Dr. Spork is on to something here. We still know so very little about our universe and by far the best way forward is to develop very large aperture telescopes.
Humans are headed toward building truly massive optical telescopes (100m scale this century, maybe 1km scale in the 2100s or 2200s). Right now we need to succeed in getting the James Webb 6.5m telescope in space and the 30m earth instruments operational. There are several ideas about how to build 100m scale optical telescopes. (See the OWL proposal, http://www.gemini.edu/science/...) One path is to use many independent small telescopes and combine the light as is done with interferometers. Maybe our photonics expertise will advance faster than our large scale construction expertise, and we will do all huge telescopes with many mass produced smaller ones. Maybe their relative positioning doesn't even need to be super stable because we can measure it accurately enough and correct for their relative motion. But it is fun to imagine a 1km scale truss system to hold a massive segmented mirror in space. Picture 250,000 mirrors, each 2m in diameter. It will not remove all stresses to be in space. At that scale, you may start to get tidal stresses and thermal stresses that are almost as hard to engineer as wind and gravity stresses on earth. Likely this huge space telescope will require asteroid or moon mining technology and much better robotic construction tools. But it is a much more likely future of massive spacecraft than the km scale star destroyers of Star Wars fame.
There is going to be increasing surveillance of public behavior in the decades to come. Police cameras seem inevitable. Human memory is just so unreliable that recording what actually happened will be overwhelmingly attractive. And sports have shown us what regular review of video can do to enhance performance. We'll have to be strategic in regulating access to these videos since they could become another piece of a comprehensive surveillance network that could enable those in power to suppress dissenters. It seems that recording is going to happen. The question is about how those in power are held accountable for how they use recordings.
Yes, the Cyc project was an attempt. You refer to a small number of optimists, but many already knew in the 1980s that its methods were far from adequate to the challenge. (My personal assessment was that common sense knowledge is much more approximate and context dependent than Cyc allows). I view Watson as largely a repudiation of the philosophy of Cyc. Rather than implement a self-consistent knowledge base, Watson does statistical analysis of a much less structured data base. Some of the people who dismiss Watson as 'not understanding' are taking the viewpoint that a self-consistent structured knowledge base is the essence of intelligence. They still view that goal as nearly impossible to attain. But Watson shows that many behaviors characteristic of intelligence can be achieved without this self-consistent knowledge base. It will be interesting to see whether the next steps involve moving back toward Cyc like philosophy or going further along the path of statistical analysis without constructed coherence.
Not sure why you say prediction is not relevant to neuroscience. It is central to some successful models of motor control, for example Shadmehr et al, "Error Correction, Sensory Prediction, and Adaptation in Motor Control", Annual Review of Neuroscience Vol. 33: 89-108, 2010.
This is crazy. If you can't make sense of your observations and connect them to our understanding, then they are unlikely to be useful. Existing journals will publish observations that are not explained if they are accompanied by a careful explanation of what is and is not understood about the problem. We definitely do not need more publication of observations disconnected from understanding.
Thanks PM. I think you are exactly right. I have a hard time understanding why so many take the Chinese room argument so seriously. (I wouldn't call it stupid...it is (smart) philosophers oversimplifying reality so they can cope with it using the tools at their disposal). The kind of processing done in the Chinese room is just a tiny piece of what is required to be intelligent.
Yes, this is a good way to think about it. Any AI is an expression and outgrowth of human intelligence. And Watson is totally amazing. People who dismiss it in hindsight do not realize how impossible such a system seemed in the 1980s. Of course the complex issue is that AI opens the possibility of intelligence very very different than human intelligence developing as an outgrowth of human intelligence. And we know so little about the kinds of intelligence that are possible that it is very hard to predict what interactions between very different kinds of intelligence might be.
Siri is a pretty useful bit of code. Are you sure its ability to get the job done is fundamentally different than a humans ability to get the job of survival and reproduction done? Intelligence is just a really hard problem. It will be decades, maybe centuries before we have a human level intelligence. But having gone from ELIZA to Siri and Watson in 50 years, it seems presumptious to assume that AI will not dramatically change our world in the next 50 or 100 years.
This article is on a useful track but suffers from a simple confusion. The author argues that intelligence should be defined by understanding and not by behavior, but then proposes that we use successful prediction as the measure of understanding. The confusion is that prediction is also a behavior. I agree with the direction the author is going. Most successful understanding (or intelligence if you like that word) can be connected to the use of a (partially) coherent set of ideas to predict observations. Intelligent agents have complex methods of adjusting their set of ideas to improve their ability to predict, and when an agent becomes able to predict a useful segment of the phenomena it encounters, we call it intelligent. But this is still a behavioral definition of intelligence. In fact, unless you hold onto some kind of dualism, all intelligence is a physical behavior of organisms. Without dualism, it makes no sense to distinguish 'understanding' from 'a behavior of a network of molecules, neurons, organs, and organisms'.
Of course, empirical verification has to be the deciding factor. But right now there is very little way to communicate the insights that many scientists have about which studies are reliable. We could be much more efficient in doing future studies if there were a way to collect assessments of those who work in closely related fields without taking years to get to know people well enough to get them to tell you their actual opinions even though they haven't been able to focus on that question to publish an article on the subject.
I don't think 'gambling' is the right word. And the devil is in the details. Any system like this is likely going to be gamed since careers and money are at stake. But the current publishing system is gamed all the time, so the new system doesn't have to be perfect. I just need a way to access some evaluations of a paper by others who have read it rather than figuring out all its problems on my own.
Has Unzicker really argued that QED is a grotesque kludge? He argues that string theory and other branches of recent high energy theory are modern religion, but I haven't heard that he attacks QED, which was developed around 1950 and is one of the most precisely experimentally confirmed theories of all time. Check out the Lamb Shift, the anomalous magnetic moment of the electron or the casimir effect for experiments that are accurately described by QED.
There is a simple reason this is controversial. Any Electromagnetic drive that produces more than 3.34 nanoNewtons per Watt by EM emission is a demonstration of new physics that is not included in our amazingly successful theory of quantum electrodynamics (QED). (The simple calculation is here: https://www.physicsforums.com/... They use a reflecting mirror, so an emitting craft would have half the force.) QED has been very precisely corroborated, sometimes to more than 10 digits (see https://en.wikipedia.org/wiki/... ). Claims of macroscopic objects that violate quantum electrodynamics simply have an extremely high prior probability of being false. (Just like claims of perpetual motion etc.). It doesn't mean we know a priori that they are false. By all means, do the experiments more precisely. But this is a claim that requires extraordinary proof because if it is true it will upset a lot of what we have good reason to think we understand about how the universe works.
Yes, it is absolutely true that too much choice is a problem. Apple's success is a clear example of this. They offer minimal choices. You are expected to take what they give you and not try to find just the right tweaks for you. It is a very delicate balance. We need expert engineers to make the best choices for us in many cases, but that quickly slides into corporations controlling individuals and customers rebel. If you really know better than the customers what is good for them, and they regularly find that trusting you is in their best interest. Then by all means, reduce choice. But most marketers are trying to sell things that are not made with the consideration of the customers or the environment. What we need is more choices among suppliers so that we can find some suppliers who are actually looking out for customer interests.
It would open up the possibility of observing the effects of quantization of gravitational interaction in the low field limit. Up to now, no one has observed any quantization of gravity. This is a really tiny effect, so you might argue that you don't care, but it would be a small clue in the big mystery of how to reconcile quantum mechanics and general relativity. In the history of physics, this has happened before. We had quantum mechanics in the 1920s through 1940s, but we didn't know how to quantize the electromagnetic field. We simply used classical interactions between charged particles and quantized their motion since we didn't know how to quantize the electromagnetic fields themselves. Then in the late 1940s and early 1950s, Schwinger, Feynman, and Tomonaga figured out how to quantize the electric and magnetic fields. It made only tiny changes in the predictions of quantum mechanics for atoms, but it has turned out to be critical to modern precision measurement and definition of the units we use. Their Quantum Electrodynamics has proved to be one of the great triumphs of theoretical physics.
Now quantization of gravity is a much much smaller effect in conditions that we can study on earth. This proposes that we might be able to observe some effects. Unfortunately, in this low field limit, I think most physicists expect that perturbative methods will give the right answer. In this case, the experiments will not be much help in building a self-consistent quantum gravity theory because perturbative methods are known to fail in the high field regime where the inconsistency between quantum mechanics and general relativity becomes important. But we definitely should make these measurements to see if the effects can be observed. Precision measurements often yield new insights, often unexpected ones.
Alfred Nobel's will says that his estate should fund 'prizes to those who, during the preceding year, shall have conferred the greatest benefit to mankind'. He lived in an age when physics was the study of the fundamental problems facing engineers of his day. Look at the careers of Kelvin or Helmholtz or Maxwell to see how closely tied these areas were. (Kelvin built transatlantic telegraph equipment, Maxwell developed color photography and studied bridge design, Helmholtz worked on physiology and thermodynamics inspired by applied science). I suspect the distance between modern fundamental particle physics and practical benefits to humanity might seem very foreign to Nobel were he alive to see it.
My concern is not actually for a subfield of physics. Applied research is often better funded than traditional reductionist physics. My concern is for physics as a discipline, and for the career path our brightest young aspiring physicists are directed down. We are at a cross-roads. Either physics will be the search for ever more fundamental models of the constituents of matter that become ever more irrelevant, and all the useful work will be done by people who call themselves something else. Or physics will become the application of quantitative models to fundamental problems in wide areas of science, and much of modern science will become ever more indistinguishable from applied physics. In the former case physics drifts into obscurity. In the latter case, physics strengthens its place as the central and fundamental science.
By dominate I mean 2 of the last 3 prizes and a fraction of recent prizes that is far larger than the fraction of physicists that work on high energy physics. (I count 6 out of 21 prizes since 1995 in high energy physics, or 28% to a community that is something like 15% of the American Physical Society). We simply haven't figured out how to recognize the more important contributions in less reductionist and more applied areas of physics. Last year's prize for semiconductor LED breakthroughs was a step in the right direction. But going back to neutrinos so quickly reflects the prize committee doesn't really get it.
It is disappointing to see the high energy physicists continue to dominate the nobel prize. Since the 1930s, anyone who discovers some new quirk about some fundamental particle gets the prize. Nevermind the fact that the discovered particle properties steadily become less and less relevant to anything that affects humans in any practical way. When Chadwick discovered the Neutron, that deserved a Nobel Prize. In many ways it was the beginning of nuclear physics. Neutrino mass was a nice puzzle. They (partly) solved it. Good work. But who cares? The future of physics is either to become solvers of interesting but irrelevant puzzles (which will be funded at the level that society funds other impractical but fascinating fields like poetry or classical music), or to become solvers of fundamental quantitative problems in materials science, environmental science, engineering, biology, and computational science that society needs solved. The overwhelming majority of actual physicists have already realized this fact and moved to work on relevant problems. But the nobel prize committee is stuck on the old path to irrelevance. The physiology prize is given to scientists who developed cures for parasitic diseases that affects many people while the physics prize goes to the discovery of a property of particles that can only be indirectly measured with massive detector. And we wonder why physics seems uninteresting to the next generation?
I recommend you read the history more carefully. In 1900 the best scientists knew that they did not understand what matter was or why it had such strange spectra (Kelvin even spent quite a bit of effort trying to develop vortex descriptions of atoms.) They knew that they didn't know where the sun's energy came from. They knew that they knew almost nothing about how cells worked. But surprisingly, it turns out that they already did know most of what we currently think is important to teach to engineering students. There is no quantum mechanics or relativity in mechanical or civil engineering curricula, and it is a tiny part of EE and chemical engineering curricula. Even something on the order of half of introductory chemistry would be familiar to a chemist in 1900. Michaelson said "The more important fundamental laws and facts of physical science have all been discovered" in 1903, and you would have a very difficult time today making the argument that he was wrong. Engineers use classical physics and empirical materials science much much more than they use the quantum physics discovered this century.
Now fast forward to 2015. The anomalous experiments we have involve dark matter on the scale of galaxies and dark energy on the scale of the visible universe. What part of our solar system is not described adequately enough by the standard model and general relativity that you have reason to expect it will allow new technology? What do you expect the lifetime of particles beyond the standard model to be? (The Higgs Boson lifetime is 10^-22 s and I don't know of any technology that uses particles with lifetimes shorter than muons with lifetime of 10^-6 s.) There will be major breakthrough in science and technology in the future. Much of it will be in understanding and designing complex systems. But practical engineering is going to be done using things whose fundamental constituents are adequately described by the standard model and general relativity. It is simply wishful thinking based on a faith in progress that makes people so certain that fundamental science is going to keep revolutionizing our lives the way it did in the 19th and 20th centuries.
When we look back from the year 5000, the years between 1800 and 1980 will stand out as the time period during which we figured out the main fundamental science to understand how the world around us works. We are not at the end of particle physics. There will be lots more to learn from higher energy and higher luminocity colliders, as well as studies of extremely high energy cosmic rays and astronomical data. But even if a bunch of new particles with masses in the TeV range are found, they won't change the models we use to describe materials, biology, planet formation, or neuroscience. Particle physics may make new discoveries or may turn cold, it is hard to say, but you can be essentially certain that it will not be practically useful.
I think the recent estimates say that habitable planets are quite plentiful. You are right that we have only found a few of them, but from the sensitivity of our search and the number we have found, we can estimate the total number, and it is very large, likely billions or tens of billions, maybe more. (https://medium.com/starts-with-a-bang/how-many-habitable-planets-are-in-our-galaxy-5bcf6db80c7f) The difference between the real world and Star Trek is that they can travel to a large fraction of the galaxy in a lifetime while there are no stars except the sun within 10000 years of earth at current spacecraft speeds.
Yes, it is a new era. About Amazon: they are seriously competing with Netflix, Apple, the cable companies and others to control the new media distribution system. Many of us that have used Amazon Prime just fell into instant video subscriptions and it is a low cost alternative that might beat out Netflix if they stumble again. It is going to take time for culture to evolve to effective use video on demand. We are at a moment of very rapid change in technological possibilities but people haven't figured out how to use the new possibilities effectively.
At the root of this is the notion that television is entertainment that is separate from life. It has always been an empty paradigm, and for decades thoughtful and ambitious people have avoided that kind of television. But while we were stuck with a broadcast model, it was the only option that could attract a large enough audience at a specific time to work economically. Now, with content on demand, it is possible for television content to be selected so that it is much more useful and relevant to people's lives. So that Kids programming can be chosen to match the developmental stage and educational goals for a child at a specific time. And dramas can be chosen to match the interests and psychology of a viewer. What we haven't yet figured out how to do is how to produce quality programming at the high volumes this model requires and we also haven't yet figured out how to index the content to allow new content to connect effectively with viewers. We are still stuck with a ratings system that is based on the old broadcast model rather than rewarding shows that effectively connect with a niche audience over many years.
When the instructor effectively places the material they are presenting in a larger framework including unknowns, it is often quite inspiring. Textbooks in mathematics and physics are the worst in this regard. They try to paint their presentation as the complete story on the subject and that leaves students bored. Even just a little bit of explaining the complex problems that are being sidestepped by the way the course material was chosen can greatly enliven a course. Even better, the students come out with an understanding of where the methods they learned will work and where they will not.
No, you are pretty precisely wrong. Elon and Gates made their fortunes in the software business but don't work in exactly the niche of AI. Exactly the same as Einstein worked in relativity and a bit in quantum mechanics, not nuclear physics. While AI and nuclear explosions are totally different, the level of understanding of the possibilities comparing now to 1939 is not all that different. At least you give no reasons beyond personal incredulity for the claim that there is no feasible way for this to happen in the next century.
The comparison to the risk of atmospheric fire is also precisely wrong. That was brought up as a possibility in the 1940s. The experts evaluated it, and concluded it was extremely low probability. Strong AI on the other hand is estimated by many experts to be very likely over the next century. (https://nakedsecurity.sophos.com/2015/05/27/1-in-5-experts-believe-artificial-intelligence-will-pose-an-existential-threat/ ) The main question is whether it will be a threat.
It is the next century or two that Musk, Gates, and others are warning of. And it is quite short sighted to dismiss the threat with 'there is no feasible way for this to happen' right now.
It has happened before that the smartest people in the world warn that technological advances may present major new weapons and threats.
Last time it was Einstein and Szilard in 1939 warning that nuclear weapons might be possible. The letter to Roosevelt was three years before anyone had even built a nuclear reactor and 6 years before the first nuclear explosion. Nuclear bombs could easily have been labelled a "problem that probably does not exist." And if someone could destroy the planet, what could you do about it anyway? The US took the warning seriously and ensured that the free world and not a totalitarian dictator was the first capable of obliterating its opponents.
This time Elon Musk, Bill Gates, and Stephen Hawking are warning that superintelligence may make human intelligence obsolete. And they are dismissed because we haven't yet made human level intelligence and because if we did we couldn't do anything about it. If it is Musk, Gates, and Hawking vs Edward Geist, the smart money has to be with the geniuses. But if you look at the arguments, you see you don't even have to rely on their reputation. The argument is hands down won by the observation that human level artificial intelligence is an existential risk. Even if it is only 1% likely to happen in the next 500 years, we need to have a plan for how to deal with it. The root of the problem is that the capabilities of AI are expanding much faster than human capabilities can expand, so it is quite possible that we will lose our place as the dominant intellect on the planet. And that changes everything.
Slashdot Mirror
Dr. Spork is on to something here. We still know so very little about our universe and by far the best way forward is to develop very large aperture telescopes. Humans are headed toward building truly massive optical telescopes (100m scale this century, maybe 1km scale in the 2100s or 2200s). Right now we need to succeed in getting the James Webb 6.5m telescope in space and the 30m earth instruments operational. There are several ideas about how to build 100m scale optical telescopes. (See the OWL proposal, http://www.gemini.edu/science/...) One path is to use many independent small telescopes and combine the light as is done with interferometers. Maybe our photonics expertise will advance faster than our large scale construction expertise, and we will do all huge telescopes with many mass produced smaller ones. Maybe their relative positioning doesn't even need to be super stable because we can measure it accurately enough and correct for their relative motion. But it is fun to imagine a 1km scale truss system to hold a massive segmented mirror in space. Picture 250,000 mirrors, each 2m in diameter. It will not remove all stresses to be in space. At that scale, you may start to get tidal stresses and thermal stresses that are almost as hard to engineer as wind and gravity stresses on earth. Likely this huge space telescope will require asteroid or moon mining technology and much better robotic construction tools. But it is a much more likely future of massive spacecraft than the km scale star destroyers of Star Wars fame.
There is going to be increasing surveillance of public behavior in the decades to come. Police cameras seem inevitable. Human memory is just so unreliable that recording what actually happened will be overwhelmingly attractive. And sports have shown us what regular review of video can do to enhance performance. We'll have to be strategic in regulating access to these videos since they could become another piece of a comprehensive surveillance network that could enable those in power to suppress dissenters. It seems that recording is going to happen. The question is about how those in power are held accountable for how they use recordings.
Yes, the Cyc project was an attempt. You refer to a small number of optimists, but many already knew in the 1980s that its methods were far from adequate to the challenge. (My personal assessment was that common sense knowledge is much more approximate and context dependent than Cyc allows). I view Watson as largely a repudiation of the philosophy of Cyc. Rather than implement a self-consistent knowledge base, Watson does statistical analysis of a much less structured data base. Some of the people who dismiss Watson as 'not understanding' are taking the viewpoint that a self-consistent structured knowledge base is the essence of intelligence. They still view that goal as nearly impossible to attain. But Watson shows that many behaviors characteristic of intelligence can be achieved without this self-consistent knowledge base. It will be interesting to see whether the next steps involve moving back toward Cyc like philosophy or going further along the path of statistical analysis without constructed coherence.
Not sure why you say prediction is not relevant to neuroscience. It is central to some successful models of motor control, for example Shadmehr et al, "Error Correction, Sensory Prediction, and Adaptation in Motor Control", Annual Review of Neuroscience Vol. 33: 89-108, 2010.
This is crazy. If you can't make sense of your observations and connect them to our understanding, then they are unlikely to be useful. Existing journals will publish observations that are not explained if they are accompanied by a careful explanation of what is and is not understood about the problem. We definitely do not need more publication of observations disconnected from understanding.
Thanks PM. I think you are exactly right. I have a hard time understanding why so many take the Chinese room argument so seriously. (I wouldn't call it stupid...it is (smart) philosophers oversimplifying reality so they can cope with it using the tools at their disposal). The kind of processing done in the Chinese room is just a tiny piece of what is required to be intelligent.
Yes, this is a good way to think about it. Any AI is an expression and outgrowth of human intelligence. And Watson is totally amazing. People who dismiss it in hindsight do not realize how impossible such a system seemed in the 1980s. Of course the complex issue is that AI opens the possibility of intelligence very very different than human intelligence developing as an outgrowth of human intelligence. And we know so little about the kinds of intelligence that are possible that it is very hard to predict what interactions between very different kinds of intelligence might be.
Siri is a pretty useful bit of code. Are you sure its ability to get the job done is fundamentally different than a humans ability to get the job of survival and reproduction done? Intelligence is just a really hard problem. It will be decades, maybe centuries before we have a human level intelligence. But having gone from ELIZA to Siri and Watson in 50 years, it seems presumptious to assume that AI will not dramatically change our world in the next 50 or 100 years.
This article is on a useful track but suffers from a simple confusion. The author argues that intelligence should be defined by understanding and not by behavior, but then proposes that we use successful prediction as the measure of understanding. The confusion is that prediction is also a behavior. I agree with the direction the author is going. Most successful understanding (or intelligence if you like that word) can be connected to the use of a (partially) coherent set of ideas to predict observations. Intelligent agents have complex methods of adjusting their set of ideas to improve their ability to predict, and when an agent becomes able to predict a useful segment of the phenomena it encounters, we call it intelligent. But this is still a behavioral definition of intelligence. In fact, unless you hold onto some kind of dualism, all intelligence is a physical behavior of organisms. Without dualism, it makes no sense to distinguish 'understanding' from 'a behavior of a network of molecules, neurons, organs, and organisms'.
Of course, empirical verification has to be the deciding factor. But right now there is very little way to communicate the insights that many scientists have about which studies are reliable. We could be much more efficient in doing future studies if there were a way to collect assessments of those who work in closely related fields without taking years to get to know people well enough to get them to tell you their actual opinions even though they haven't been able to focus on that question to publish an article on the subject.
I don't think 'gambling' is the right word. And the devil is in the details. Any system like this is likely going to be gamed since careers and money are at stake. But the current publishing system is gamed all the time, so the new system doesn't have to be perfect. I just need a way to access some evaluations of a paper by others who have read it rather than figuring out all its problems on my own.
Has Unzicker really argued that QED is a grotesque kludge? He argues that string theory and other branches of recent high energy theory are modern religion, but I haven't heard that he attacks QED, which was developed around 1950 and is one of the most precisely experimentally confirmed theories of all time. Check out the Lamb Shift, the anomalous magnetic moment of the electron or the casimir effect for experiments that are accurately described by QED.
There is a simple reason this is controversial. Any Electromagnetic drive that produces more than 3.34 nanoNewtons per Watt by EM emission is a demonstration of new physics that is not included in our amazingly successful theory of quantum electrodynamics (QED). (The simple calculation is here: https://www.physicsforums.com/... They use a reflecting mirror, so an emitting craft would have half the force.) QED has been very precisely corroborated, sometimes to more than 10 digits (see https://en.wikipedia.org/wiki/... ). Claims of macroscopic objects that violate quantum electrodynamics simply have an extremely high prior probability of being false. (Just like claims of perpetual motion etc.). It doesn't mean we know a priori that they are false. By all means, do the experiments more precisely. But this is a claim that requires extraordinary proof because if it is true it will upset a lot of what we have good reason to think we understand about how the universe works.
Yes, it is absolutely true that too much choice is a problem. Apple's success is a clear example of this. They offer minimal choices. You are expected to take what they give you and not try to find just the right tweaks for you. It is a very delicate balance. We need expert engineers to make the best choices for us in many cases, but that quickly slides into corporations controlling individuals and customers rebel. If you really know better than the customers what is good for them, and they regularly find that trusting you is in their best interest. Then by all means, reduce choice. But most marketers are trying to sell things that are not made with the consideration of the customers or the environment. What we need is more choices among suppliers so that we can find some suppliers who are actually looking out for customer interests.
It would open up the possibility of observing the effects of quantization of gravitational interaction in the low field limit. Up to now, no one has observed any quantization of gravity. This is a really tiny effect, so you might argue that you don't care, but it would be a small clue in the big mystery of how to reconcile quantum mechanics and general relativity. In the history of physics, this has happened before. We had quantum mechanics in the 1920s through 1940s, but we didn't know how to quantize the electromagnetic field. We simply used classical interactions between charged particles and quantized their motion since we didn't know how to quantize the electromagnetic fields themselves. Then in the late 1940s and early 1950s, Schwinger, Feynman, and Tomonaga figured out how to quantize the electric and magnetic fields. It made only tiny changes in the predictions of quantum mechanics for atoms, but it has turned out to be critical to modern precision measurement and definition of the units we use. Their Quantum Electrodynamics has proved to be one of the great triumphs of theoretical physics.
Now quantization of gravity is a much much smaller effect in conditions that we can study on earth. This proposes that we might be able to observe some effects. Unfortunately, in this low field limit, I think most physicists expect that perturbative methods will give the right answer. In this case, the experiments will not be much help in building a self-consistent quantum gravity theory because perturbative methods are known to fail in the high field regime where the inconsistency between quantum mechanics and general relativity becomes important. But we definitely should make these measurements to see if the effects can be observed. Precision measurements often yield new insights, often unexpected ones.
Alfred Nobel's will says that his estate should fund 'prizes to those who, during the preceding year, shall have conferred the greatest benefit to mankind'. He lived in an age when physics was the study of the fundamental problems facing engineers of his day. Look at the careers of Kelvin or Helmholtz or Maxwell to see how closely tied these areas were. (Kelvin built transatlantic telegraph equipment, Maxwell developed color photography and studied bridge design, Helmholtz worked on physiology and thermodynamics inspired by applied science). I suspect the distance between modern fundamental particle physics and practical benefits to humanity might seem very foreign to Nobel were he alive to see it.
My concern is not actually for a subfield of physics. Applied research is often better funded than traditional reductionist physics. My concern is for physics as a discipline, and for the career path our brightest young aspiring physicists are directed down. We are at a cross-roads. Either physics will be the search for ever more fundamental models of the constituents of matter that become ever more irrelevant, and all the useful work will be done by people who call themselves something else. Or physics will become the application of quantitative models to fundamental problems in wide areas of science, and much of modern science will become ever more indistinguishable from applied physics. In the former case physics drifts into obscurity. In the latter case, physics strengthens its place as the central and fundamental science.
By dominate I mean 2 of the last 3 prizes and a fraction of recent prizes that is far larger than the fraction of physicists that work on high energy physics. (I count 6 out of 21 prizes since 1995 in high energy physics, or 28% to a community that is something like 15% of the American Physical Society). We simply haven't figured out how to recognize the more important contributions in less reductionist and more applied areas of physics. Last year's prize for semiconductor LED breakthroughs was a step in the right direction. But going back to neutrinos so quickly reflects the prize committee doesn't really get it.
It is disappointing to see the high energy physicists continue to dominate the nobel prize. Since the 1930s, anyone who discovers some new quirk about some fundamental particle gets the prize. Nevermind the fact that the discovered particle properties steadily become less and less relevant to anything that affects humans in any practical way. When Chadwick discovered the Neutron, that deserved a Nobel Prize. In many ways it was the beginning of nuclear physics. Neutrino mass was a nice puzzle. They (partly) solved it. Good work. But who cares? The future of physics is either to become solvers of interesting but irrelevant puzzles (which will be funded at the level that society funds other impractical but fascinating fields like poetry or classical music), or to become solvers of fundamental quantitative problems in materials science, environmental science, engineering, biology, and computational science that society needs solved. The overwhelming majority of actual physicists have already realized this fact and moved to work on relevant problems. But the nobel prize committee is stuck on the old path to irrelevance. The physiology prize is given to scientists who developed cures for parasitic diseases that affects many people while the physics prize goes to the discovery of a property of particles that can only be indirectly measured with massive detector. And we wonder why physics seems uninteresting to the next generation?
I recommend you read the history more carefully. In 1900 the best scientists knew that they did not understand what matter was or why it had such strange spectra (Kelvin even spent quite a bit of effort trying to develop vortex descriptions of atoms.) They knew that they didn't know where the sun's energy came from. They knew that they knew almost nothing about how cells worked. But surprisingly, it turns out that they already did know most of what we currently think is important to teach to engineering students. There is no quantum mechanics or relativity in mechanical or civil engineering curricula, and it is a tiny part of EE and chemical engineering curricula. Even something on the order of half of introductory chemistry would be familiar to a chemist in 1900. Michaelson said "The more important fundamental laws and facts of physical science have all been discovered" in 1903, and you would have a very difficult time today making the argument that he was wrong. Engineers use classical physics and empirical materials science much much more than they use the quantum physics discovered this century.
Now fast forward to 2015. The anomalous experiments we have involve dark matter on the scale of galaxies and dark energy on the scale of the visible universe. What part of our solar system is not described adequately enough by the standard model and general relativity that you have reason to expect it will allow new technology? What do you expect the lifetime of particles beyond the standard model to be? (The Higgs Boson lifetime is 10^-22 s and I don't know of any technology that uses particles with lifetimes shorter than muons with lifetime of 10^-6 s.) There will be major breakthrough in science and technology in the future. Much of it will be in understanding and designing complex systems. But practical engineering is going to be done using things whose fundamental constituents are adequately described by the standard model and general relativity. It is simply wishful thinking based on a faith in progress that makes people so certain that fundamental science is going to keep revolutionizing our lives the way it did in the 19th and 20th centuries.
When we look back from the year 5000, the years between 1800 and 1980 will stand out as the time period during which we figured out the main fundamental science to understand how the world around us works. We are not at the end of particle physics. There will be lots more to learn from higher energy and higher luminocity colliders, as well as studies of extremely high energy cosmic rays and astronomical data. But even if a bunch of new particles with masses in the TeV range are found, they won't change the models we use to describe materials, biology, planet formation, or neuroscience. Particle physics may make new discoveries or may turn cold, it is hard to say, but you can be essentially certain that it will not be practically useful.
I think the recent estimates say that habitable planets are quite plentiful. You are right that we have only found a few of them, but from the sensitivity of our search and the number we have found, we can estimate the total number, and it is very large, likely billions or tens of billions, maybe more. (https://medium.com/starts-with-a-bang/how-many-habitable-planets-are-in-our-galaxy-5bcf6db80c7f) The difference between the real world and Star Trek is that they can travel to a large fraction of the galaxy in a lifetime while there are no stars except the sun within 10000 years of earth at current spacecraft speeds.
Yes, it is a new era. About Amazon: they are seriously competing with Netflix, Apple, the cable companies and others to control the new media distribution system. Many of us that have used Amazon Prime just fell into instant video subscriptions and it is a low cost alternative that might beat out Netflix if they stumble again. It is going to take time for culture to evolve to effective use video on demand. We are at a moment of very rapid change in technological possibilities but people haven't figured out how to use the new possibilities effectively.
At the root of this is the notion that television is entertainment that is separate from life. It has always been an empty paradigm, and for decades thoughtful and ambitious people have avoided that kind of television. But while we were stuck with a broadcast model, it was the only option that could attract a large enough audience at a specific time to work economically. Now, with content on demand, it is possible for television content to be selected so that it is much more useful and relevant to people's lives. So that Kids programming can be chosen to match the developmental stage and educational goals for a child at a specific time. And dramas can be chosen to match the interests and psychology of a viewer. What we haven't yet figured out how to do is how to produce quality programming at the high volumes this model requires and we also haven't yet figured out how to index the content to allow new content to connect effectively with viewers. We are still stuck with a ratings system that is based on the old broadcast model rather than rewarding shows that effectively connect with a niche audience over many years.
When the instructor effectively places the material they are presenting in a larger framework including unknowns, it is often quite inspiring. Textbooks in mathematics and physics are the worst in this regard. They try to paint their presentation as the complete story on the subject and that leaves students bored. Even just a little bit of explaining the complex problems that are being sidestepped by the way the course material was chosen can greatly enliven a course. Even better, the students come out with an understanding of where the methods they learned will work and where they will not.
No, you are pretty precisely wrong. Elon and Gates made their fortunes in the software business but don't work in exactly the niche of AI. Exactly the same as Einstein worked in relativity and a bit in quantum mechanics, not nuclear physics. While AI and nuclear explosions are totally different, the level of understanding of the possibilities comparing now to 1939 is not all that different. At least you give no reasons beyond personal incredulity for the claim that there is no feasible way for this to happen in the next century.
The comparison to the risk of atmospheric fire is also precisely wrong. That was brought up as a possibility in the 1940s. The experts evaluated it, and concluded it was extremely low probability. Strong AI on the other hand is estimated by many experts to be very likely over the next century. (https://nakedsecurity.sophos.com/2015/05/27/1-in-5-experts-believe-artificial-intelligence-will-pose-an-existential-threat/ ) The main question is whether it will be a threat.
It is the next century or two that Musk, Gates, and others are warning of. And it is quite short sighted to dismiss the threat with 'there is no feasible way for this to happen' right now.
It has happened before that the smartest people in the world warn that technological advances may present major new weapons and threats. Last time it was Einstein and Szilard in 1939 warning that nuclear weapons might be possible. The letter to Roosevelt was three years before anyone had even built a nuclear reactor and 6 years before the first nuclear explosion. Nuclear bombs could easily have been labelled a "problem that probably does not exist." And if someone could destroy the planet, what could you do about it anyway? The US took the warning seriously and ensured that the free world and not a totalitarian dictator was the first capable of obliterating its opponents.
This time Elon Musk, Bill Gates, and Stephen Hawking are warning that superintelligence may make human intelligence obsolete. And they are dismissed because we haven't yet made human level intelligence and because if we did we couldn't do anything about it. If it is Musk, Gates, and Hawking vs Edward Geist, the smart money has to be with the geniuses. But if you look at the arguments, you see you don't even have to rely on their reputation. The argument is hands down won by the observation that human level artificial intelligence is an existential risk. Even if it is only 1% likely to happen in the next 500 years, we need to have a plan for how to deal with it. The root of the problem is that the capabilities of AI are expanding much faster than human capabilities can expand, so it is quite possible that we will lose our place as the dominant intellect on the planet. And that changes everything.