Replication of the above two in functioning electronic form.
Once you have this put in place, it's not hard to see that the subsequent investigation of calibration and testing of such a system, would generate new body of knowledge at an unprecedented rate. We may not immediately convert such a working system into macroscopic behavior resembling its biological counterpart, but it'll happen within a matter of years after that.
What ML/DL/RL folks are doing is only going to hasten this, by eliminating the need to carry out all of the above mentioned steps.
Perhaps, but that's a big if. We can't even say for sure whether a complete reverse-engineering of neurons needs to take quantum effects into account, as Penrose has suggested.
All you're saying is that if we knew exactly how humans work, we could build one. Seems like a tautology to me.
If I knew the exact quantum state of the Universe at the Big Bang, I could figure out exactly how the Universe evolved, but that's never going to happen either.
I think the complete reverse engineering the way you are describing will not be possible. We can only try to reproduce the same outputs for the same inputs. But I don't think we'll be able to fully define what happens in the black box in between.
We might come up with something that works similarly, and can do great things with it, but I don't think AI can be invented the way you are describing.
What you describe is a very high-dimensional non-linear system. We don't have the mathematical tools 'know' such systems. We 'know' a system when we can describe it by a much simpler (preferably mathematical) model. This is why linear systems are easy - we have mathematical tools to break it down to simpler parts (reductionism) and then understand the whole.
If the best we can do with the brain is simulate it at the the level of connectome/neurons/synapses/ thereby creating a system as complex as the brain - then do we really 'know' it ?
I feel the right word is 'Conscious AI'. AI in it's current form is very nascent, but does pretty well in specific use cases (image/speech recognition etc.). The true AGI will be the 'Conscious AI', which will be like a toddler, but learn about the world at exponential rate and perhaps become an adult in a month.
What is difference between Conscious AI and General AI?
There is a fundamental reason. Read Penrose on Gödel.
Even Penrose will tell you he's not certain on this being "fundamental". It's a theory he is working with.
Computer people really hate Searle and Penrose, because they spoil the beautiful picture of Strong AI.
We are born with pre-programmed information about the structure of our environment, IE instincts. Why should an AI be held to higher standards?
The manner of human "pre-programming" is so abstract that it falls outside the definition of what is normally meant by "pre-programming" in the context of AI. If you had some basic "instincts" codified into your AI that allowed for "self-perseveration" or was manifested as some type of reasoning-engine-BIOS that's one thing, but applying arbitrary training data sets onto your AI (as a pre-requisite for it to function maximally) is the type of "pre-programming" I'm getting at.
1 reply →
Primates have relatively few instinctive behaviours. Almost everything we do is learned to one degree or another.
8 replies →
> realistic roadmap towards such a thing.
Probably continuing with Deepmind's work shown here
https://news.ycombinator.com/item?id=17313937
OK it's not at human levels but it shows networks figuring out a 3d model of their environment
I am seriously interested in understanding why a lot of people have no issue accepting the idea of humans evolving from simple immaterial matter into biological beings and all the way up to our current reality but have a hard time believing that can be done via computers.
I would at least put it in the very likely box that computers can learn to be the same kind of pattern recognizing feedback loops that we are even without humans understanding the brain completely just as we became conscious and self aware without any "programmer".
"Computers" don't need to be like us to become intelligent, they don't actually need to reproduce the lungs and the intestine as we have it, they are in many ways free of those restrictions.
They might be "concerned" with very different things than we are and might even not really care about nature to survive. All they need is energy.
I don’t think anyone here disputes the logical possibility, but rather that such systems are still far off, not as near as the singulariati would have us believe.
From the perspective of embedded cognitive systems, intelligence is not an abstract property but a faculty derived from an agent embedded in a world with an ability to act upon that world in specific ways. This is reiterated in cognitive linguistics by Lakoff’s metaphor work et al. I take this to mean that the question of AGI itself is missing the point. To wit all recent provoking demos of AI (Alpha Go, DOTA) are embedded within specific environments and actions of a model-specified agent. It is then extrapolated that AGI will appear automatically induced by a series of more and more competent AIs.
Incremental evolution as a means of indiction as we know from biology can take a long time, and the landscape is complex. Hence the view that it might be coming but not soon and not with a roadmap.
> not as near as the singulariati would have us believe.
I've seen the singularity described as "The Rapture for nerds", which is pretty apt. It's going to be amazing, the future is so bright, we won't have to worry about any problems plaguing mankind now, and it's just around the corner! Any day now! Aaaaaaany day now!
Exhausting. Maybe we'll get a future AI-run utopia a la The Culture. Maybe not.
> singulariati
I like this term. Going to use it any chance I get now.
Biology works slower generation wise than technology. If selfawarenes is based on emergent complexity in pattern recognizing feedback loops then technology can play through billions of scenarios in a short time. We dont need AGI for technology to be a problem. Technology at the intelligence level of a mouse with all its potential physical power through the systems it might have access to can be quite dangerous for humans.
> but have a hard time believing that can be done via computers.
Much of the skepticism isn't about whether this can be done in theory. It's about "how close" we are given the current state of the art (beating Go, chess, deep NN, etc).
One can be (I am) simultaneously blown away by these accomplishments and believe that we have hardly scratched the surface of AGI.
And while this isn't necessarily a deal breaker for AGI, it's not encouraging that we still understand close to nothing about consciousness.
> .. but have a hard time believing that can be done via computers.
> I don’t think anyone here disputes the logical possibility
> Much of the skepticism isn't about whether this can be done in theory
The math tells us it cannot be done with computers. Read Penrose.
3 replies →
I am seriously interested in understanding why a lot of people have no issue accepting the idea of humans evolving from simple immaterial matter into biological beings and all the way up to our current reality but have a hard time believing that can be done via computers.
You seem to be missing a key ingredient: billions of years and endless branching iterations with the totality of all non-sequestered resources available. Even if you scale that to account for the more rapid iteration theoretically possible with machines, it doesn’t paint a rosy picture for any of our lifetimes.
Plus, in organisms the software and hardware are both subject to mutation, and exist in the context of a global system of competition for resources. That only tangentially resembles work done on AGI, and only in the software component. We can only design and build hardware so quickly, and that adds more time. I’m not hearing pushback against the possibilitY of AGI, just singularity “any day now” claims that seem mostly calibrated to sell books and seminars.
> I am seriously interested in understanding why a lot of people have no issue accepting the idea of humans evolving from simple immaterial matter into biological beings and all the way up to our current reality but have a hard time believing that can be done via computers.
And similarly, I’d love to understand why you’d believe in something for which there’s zero evidence and ask with incredulity why others don’t also believe it. Isn’t it reasonable to be skeptical until it actually happens? Isn’t it possible that we don’t understand what intelligence is, and that there could be an undiscovered proof that computers can never attain it? It seems too early to call.
Computers haven’t to date ever done anything on their own, what reason is there to think they ever will? We made them. Neural networks are nothing more than high dimensional least squares. Maybe humans are also least squares, but my money is on there being more to it than that.
> ...they don’t actually need to reproduce the lungs and the intestine...
Obviously, but the one thing that all intelligent life does is reproduce and contain a survival instinct. No computers have a survival instinct, and thus no reason to get smarter that we don’t imbue them with artificially.
There is a noticeable bias among computer scientists to believe in simplistic models of intelligence and minds. I agree with you, though, and I also sympathize to the ideas of Searle and Penrose.
> Computers haven’t to date ever done anything on their own, what reason is there to think they ever will?
Interestingly, people that spend a lot of time studying their own brain by meditating are usually very confident that what most people call "free will" does not exist - is merely an illusion.
Not to mention that physics tells us that we should be simul-able, after all physics knows only about deterministic and random processes.
1 reply →
>and that there could be an undiscovered proof that computers can never attain it?
It's already implemented in at least one physical system.
5 replies →
Also ponderable - there aren't many other intelligent animals around, but that is not actually evidence that human intelligence is hard to achieve. It is evidence that it is hard to turn a small increase in sub-human intelligence into an evolutionary advantage.
There is solid evidence that teams of mathematicians, scientists and engineers will be able to catch up with what nature has wrought in the next few centuries, if not much faster (decades on the optimistic end). Intelligence could be one of the easiest part of biology to replicate, given how shonky human intelligence is when tested against objective standards.
I question what objective standards are here. Human intelligence sucks in terms of consistent long term memory, and possibly in terms of not being influenced by emotions/outside forces. It's pretty good at a number of things computers are terrible at however, including ability to generalize from very limited "training" examples, combining "models" i.e. general object recognition with physical movement/obstacle avoidance, lots of things having to do with natural language. What kinds of objective standards were you referring to?
For my part I have faith that computers are great at memorization, and will continue to improve on that front. However, I'm less convinced on their ability to "understand", which is admittedly poorly defined, but intuitive. It seems to me there's still a missing piece between machine learning (essentially all things that are called "AI" these days) and the kind of generalization that we expect out of even a human 1 year old or your average vertebrate.
1 reply →
I agree with your first paragraph, but your second one is definitely wrong. We aren't anywhere near even understanding "what nature has wrought" the possibility of "catching up" within the foreseeable future is not realistic.
We are a part of nature too. So when humans create AI isn't it still just nature creating AI?
...is strong AI a more sophisticated invention than cells? Is AI more sophisticated than DNA? Is AI a more sophisticated invention than a habitable orb of magma orbiting a nuclear explosion in space? It's all crazy. It's all amazing. ...We are all lucky enough to be here for a little while to see whatever this is. We're a point on a stream.
But let's not forget evolution of the magnitude you speak of took hundreds of millions of years. I'm not sure why humans think we can beat that record.
Because they aren't starting from nothing? They're starting from humans -- the more we can usefully encode about our knowledge of ourselves, the more time we can skip in evolutionary effort. The mutation rate is also rapidly accelerated and although we drop the fidelity in simulation, for the most part we can run magnitudes faster than real-time (and certainly if a limiting factor was human decision making time).
Life evolves at linear rates, while computers at geometric.
Even more to the point the cycle time for human evolution is at a minimum 14 years (biological limit to reproduce) and around 25 years (today's societal constraints). Whereas cycle time for new computer generations is what 12-18 months?
So at best every 14 years we get a new 'human model' with some random variations. Meanwhile in that time frame our computing hardware is 2000x improved (hopefully improvements in software accelerate that improvement further).
I don't think humans can beat the record, but technology? That's another matter all together.
Again our own consciousness wasn't created it emerged which means that we are a product of this emergence.
2 replies →
Flying took billions of years to evolve, but we created machines that fly faster than the speed of sound...
> but have a hard time believing that can be done via computers
Remember that we are also constrained by our own intelligence. Computers might not be the problem.
I am seriously interested in understanding why a lot of people have no issue accepting the idea of humans evolving from simple immaterial matter into biological beings and all the way up to our current reality but have a hard time believing that can be done via computers.
You just described AI in the 90's: Artifical Life, subsumption architecture, evolutionary algorithms. Theoretically we should be able to evolve intelligence but the search space is impossibly high and we don't understand why life is evolvable. Even if we did evolve an AI through a simulation of evolution - there's a small chance we would understand it.
> I am seriously interested in understanding why a lot of people have no issue accepting the idea of humans evolving fromm immaterial matter into biological beings and all the way up to our current reality but have a hard time believing that can be done via computers.
It took nature what like 4 billion years to come up with the human brain in a single species... We including Kurweil think that this can be replicated in say 100, I mean not even that according to the singularity, 40 years. I seriously doubt it. Can the human brain eventually be surpassed, yes, but it's very likely that we will go extinct before then.
It also took billions of years to evolve heavier than air flight, but it took technology much less time than that.
How long did it take evolution to build a structure taller than 100 m?
Technology just moves on different timescales, with different methods and different objectives. Evolution is slow and dumb, and not goal oriented.
In some domains, technology might never catch up to biology. In others we've beaten it soundly.
1 reply →
One reason is that technology has the ability to do trial and error through billions of generations in a very short time.
1 reply →
> all they need is energy
A comment above about being as intelligent as a mouse seems to touch on this problem domain rather elegantly : "navigate the forest floor looking for food and avoid predators".
> the idea of humans evolving from simple immaterial matter into biological beings
We have no evidence that life evolved from inorganic matter (i.e. that life 'began' at some point). That's a widely made assumption but it remains only that. The alternative is that life has always existed (and continually evolved).
The more you learn about Kurzweil and what he bases his predictions on, the more you realize he's a one-trick pony. They only work on things governed by Moore's law (advancing transistor density) and that in turn depends on a variety of things. Moore's law is expected to wind down around 2025.
Also, a lot of what he bases his claims on are unexamined junk science (like his nutty health books, but also extending into specific technologies). Let's not swallow everything he says just because he helped invent OCR. https://en.wikipedia.org/wiki/Ray_Kurzweil#Criticism
You are much too kind. Kurzweil is a loon, full stop. The fact he once made brilliant contributions to computer science is quite irrelevant to the essential craziness of his more recent delusions.
In 2005, Kurzweil published The Singularity Is Near and predicted this would be the state of the world in the year 2030: "Nanobot technology will provide fully immersive, totally convincing virtual reality. Nanobots will take up positions in close physical proximity to every interneuronal connection coming from our senses. If we want to experience real reality, the nanobots just stay in position (in the capillaries) and do nothing. If we want to enter virtual reality, they suppress all of the inputs coming from our actual senses and replace them with the signals that would be appropriate for the virtual environment. Your brain experiences these signals as if they came from your physical body."
That is not happening by the year 2030. It is so starkly delusional that anyone who seriously affirms a belief that it will happen probably needs psychiatric help.
It is akin to Eric Drexler's loony visions back in the 1980s that nanobots would cure all diseases and continually restore our bodies to perfect health. We were supposed to all be immortal by now.
None of this is happening, probably not ever, and certainly not in the lifetime of any human being currently living. Kurzweil is going to die, Drexler is going to die, everybody is going to die. Adopting a pseudo-scientific religion to avoid facing mortality is kind of sad.
>Loon...craziness...delusions...needs psychiatric help.
I agree many of his predictions are bad but you should calm down with the gaslighting, it's ignorant of science history (the same was said of Aristotle, Semmelweis, Wright Brothers...) and is an impotent way of debating, especially in the context of science.
2 replies →
It's interesting to consider the parallels between this stuff and the fountain of youth, or alchemists turning lead into gold. Explorers were constantly uncovering unimaginable new things with no real idea of where it might end. Similarly alchemists who were finding that various combinations of compounds created ever more unimaginable results. So they, too, simply extrapolated outward.
I grew up with Drexler as a sort of hero. It's amazing how rapidly nanotech went from the imminent thing to change all things to, 'huh - what's that supposed to be about?' Wonder if 20 years down the line we might look at AI similarly.
Your last paragraph reminds me of FM-2030. A "transhumanist" born in 1930, he hoped to live until 2030 and wrote that "in 2030 we will be ageless and everyone will have an excellent chance to live forever." He died in 2000 from pancreatic cancer.
https://en.wikipedia.org/wiki/FM-2030
You can differentiate the Moore's law / AI stuff which seems fairly sensible to me and the nanobots and vitamin pill stuff which I've always thought a little nuts. Hans Moravec did a much more down to earth analysis on the Moores/AI stuff if you'd rather avoid Kurzweil.
3 replies →
Hennessy and Patterson said that Moore's law ended in 2015.
The singularity is nigh! This trope might make for great fiction but the on the ground reality is far different. Intelligence is multi-dimensional. No machine intelligence has yet shown an ability to match humans in multifaceted intelligence and the day when such intelligences can outpace humans is as far off as it was when the singularity was first posited.
> No machine intelligence has yet shown an ability to match humans in multifaceted intelligence
This is pretty absurd to point to as an intermediate goalpost; that's basically game over when it happens.
It's not an intermediate goalpost. I'm referring to what essentially we're promised by those decrying AI who insist that it will match and exceed humans in terms of cognitive abilities.
As for it being 'game over', why? Is there something inherent in AI that would necessarily be inimical to human beings?
1 reply →
It's game over when machine intelligence controls all the resources needed to ensure their continuation. As long as they are trapped and subject to our control of the power switch and reproduction they can't do much that we don't want them to.
16 replies →
In Kurzweil's book 'The Singularity it Near' written in 2005 he predicted it for 2045 and has never changed the predicted date so we're still 26 years off. We've recently had AlphaZero beating us at games and Waymo cars kind of driving so there's some progress on the ground. Give it time.
What does multi-dimensional even mean, in the context of intelligence? Sounds like a buzz word.
In the field this is the idea of General Intilligence versus narrow intelligence.
Alexa, for example, is an Artificial Narrow Intelligence. It can process speech and then follow different scripts with instructions derived from that speech, but it often fails comically so you as the human have to talk to it just right for it to work. Not too different from a verbal command line.
Meanwhile a human personal assistant has general intelligence. You can just tell them what you want and they can understand and figure it out.
2 replies →
> What does multi-dimensional even mean, in the context of intelligence?
Emotional
Logical
Intuitive
Social
Computational
Instinctive
Sexual
Experiential
and I can keep going, but you should get the point.
3 replies →
Aren't you a little bit worried ? I mean there has been a long history of "you can't do this with computers".
It started with carrying out list of simple instructions, then doing anything with images, then recognizing images, then drawing images, then simple games chess, recognize audio (though people realized this was stupid in the 60s with IBM speech synthesizers and so today people don't remember), and then Go (and less well known: backgammon, video games, ...). For most of those we now roll our eyes and go "how could they have been so stupid ?".
As for practical applications: robots run something between 3/4ths and 5/6ths of the stock market. Maybe you're unaware of this but that's the thing that decides how and where humans work (not just the ones also doing stocks, anyone working at any public company is partially directed by it, which in practice is nearly everyone).
AIs talk to humans more than humans do. AIs produce more writing than humans do. AIs judge more humans (for insurances, or creditworthiness) than humans do. Despite what you think AIs actually drive just a little under 1 millionth of total miles driven in the US today, going up exponentially. AIs currently drive a little more than a 1000 person city does.
In experimental settings, AIs have beat humans at convincing other people that they're human. At "chatting up" humans. Seriously. Not that it seems to take much at all to convince humans you're sentient: iRobots have convinced soldiers to threaten army technicians with guns into repairing them. There wasn't even that much AI involved, but that's got to count for something.
In research, I would argue there are already multifaceted artificial intelligences in reinforcement learning. There is nothing in those atari game playing AIs about the games. There used to be the score of the game, but the modern ones don't even have that. They can play any atari game, from Montezuma's revenge to pacman, which I'd argue are very different indeed. There must be some measure of "multifaceted" in there, surely ?
But let's keep it simple: could you make this problem a bit more accurate ? For example, which animals would you say exhibit an acceptable level of "multifaceted" intelligence ? Why do those animals qualify ? What would a good test be ? I'd love to find an interesting test for this.
>I mean there has been a long history of "you can't do this with computers".
Historically, people have both grossly overestimated and grossly underestimated future technological progress. Some people thought computers would never be able to play chess; others thought we'd have superhuman AGI in 2001. The bottom line is that people's past and current ignorance tells us nothing about what the future is actually going to be like.
The idea of the singularity is based upon the idea of recursive self improvement. I'm not sure how your claims are relevant.
Recursive self improvement in which field of intelligence? Does getting better and better pattern matching eventually lead to human level intelligence? Does improving pattern matching accelerate our ability at improving pattern matching even?
We can solve any board game now with alpha-zero but will that necessarily improve how fast we develop other types of intelligence?
I used to be in Rays camp but my girlfriend got into Computational Neuroscience and I talked with her and her colleagues got the impression very very few people think we’re close to general intelligence.
Human intelligence is a lot of things put together and we may get better at pieces but we don’t even have all the pieces and when we do try to put them together it doesn’t work. Look at criticisms of Europe’s Himan Brian project[1] (another example I had seems to be outdated), some believe we don’t understand too much to even begin to attempt modeling the brain.
[1] https://en.m.wikipedia.org/wiki/Human_Brain_Project
7 replies →
There are actually a bunch of different ideas that all go under the "singularity" label making the term fairly confusing. You describe one. Another is the idea that if AIs start designing computers you'll have a positive feedback loop in improvements in computer speed. Another is that we can't predict the actions of a smarter-than-human intelligence. And then there's Kurzweil's idea that progress tends to speed up and at some point we call it the Singularity for some reason. I just wanted to point this out because I've seen a number of arguments caused by people using the same words for very different ideas of "the singularity" without realizing it.
I think (human)intelligence almost guarantees recursive self improvement; thus, if AGI = human intelligence or greater, it would also guarantee recursive self improvement.
2 replies →
It’d be a lot easier to believe not coming from a company selling ads. Google would be far more of a self improvement tool if it were not incentivized to sell you things you don’t want.
8 replies →
Despite the cynicism and his black and white predictions, I think his rhetoric still has valuable contributions. It forces others to take a true account of what intelligence is and what kind of intelligence is AI capable of in the medium term (evidence: this thread).
For some reason, this doesn't get enough attention and we have people like Elon and Stephen Hawking making dire predictions all over the place.
I disagree. I think his rhetoric feeds the 'we already understand it' mythology around the nature of AI technologies, and worse the state of our own understanding of the system they claim to model, and directly contributes to the harmful and frustrating boom-and-bust cycles that AI goes through. I'd lump him in the same category as Elon Musk and Stephen Hawking in this space; sellers of fantasy.
I admire his optimism, but I think it's irresponsible to sell it like he does.
Marvin Minsky ("Father" of AI at MIT) made the bold claim in 70s that we will be marrying robots by 2000. After it didn't come to fruition, it led most AI researchers to take stock of the situation and realize these goals aren't that trivial.
Similarly, if we don't have bold predictions like these that we can actually measure within our lifetimes, we fall prey to fantasies that cannot be measured. Once this prediction fails miserably (I think) it helps many others to re-calibrate all their BS.
4 replies →
I can't believe there are people genuinely afraid of a hypothetical powerful malevolent AI, yet seemingly not that concerned by actual climate change.
> I can't believe there are people genuinely afraid of a hypothetical powerful malevolent AI
I don't think even the AI doomsayers, deep down, actually believe what they preach. It's just a way to signal that one is clever and informed of new tech.
If they actually believed what they say, they'd be worried of being targeted by violent protestors, like drug testing companies and crop breeding companies have to be.
Who are the people who both predict super human intelligent AI and do not believe in human caused climate change?
1 reply →
Who says you can only worry about one thing?
8 replies →
I've felt for a long time that Singularitarians are looking in the wrong place. They see accelerated technological development and assume that the endpoint will be an artificial brain in a box. What they fail to see is that these inventions and breakthroughs haven't been about increasing the intelligence of a machine... they've been about increasing the intelligence and efficiency of human systems.
The singularity isn't a brain in a box... it's us, the collective, a metasystem transition that's been underway for millennia. A movement toward a whole that transcends the parts.
That's what seems apparent to me, so far anyway. AI is more about augmenting human intelligence than it is about smart machines. All that impressive DL stuff is ultimately providing more tools for humans to be more productive.
Can we please as least try to steer it towards https://en.wikipedia.org/wiki/As_We_May_Think ?
Here's my shorter-term thinking. Current ML can't generalize well [0], can't do arbitrary & conceptual thinking, and has trouble with language.
[0] by that, I mean that it does not easily pick up higher-level patterns unless explicitly forced to (through either model architecture, data & task setup, etc). Meaning - it has a much lower success rate outside of the training data distribution compared to "flesh" intelligence. Kind of reminds me of https://www.youtube.com/watch?v=PHRvF0m3yuo
Hmm, current ML has only been worked on for a few years, we have barely scratched the surface. No signs of slowing down afaict. It has wiped the state of the art against algorithms we took decades to think up. What happens after decades of that?
It feels more like there is too many new directions to explore and we don't have enough time/creativity/insights/ideas to fully take advantage of it.
1957 - First ever man-made object launched into space.
1957 - First ever animal launched into space.
1960 - First ever animal launched into space that survives the trip.
1961 - First man goes into space.
1965 - First man goes on a space walk.
1966 - First spacecraft lands on the moon.
1967 - A probe is landed on Venus (!) which collects and sends back video and audio recordings among other data.
1968 - First spacecraft to orbit the moon and then return to Earth.
1969 - We put men on the moon who then safely return.
We've done all of this in just 12 years! We literally went from nothing to men on the moon in 12 years! We've barely scratched the surface of what's possible and there's no sign of things slowing down at all! Can you imagine where we'll be after decades of this. Where will we be in 50 years? I can't even begin to imagine what 2019 will be like.
I don't say this to be snarky, but to point out the issue. Exponential growth doesn't mean continued exponential growth, and it's always very difficult to predict the future. While thoughts such as the above were common place, there would have been hardly a person to predict that the room-sized government "computing machines" would one day be a few inches large, millions of times more powerful, and cost as little as a week or so of minimum wage work. In some cases, those exponents did keep churning for a while.
On AI, I used to be a futurist in terms of it - until I got to work with it. Now I'd be extremely surprised if we have have fully capable self driving vehicle within the next decade that's not constrained to white-listed routes. And that is an extremely softball problem since it fits perfectly into the small domain of problems that current ML/AI systems are really great at making progress on. The thing is getting to a testable model is absolutely trivial. Going that remaining 10% from model to product is many orders of magnitude more challenging.
Current ML algorithms have been worked on for decades earlier. They are gaining momentum again due to big data and computing power.
1 reply →
ML was covered in my 1972 AI course at MIT.
1 reply →
>Here's my shorter-term thinking. Current ML can't generalize well [0], can't do arbitrary & conceptual thinking, and has trouble with language.
To be honest a lot of humans struggle with the same problems.
I had a chance to interview Ray and spend some time with him before he joined Google.
The way we met was very serendipitous -- I asked a question during a movie premiere about the nature of reality. He must have enjoyed it because later someone from his team sought me out to invite me to a VIP after party.
His publicist at the time said this was the best interview he ever gave (It's possible she says that to all the girls, but, I'll take it!)
https://www.huffingtonpost.com/anthony-adams/ray-kurzweil-in...
It's much easier to make a plane that can fly than to emulate the particular way a bird flies.
They'll both solve the flight problem, so it doesn't really matter.
Flight is something extremely comparible to ai development. When it was developed, many companies were trying, and a lot of people said it was impossible. The problem space is also similar. We're not sure how to get there exactly but we may be close.
It was, in the end made to happen by an unlikely pair - not a large company with a lot of investment. I believe this might be that fate, too.
It matters tremendously. Think about the economics of bird flight vs. plane flight. Why is it that flight has been around for almost 100 years now and yet we still aren't flying everywhere? The reality is we can always come up with subpar ways of doing things, but there's a reason birds have evolved as they are today that we just can't replicate ourselves. We mastered long haul rapid flights, but bird flight generalizes much better. They might be slower in long haul, but they are much better and short haul and medium haul and have evolved to work in groups to make long haul possible. So yes, analogously we can create intelligence in a different way, which we are, but it likely won't generalize as well.
If we could reduce the density of a person by about an order of magnitude we can start replicating bird flight on a per-person basis. Our approaches to flight are constrained in all forms by the desire to put really heavy things into the air.
Look no further than drones that can operate for quite a long while, move in any dimension fairly well, and manage some pretty ridiculous speeds for not being designed to do so all because their purpose is not to carry large complicated amounts of weight.
The difference is we knew what flight was and had an easily visible and verifiable goal that was obvious to all when it was achieved. What is intelligence? How do you measure it? How will we know once we have created? We haven't even defined the problem yet, how can we know when we have achieved it?
> in the end made to happen by an unlikely pair
Who spent years studying birds...
No, the wings of the wright flyer don't flap, but that doesn't mean it wasn't designed based on birds. One key piece (controllability) came as a result of the wright brothers studying how birds adjusted the shape of their wings.
Question is, are the current approaches at AI the equivalent of the early flappy wing planes we were building to emulate birds?
Brave claim: we won’t actually invent good AI until two-way brain-computer interfaces become useful.
We need to augment human intelligence to handle systems vastly more complex than we can do today. Specifically memory creation and recall, and information assimilation rates.
How do you define 'useful'?
My keyboard and monitor are brain-computer interfaces. I'd argue that they are pretty useful.
I can imagine a way to consume information that is dramatically faster than the 500 or some odd words per minute I get reading.
But for input to the computer? I mean, I'm not saying that a keyboard can't be improved upon, but I think there are some hard limits to how quickly I can compose my thoughts; I think that it's fairly rare that the fingers are the limit. (Of course, making it so I think less about the fingers would be great)
What I'm saying is that there might be human processing speed issues that limit how much value we can get out of faster I/O.
On the other hand, maybe it's like reading was for me; before I could read at a certain speed and certain ease, I couldn't enjoy books, because I'd forget the beginning of the paragraph before I got to the end. Like I needed to load enough words at once to see the picture. It's totally possible that output would be the same way; if I could somehow output the equivalent of five hundred plus words per minute, maybe that would change my writing the same way passing that speed changed my reading?
I think most of the bandwidth would be computer -> brain, not the other way around, assuming it can teach your brain to retain knowledge and show it images/movies. The only exception would be if we could significantly alter the perception of time, but I'm not really betting on that despite how useful it'd be.
Even if I could only accurately sample 1-8 bits/second from the brain, that would be world-changing. There are a lot of clever people who could stuff a lot of utility through 8 bits/second given you are paired with a powerful smartphone and a proper "UX". Clever encoding schemes and signals could be developed. Especially if you go for sequences/chords that are executed over 1-4 seconds.
Plus the brain would likely just act as an index to fetch info that is then replayed to you. I really don't need a terabyte in the form of neuron connections when storage and computing would be ubiquitous.
3 replies →
So brave.
IDK, Google's voice assistant thing still can't figure out my wife's name because it's not an English name and I pronounce it as it should be pronounced.
Our receptionist can't figure out any of the names of our offshore contractors either.
lack of ability and lack of caring are not the same.
Yet.
What if human beings were physically connected to a growing network of perfectly organized facts and information? What if that network could even assist with basic logical reasoning tasks? Would that be enough to say we’ve reached the singularity?
It doesn’t seem like it now because our phones aren’t quite physically connected to our brains, and the internet is far from organized, but I’d venture to say we’re already pretty close. A more interesting question to me is whether this is really anything we want.
If human machines were infinitely intelligent then they would all agree. So you wouldn’t need more than one. Sure, the ultimate human might need 7 billion pairs of eyes but those eyes wouldn’t need any more brains than to fuel up and locomote. Furthermore any human-like tendency to doubt oneself or fall in love or “rock out” would be deemed useless and therefore overridden.
Understanding that the purpose of a human is to pass on genes, and that there’s little human or genes left in her, the ultimate human might therefore conclude that her sustenance only disrupts the purposes of organic life forms. Her next and final act would be to destroy herself.
This is how I see the end of humanity. Not in a violent clash between artificial and organic intellects, not in a symphony of mushroom clouds, but with the final flip of a switch, in cold, calculated resignation.
Check out Rodney Brooks' blog to see some predictions from somebody who actually has at least some idea of what they're talking about:
https://rodneybrooks.com/predictions-scorecard-2019-january-...
However as the ML researcher Michael Jordan (one of the most important in the field) has previously stated, these sort of long-term technology predictions are just fun science fiction and there is essentially no academic rigor in this stuff:
https://medium.com/@mijordan3/artificial-intelligence-the-re...
Counterpoint: "Hey Siri, lock my phone." "Playing 'Locked Out of Heaven' by Bruno Mars..."
Betting on technology evolving faster than expected is usually a safe bet, but it feels like Ray has been working on flawed premises and refuses to revise them.
What, specifically, has he gotten wrong?
This one (especially HN commentary) is a good read: https://news.ycombinator.com/item?id=18806315
After reading it all, I don't really see him as being any more successful in predicting than you and me. You get some right, you get some kind-of-in-the-right-direction, you get some wrong, and some are just laughably wrong.
A better question is what has he gotten right?
As we do not possess a clear (or, at least, a generally accepted) definition of what constitutes the "true AI," it seems very likely that we will cross this "even horizon" without realizing it; we may have done it already, for all we know (or, rather, don't know - due to the fact that much of the progress is being done in secrecy).
On pace to do what? Add numbers really fast? Play chess? Make a best guess of which news story I want to see? (Hint: not sports) Perform automatic gear shifting? Follow the vehicle in front of it?
Kurzweil has long been preaching that there will be a qualitative revolution in what AI will achieve - and it has been 15 years away, for the past three decades.
Not actually true, the always 15 years away bit. Three decades ago (1990) his prediction was Turing test and AGI in 2020–2050. More recently he's predicted Turning test 2029. He may be wrong but doesn't keep shifting the dates 15 years away. (https://en.wikipedia.org/wiki/Predictions_made_by_Ray_Kurzwe... )
Ray Kurtzweil is certainly super smart, but I find his views about the Singularity disturbingly myopic. It's as if he has been living in a bubble for decades, and doesn't realize how the world works for real.
I still have a lot of sympathy for him, as I'm sure he means good. Still, these views distort our reality for many people.
Of course it will happen. We humans are an outcome that took billions of years and yet here we are. It might take a billion more to get sentient AI but as long there is progress then the end is inevitable.
The only debate is when it'll happen, but we probably need an advanced sentient AI to help figure that out.
I suspect human intelligence is more than just computations in the brain. A lot of computation I assume most occur 'outside' of the brain in the gut via hormones and protein signalling. I have doubts that you can replicate human intelligence without replicating human biology.
I doubt a lot of computation occurs outside the brain. I'm assuming most of the neurons, nerves, etc in the gut are used primarily for regulating gut health rather than calculating, reasoning, etc.
People have had their intestines completed removed, liver transplants, kidney transplants, stomach stapled, etc without hurting their intelligence or computational ability. I don't think we can replace the brain like we can elements of the gut.
Certainly, the gut, hormones, etc are important to your physical and emotional well being, but I think it's safe to assume that pretty much all over the computation is happening in our brain.
Or even if it were all in the brain, accurately modeling neurons and interactions of neurons, on bajillions of cudas and threadrippers, will probably still take much longer than it takes for the real-life system to do its thing:
https://newatlas.com/spinnaker-neuromorphic-supercomputer-mo...
Modeling the human brain is just one of the directions of the AI research, possibly not the most important one, even, at this point. On the other hand, the "super-intelligence" everyone is thinking about will be nothing like the human mind.
1 reply →
We also have 500 million neurons in our gut! https://en.wikipedia.org/wiki/Enteric_nervous_system
I wonder why wikipedia doesn't have more info about the function of this system... perhaps it isn't well understood yet.
> hormones and protein signalling
They are control utilities, regulators for the brain, whose generation is initiated by the brain. However I think you are right with the need for human body. Without sensory input its state would collapse, I think it needs a world with constant stimuli to live in.
https://blogs.scientificamerican.com/guest-blog/a-brief-guid...
I think we will find limited use for general AI since we won't be able to properly regulate or understand the behavior of machines that are smarter than us. Alpha Go Zero can't explain why it plays chess the way it does except for showing how it evaluates the board, but the reasoning is buried in millions of mathematical calculations accreted via millions of games of self play that we can't understand the intuition of. Of course, many smart players can't really tell you systematically how they make the moves they do either. AI will be like a dictator who makes his decisions by intuition only.
An intelligent black box can still be "useful". It could help cure diseases or get original insights in many fields for example. Even if we don't understand how it thinks.
Sure, but these are narrow AI applications. The great and powerful Oz general AI will be of limited utility unless there is a man behind the curtain.
1 reply →
Counterpoint: https://engineering.purdue.edu/kak/RobotsWillNeverHaveSex.pd...
Gave up at page five of "this is what this presentation is about", can you summarize his main argument on why robots will never have sex?
It's clickbait. The last page is:
>But What About the Title of this Presentation .... While obviously a cheap hook, it is nevertheless intended to convey the possibility that it is our emotions, our passions, our innate desires — all ingredients of our sexuality — that are the defining elements of our consciousness and, through consciousness, our intelligence. The End
1 reply →
Ray Kurzweil is still on pace to die before the Singularity arrives.
I always felt that Kurzweil’s timelines were too optimistic not only because technology may not evolve so fast, but also because it fails to account for the delays that government regulation imposes. So many of the vaunted leaps in medical technology, for example, are going to require years and years of testing before they get approval for use. Other whizbang technology of the future might be prohibited from the consumer market because governments see it as too dangerous.
This may be the case for many technologies, but clear wins that save lives get fast tracked. It just happened to CTX001. Also older researchers are doing self testing all the time, as they know that they don't have the time to go through regulations.
https://www.investors.com/news/technology/crispr-vertex-gene...
Not if he eats any more pills
I think Moravec's arguments on the same subject "When will computer hardware match the human brain?" are a little less cranky than Kurzweil's and his prediction from 1998 "it is predicted that the required hardware will be available in cheap machines in the 2020s" seems to be panning out. https://jetpress.org/volume1/moravec.htm
The Singularity is just millennialism [0] for secular technolgists. We still have a pretty paltry understanding of human intelligence. It seems unlikely to me that more and more iterations of AlphaGo will spontaneously produce strong general AI. My bet is no AGI within my lifetime.
[0] https://en.wikipedia.org/wiki/Millennialism
Alpha zero is a generalized AlphaGo, so they are headed in the right direction!
AlphaGo -> win this boardgame
AlphaZero -> win any boardgame (well, three right now)
AlphaMinus1 -> win any game
AlphaMinus2 -> win anything
AlphaMinus3 -> win winning. so much winning.
But you get my drift, I'm extrapolating hugely off one abstraction step.
Yeah maybe. How do you define "win" in terms of general intelligence? Outsmart all top human experts at their field?
3 replies →
it depends on how old you are, but I think you may be right that AGI still needs on the order of hundreds of years to happen.
The brain and "mind" is insanely complex. There is so much we still don't understand, so of course there is so much we cannot replicate. But is it possible? If there is an incredible breakthrough in understanding, then yes of course. Personally I see little value in people saying it can or can't be done (of course it can!), it's just a matter of if and when. We shall see...
Define AI. Define intelligence. Kurzwiel's prediction are nonscientific and unfalsifiable. AGI is unscientific.
He really likes making predictions: https://en.wikipedia.org/wiki/Predictions_made_by_Ray_Kurzwe...
He predicted Turing test by 2029 which is fairly definite, falsifiable and quite likely to be falsified.
For me proper AGI is when you could say to a robot go build some better robots and then have them build even better ones and they can do so without needing us. There will be some year before which if humans disappeared, computers and robots would grind to a halt and after which they could carry on. That's also a definite falsifiable thing, harder than the Turning test and is kind of what I think of as corresponding to the singularity which tends to be rather wooly in it's definition.
I don't think this one happened:
2019: "The computational capacity of a $4,000 computing device (in 1999 dollars) is approximately equal to the computational capability of the human brain (20 quadrillion calculations per second)."
$4,000 in 1999 would be about $6,000 today. A midrange Deep Learning workstation is in the ballpark, and while it can't "match the computational capacity of the human brain", it can do plenty of nifty things - including beat the current Go world champion in a match.
1 reply →
“Kurzweil’s Phantasms”: https://www.firstthings.com/article/2013/04/kurzweils-phanta...
I don’t find this any different than a steam shovel outpacing a human digger. It’s still necessary to prevent both from destroying things and harming people.
Yawn. Of course he has to say this, he makes his living saying stuff like this. He is, after all, a ”futurist”. It just seems rather trite at this point.
AI is mechanical turks all the way down and always will be.
Then humans are as well, yes? What exactly is the magic in human cognition that is impossible to replicate in silicon?
The brain is denser, more efficient and 3 dimensional. Compared to the current limitations of silicon the combination of those three aspects is downright magical.
Ray Kurzweil's brain has already been backed up in the Google Cloud Platform. Everyone claiming he's running out of time is just wrong!
The most important part of the talk was when Kurzweil mentioned running partially-trained AIs in simulation to generate more training data than exists in the real world. That was the first clear description of imagination that I've seen, although others have hinted at it, for example the running stick figures (that I think Google's DeepMind) trained to run an obstacle course several times before they actually did it. I might be remembering that wrong, but there have been a few similar examples.
Anyway, the reason why the imagination part is important is that nobody today is talking about parallelization when it comes to AI. A few "rules" of AI computing and their limits:
* Processing power doubles every year or two for the same cost -> processing power will someday be approximately infinite (for certain classes of computation)
* The search space for even the simplest problems is effectively infinite if we don't know the hidden models -> if we know all the hidden models, then large search spaces become tractable
* The sequential portion of intelligent decision making is less complex than the parallel portion (computers have been beating us at all sequential operations since perhaps the mid 60s or 70s) -> when the parallel portions are solved, then computers will beat us at all intelligent decision making
That last point is really important and exciting, because most of us simply aren't accustomed to thinking in terms of solving problems in parallel. We're thinking "how do we train a self-driving AI to recognize pedestrians and swerve to avoid them" instead of "how do we optimize the solution to the series of equations to drive to the store and back without hitting anything, in a single pass".
AIs will soon be looking forward from their current model of the world, running some number of simulations (potentially billions) and choosing the best outcomes, then feeding back those solutions to improve their hidden models of the world. All in parallel, with virtually unlimited computing power. At that point we'll be training AIs like children and will start to see emergent behavior that mimics life.
Personally, I don't see how this can possibly be any more than 10 years away, 20 at most (even for rank amateurs like me), given the number of people tinkering on these evolutionary patterns, and the amount of open source research being generated.
It kind of haunts me actually. Like, we just started investing in our 401k at work, but I don't have the heart to tell everyone that the singularity will probably arrive before their investments mature. Like, how does one have a child when it's looking like the 20th century reality we're living in isn't going to be around much longer? Do we suffer through the daily grind, the 40 hour weeks of converting our time to money, when AGI will overshoot that within weeks of surpassing our definition of sentience? It literally might make more sense to become a beach bum, or travel the world, go to Amsterdam and spend a few years in a drug-induced haze.
Or is the correct answer to keep going through the motions of running the rat race, knowing that it's all a waste of time but constructing some kind of inner monologue to distract us from the depression induced by the real world's unfolding existential nightmare? Is there a way to un-take the red pill and go back to the blue pill?
I guess I digress, but this is literally all I think about anymore. The futility of work/culture/technology in a world of ever-increasing human subterfuge. Why can't we admit that the dystopia is playing out before our eyes and begin to escape the yoke?
I guess the one thing that keeps me going is that we might be able to break down a single day into its component parts and automate away anything that's beneath human potential so that we can do the stuff we'd like to do (like nothing). The Roomba can vacuum (done). The self-driving Uber can take us to work and back (almost done). The rooftop solar panels can pay our electric bill (effectively done). And so on, until there is no minutia left.
The last step will be to spin up a partial (of oneself) to go to work and earn the paycheck, or literally grow the food to feed us so we no longer have any external dependencies. This is the part that I think is roughly 10 years away.
His predictions are hit and miss. From his 1999 predictions: https://en.wikipedia.org/wiki/Predictions_made_by_Ray_Kurzwe...
By 2009:
Y: The majority of reading is done on displays rather than paper
N: Most text will be created using speech recognition technology.
N: Intelligent roads and driverless cars are in use
N: People use personal computers the size of rings, pins, credit cards
Y/N (premature): Though desktop PCs are still common for data storage, individuals primarily use portable devices for their computer-related tasks.
Y/N (premature): Personal worn computers provide monitoring of body functions, automated identity and directions for navigation.
N: A $1,000 computer can perform a trillion calculations per second.
... seemed a total mixed bag, many 2009 predictions just beginning to coming true today
By 2019:
N: The computational capacity of a $6,000 computing device is approximately equal to the computational capability of the human brain (claims that is ~20 petaFLOPS. Nvidia DGX-2 is 2 petaFLOPS "AI" supercomputer for $399,000). [
N: Computers are embedded everywhere in the environment (inside of furniture, jewelry, walls, clothing, etc.).
Y(easy win): Most people own more than one PC
N: Cables connecting computers and peripherals have almost completely disappeared.
Y/N: People communicate with their computers via two-way speech and gestures instead of with keyboards. (can but in my home and my office most words entered are not)
N: Most business transactions or information inquiries involve dealing with a simulated person
N: Rotating computer hard drives are no longer used.
N: Three-dimensional nanotube lattices are the dominant computing substrate.
N: The algorithms that allow the relatively small genetic code of the brain to construct a much more complex organ are being transferred into computer neural nets.(Could be decades off. Need several breakthroughs to even put this on the horizon)
N: Most roads now have automated driving systems—networks that allow computer-controlled automobiles to safely navigate.
N: Most decisions made by humans involve consultation with machine intelligence. For example, a doctor may seek the advice of a digital assistant.
... few hit but we're mostly heading that way. To me seems 5-15 years premature
[1] https://www.popsci.com/intel-teraflop-chip#page-2
>Most people own more than one PC (easy win)
What's interesting is that I think most people own just one general purpose computing device these days, their cellphone. Yeah, they maybe also have a TV that can play netflix or a game console, but my observation is that most people don't even have one PC in the old sense of "a general purpose computer with a physical keyboard that you use by sitting down" - and if you exclude laptops from the definition it's even more rare.
> Cables connecting computers and peripherals have almost completely disappeared.
This one is unfortunately happening right on schedule, at least if you use Apple hardware. "A $1,000 computer can perform a trillion calculations per second" should be considered "premature" as well, seeing as current GPGPU hardware can go well beyond 1 TFLOPS.
As with most technologists and futurists, he fails to grasp that things take time, lots of time, to make meaningful progress and adoption.
are people still listening to this guy?
Why not?
Seems so.
oh Ray... how about you stop waving your hands, sit down and shut up...
And if you don’t agree you just don’t understand exponential growth.
cant tell if downvotes are people who don't understand exponential growth or people who think I don't understand exponential growth
I didn't vote either way, but long term exponential growth in anything is basically impossible.
Think about it for a minute. At some point lithography gave us Moore's "law" (which, FWIIW has arguably already failed). Do you think that can continue forever? Pretty sure you exceed computational capacity of the universe pretty quickly. Exponential growth in cell phone sales happened at some point as well. As did exponential growth in transportation speed.
The idea that we've had exponential growth in anything but lithography technology in recent years is also laughably insane. Do you think programming languages are exponentially better than they were 30 years ago? Do people live exponentially longer? Cars, autos exponentially cheaper or better in some way? Are machine learning algos exponentially better? No, no, no, and no.
4 replies →
I don't think anything like the singularity will ever be possible, because the ability to speak language for oneself and invent complex reasoning is spiritual, not merely physical. Humans can do it because we are inherently spiritual beings, a union of body and spirit. (I know most folks here don't subscribe to that assertion... I'm not here to argue it anyway.)
But even if that isn't true, humans are not merely the product of our biology. Each of us has undergone an $age-year period of training, 18 years of which was on someone else's effort and time. I doubt someone is willing to put in that much personal investment to raise a neural net like his/her own child.
Your unusual argument recalls a discussion I once had with members of the Go club at my alma mater, nearly 10 years ago. On the question of whether Go would ever be mastered by a computer, one asserted that the game needed to be understood at a spiritual level - not simply as a computational problem - and that’s why computers would never reach the level of humans. The game was simply too sublime and complex, came the claim.
At the time this might have seemed reasonable - after all, Go software struggled mightily for years to achieve even beginner levels of play, and seemed incapable of mastering the intricacies of high-level play.
Well, we know now in retrospect that spirit was unnecessary to play Go. We know now that it was, in fact, a computational problem. There’s really no reason to expect that “spirit” is necessary to perform the tasks of language or cognition - tasks which are demonstrably executed by impulses of chemicals and electricity pulsing through the brain.
I know spiritual reality is sometimes invoked in discussions about things we haven't yet mastered or understood in science and engineering (one might even accuse me of doing that). But your story is kind of a straw man example, because even 50 years ago we knew of algorithms that could theoretically render Go a solved game, given enough computing resources. You could just as easily say modern cryptography is spiritual in its essence, which would be nonsensical.
Emulating a human mind with computers is an entirely different class of problem. We don't even have a mathematically rigorous definition of what constitutes a mind, much less know of any algorithm that could implement or solve it, even with any given supply of computing resources.
3 replies →
> Each of us has undergone an $age-year period of training, 18 years of which was on someone else's effort and time. I doubt someone is willing to put in that much personal investment to raise a neural net like his/her own child.
Unlike humans, software is reproducable and doesn't expire. So even if such a software took 18 years to develop, it is still a worthwhile endeavor since it has the potential to be used by countless people well into the future. Furthermore it is not necessary that it is developed by just one person, nor is it necessary that it is developed at the same speed that humans develop. The creation of such a software could take 100 years and it would still be worthwhile.
But if something is spiritual, there must be some way for it to interact with the physical world, otherwise you would not know about it. Then that way can be measured, which makes it physical...
Unless you are calling spiritual those things that can't be measured, yet?
For the sake of argument, "spiritual" could be a class of physical objects that can only be produced by other spiritual objects. This is the old "vitalism" argument, which is generally considered discredited, but as the "hard problem of consciousness" is still unanswered, this is a small loophole that proponents can use to hang on to it.
2 replies →
> there must be some way for it to interact with the physical world
that's the body
We humans are blackbox physical machines that processes physical inputs to produce physical outputs.
Whether that input goes through processing that is "spiritual" is besides the point. Reproducing "spirituality" is not the goal. The goal is to replicate what is "observable." To produce a machine that when given the same inputs as that given to a human will produce identical (or "better") outputs as the human.
This is definitely an achievable goal as for every set of inputs and output pairs in existence an infinite amount of different machines can be constructed to replicate that exact input/output functionality.
The "spirituality" that goes on within this blackbox machine or humans is irrelevant. It's not even a word that is clearly defined. We aren't trying to replicate "spirituality." We are trying to make a machine that only from an observational standpoint is clearly superior or indistinguishable to a human intelligence.
The understanding the theoretical possibility of singularity is purely within the grasp of a typical person.
Imagine a chat bot designed to be indistinguishable from a human chatter. The inputs are text queries and the outputs are text queries. The domain is every possible text query and the range is every possible text response.
For purposes of simplicity lets say the text input and output is limited to 500 chars. This makes the possible input and output pairs finite. To create a chat bot identical to a human you simply need to create a machine that maps all inputs to certain outputs. Think of it like a giant hash map of strings to strings. Given enough time and enough memory this can be done.
The goal of current AI is to solve this problem more efficiently. To produce logical cores that can dynamically generate outputs from inputs rather than following a gigantic mapping. It can be thought of as form of data compression. The data is the mapping; the compression of that data is the logical core.
The fact that a mapping can even be constructed at all is by logical induction, a proof that a singularity is possible and thus completely invalidating your initial statement of:
>I don't think anything like the singularity will ever be possible
Only that the chat response is not just function of the 500 character query but also of all previous queries:
You could ask the bot things like "Did I ever tell you about [...]?" or "What was the last time we talked about X?"
2 replies →
Raising a kid without having to change diapers sounds pretty cool.
But a fan will be there for sure.
hardware - the body.
software - the mind aka the spirit.
hardware together with software is the runtime
the body with the mind make a human person, maybe that's the soul?? ️
[flagged]
I love Ray Kurzweil and he's an undisputed genius, but like many other believers in The Singularity, he's convinced of something that has never been proven to be true.
No computer will ever demonstrate actual intelligence, which is the ability to bring a new and unforeseen solution to an existing problem. I need only a cite a few examples to show this is true: Uber cars run over people. Facial recognition claims black people are criminals just for being black. Facebook's AI software lets in fake news.
These are all errors that could be avoided by actual human intelligence. The fact of the matter is that we don't really know where intelligence or creativity comes from. It's unlikely that we ever will. (See "Heisenberg Uncertainty Principle" for one reason.) While machines are great at doing things faster than people can do them, they're not good at inventing new solutions or doing something that humans couldn't do, given enough time.
> No computer will ever demonstrate actual intelligence, which is the ability to bring a new and unforeseen solution to an existing problem. I need only a cite a few examples to show this is true: Uber cars run over people. Facial recognition claims black people are criminals just for being black. Facebook's AI software lets in fake news.
Burden of proof is not on your claim, but rather on Kurzweil's, but since you made a claim and tried to back it up, I feel compelled to point out that it's a non-sequitur. Just because something hasn't been done before, doesn't mean it's not possible. Many people said going to the moon was impossible. Might be interesting to look at Russell's Teapot.
> doesn't mean it's not possible
In general, that which nature has demonstrated is can usually be replicated. A bird (flight), a floating log (ships), a fish (submarine), an asteroid (space travel), etc. Nature has demonstrated intelligence: a human.
However, just like nature has not demonstrated superluminal travel, it has not demonstrated super-intelligence; so that is still a question.
8 replies →
Just FYI, going to the moon is also impossible. See "Van Allen Belts."
> No computer will ever demonstrate actual intelligence, which is the ability to bring a new and unforeseen solution to an existing problem
It's easy to disprove absolute claims such as yours. Yes AI can create new things and it's only getting better. You just need 1 example. Read up on AlphaGO. It was able to beat humans by coming up with techniques that no human has ever and possibly would never have come up with because of how illogical they seem to even the masters of Go.
Look into GANS. AI can create new things.
Contemporary AIs don't create new things via creative insight or intuitiveness. A trained NN is only good at the one thing it has been trained at. That isn't true intelligence, no matter how astounding the results.
4 replies →
The means to do those things were programmed by humans. While the computer is faster at assessing all of the possibilities, it isn't coming up with anything that wasn't made by possible by the human-programmed software.
We now have to have safety belts for Amazon workers to keep them from being run over by robots!
1 reply →
"Ever" is a very long time. But I'd be comfortable saying AGI won't arrive in my lifetime, at least not electronic AGI. Note that I single out electronic AGI here due to both our complete lack of understanding of how to model intelligence, and the prohibitive energy cost of modelling the brain if we did understand how to do so.
OTOH advances in genome understanding could yield e.g. a super smart chimp, or dolphin, something like that. That would also be "artificial" intelligence then, and it would be general, just not electronic.
Unfortunately we are still vastly better at making a worse mouse (or cow) than a better one, to paraphrase biologist Anne McLaren. Sure, evolution did eventually make a super smart chimp, but it took 7 million years of tinkering.
1 reply →
haha. maybe no computer by the current definition of computer. but I would say that it's a bit extreme to say humans are going to be the only higher intelligence species.
the only thing that I think we should have disagreement on is if we want it to happen and of course the timeline.
for the if question: I believe we (humans) want to do this. Long term it is going to basically mean immortality and I would not think about intelligent machines as a new rival species, I would think of it as an evolution of our species - it's actually a great way of transcending our biological limitations.
for the timeline piece: I don't think it's going to happen within our lifetimes. I also think it's going to be a process where for a very long period of time we will be hybrids (we will augment ourselves with hardware that will make us smarter/more capable and this in turn will lead to us building even better and more comprehensive hardware. At some point we'll figure out that the biological part is just getting in the way and we'll completely remove it).
>No computer will ever demonstrate actual intelligence, which is the ability to bring a new and unforeseen solution to an existing problem.
This is already false, given the many examples of machine learning out performing humans in a variety of tasks. Google's data center efficiency comes to mind (IIRC saving 30% in energy usage over human-found solutions). So does AlphaGO.
Outperforming in these cases just means "doing it faster," as I said. There is no algorithm which can produce results which were not inherent in the algorithm itself. There is no software which can "jump the box" and come up with a solution which was not already present in the software as it was written. The computer is merely processing all the options faster than a human could.
When we say that software "found" something new, it's only that it found it faster than we would, using the algorithm we wrote.
5 replies →
> These are all errors that could be avoided by actual human intelligence.
Do you think that dogs are intelligent? Can you imagine some situation where the dog intelligence is not enough to solve it?
Unlike dogs intelligence, the intelligence of computers is improving. Now they can play chess and go like a grand master, but 30 years ago it was an impossible task.
The problems you cite are real, but sometimes people fail at them too (drivers regularly run over people, witness misidentify suspect because they have a skin color that is not the skin color of most of the people they see, people actually believe and like fake news). Just give some years until the interface with the real world is improved and computers have more "street" knowledge to apply in difficult situations.
> See "Heisenberg Uncertainty Principle" for one reason.
I have now idea about how this is related.
It's related because it says that the observer cannot escape the observation; he's creating it. We cannot objectively assess our intelligence or creativity because we are the ones observing them. If we can't understand or assess how they came to be, we can't replicate them.
1 reply →
Humans make all those same mistakes.
Exactly. Software inherits the limits of the people who program it. It doesn't expand upon or escape them. Actual human intelligence, in contrast, can "jump the box" and come up with something that didn't exist in the original materials as presented to the person.
> Uber cars run over people. Facial recognition claims black people are criminals just for being black. Facebook's AI software lets in fake news.
Humans do all of those things too. We might not believe they are the right things to do, but doing them certainly doesn't rule out the presence of a human-like intelligence.
In fact, it does. It shows that human bias (facial recognition) and the inability to posit all possible outcomes (Uber; halting problem) are built-in to the software we write.
While another human may be able to overcome one's racial basis or another driver wouldn't have run over that biker, the fact remains that the software did which shows that it inherits human frailties and limitations. It doesn't surpass them.
1 reply →
> No computer will ever demonstrate actual intelligence
Such statements are impossible to be proven right, as, assuming you consider yourself actually intelligent, that implies we are not inside a computer ourselves, which I'd argue is impossible to prove.
If we're running in a simulation, the host computation hardware may be different from what you imagine a computer to be. For example, we could be an experiment of "is life possible in a euclidean universe".
1 reply →
> that implies we are not inside a computer ourselves, which I'd argue is impossible to prove.
Perhaps not. But some statements cannot be disproven and yet still should not be taken seriously.
But I think you are right that, while correct in claiming that the Singularity has never been proven, mimixco is also making statements that cannot be proven.