Perspectives on Science and Culture: February 2017

Dorn Cox, FarmOS

Collective action, Political will, Artificial Intelligence: Which lever to pull?

In 1938, Aldo Leopold, the father of conservation, wrote “We end, I think, at what might be called the standard paradox of the twentieth century: our tools are better than we are, and grow better faster than we do. They suffice to crack the atom, to command the tides. But they do not suffice for the oldest task in human history: to live on a piece of land without spoiling it.” Does it take more technology to live sustainably? Of course not; this is a fact. Jared Diamond described a few societies in his book "Collapse: How Societies Choose to Fail or Succeed" that did learn to live sustainably despite having few of the technological tools available today, as has Dan Dagget in Gardeners of Eden. Elsewhere Joe Romm has made it clear that we have the tools today to eliminate dependence of fossil fuels. So no, we don't need more technology. We need a change in thinking about how to make intelligent use of what is currently available to us.

How many of our technologies are really put toward the end of reducing environmental harm instead of causing it? We need to evaluate them according to that criteria and reassess the values we hold as a society, as expressed in our choices and behavior. Without responsible values at the foundation to shape how we use our tools, of what use are they? I would suggest that the adaptive potential of machine learning to increase efficiency and address land management issues makes a behavioral shift easier and lessens the impact overall. Making sustainability not only the right choice, but also the easy (or affordable) choice, is important. There is a subtle give and take: Our behavioral preferences dictate the tools we invent and use, but our tools also influence our behavioral choices. Who knows where and how to use the tools we have today to acheive net zero?

Do we need a different set of values? Said Jared Diamond, in ‘Collapse’: “That proves to be a common theme throughout history. The values to which people cling most stubbornly under inappropriate conditions are those values that were previously the source of their greatest triumphs over adversity.” That said, I don't think our values are the problem, so much as we cannot effectively apply them at larger scales. What in human history ever prepared us to be the most powerful global force? This isn't a role evolution predisposed us for. As a society we tend to be aimless, petty, and struggle to meet global challenges during the best of times (during the worst of times... we sabotage ourselves). We need to learn to cooperate at a global scale much better than we do today in order to actually realize an ethical world. I'd argue cooperation at that scale requires augmented intelligence for a number of reasons.

People are very good at identifying important things, on a case by case basis, but less so at identifying the interactions between multiple things and how that contributes to a unified whole. And related to that we tend to be not good at making transitions from an inefficient state to a more optimal state. That’s where ambient intelligence comes in. To enable us to strategically make, analyze, and prioritize long-term plans by maintaining sustained focus on multiple things at the same time and their large scale (and small scale) interactions. And synchronizing interactions at the individual, social, and environmental levels. We still don’t do this well enough. Roxanne Bauer: "The Internet of Things and the ambient intelligence it generates can solve many problems by getting the right information to the right place at the right time. The potential benefits are huge - both in terms of improved living standards and in cost savings through greater efficiency."

Should we focus on political action? Yes, but... I think we need to learn to operate outside the political framework to some extent and localize technical recommendations, specify individual actions, while at the same time keep the national/global context in view. We need to make "what we need to technically do" less a matter of politics and at least, if not more, a matter of economics. From what I've seen, policies tend to follow and trail behind developments in technology and economics, and less so the other way around.

To illustrate my point, why did Deepmind find a path to a 10% reduction in energy use through optimisation alone, a path we were unaware of before? There is no simple solution to these large scale problems like those encountered when design a smart grid or tackling climate change. The only real solutions involve a combination of many things, some we've known for a long time, some we are only beginning to understand. All these individual actions - slight and great, in social structures and technological capabilities - must come together to allow us to realize the swiftest and most sure path to net zero emissions. This is big and almost impossibly difficult, but combining existing technologies in new, and novel, ways may be all it takes. Artificial intelligence is peerless in solving complex but defined problems. AI scientists predict computers will increasingly be able to search through thickets of alternatives to find patterns and solutions that elude the human mind. We need an "algorithm for Earth." Or as Drew Purves put it, an "Intelligent Biosphere." This is how we confront power, by indirectly undermining BAU and changing the foundation it rests on. Linear antagonism, a direct attack to the fortified head of plutocratic petrostates, hasn't achieved nearly as much to date.

As David Bowie said, ‘Tomorrow belongs to those who can hear it coming.' The problem is, most of us can't hear it. We need better ears. To evolutionary algorithms, climate change is just another fundamentally solvable problem. Demis Hassabis is at the helm of DeepMind, and they've proposed using these methods to increase grid efficiency dramatically, and without building new infrastructure. That's just one organization, one application. Algorithms like these are not inherently limited by any single worldview or language; the bounds of what's possible begin to dissolve. The single greatest tool we have to leverage is intelligence, and I think we need to apply this in every sector, reiterated many times over.

In a departure from pure climate/energy issues, this current post is intended to serve as a brief outline of topics related to AI, including some of the more speculative applications of these technologies, and includes a few of the factors leading to the growth and evolution of energy management systems. As many people have pointed out, the implications of AI for society are potentially as disruptive as climate change, and present their own challenges. And like climate change, this disruption is already being felt. But as a component of an energy transition, the appropriate use of these technologies is part of a renewable, sustainable future. In fact, such a future may be be almost impossible without AI. In previous posts the ability of renewable energy resources to replace carbon intensive fuels and thereby form a key part of regional and global efforts to respond to the threat of climate change was examined. In this post I also take a closer look at the tech-driven transition involving both energy specialists, who can propose alternate scenarios to carbon intensive fuels, and the energy management systems capable of integrating distributed energy. But the main focus is on agriculture as the sector of greatest potential.

AI: Agricultural Intelligence?

As David Roberts has pointed out, climate change is an environmental problem, an energy transition, a national security threat, a market failure, an economic opportunity, an obligation to our children, a political dispute, and a question of justice. This makes addressing it effectively a very difficult proposition, but we know that we must realize Oliver Geden's "net zero emissions" target within a very short span of time to prevent the worst consequences from occurring.

Consensus has emerged that negative emission technologies (NETs) have become an unavoidable part of efforts to stabilize the climate, and increasing soil carbon sequestration with regenerative agriculture is one proposed way to achieve this. But it is not our only goal, at the same time we must enhance the productivity of the farm ecosystem and reduce agricultural inputs. How can a farm achieve these multiple functions? One promising approach involves "field robotics," an area of research in precision agriculture that is seeing considerable development. Field robots can plant, irrigate, weed, identify individual pieces of fruit and nuts, autonomously harvest at the right time, manage pests, and work day or night. The fine grained data they collect can also be analyzed to improve productivity and suggest new methods. Optimizing agriculture for negative emissions would have a substantial mitigating impact on climate change, as a third of all GHG emissions come from agriculture. Improved cultivation practices could eliminate them and sequester as much again in increased organic content of soil, resulting in falling atmospheric CO2 levels if carried out on a sufficiently large scale.

In Gardeners of Eden, Dan Dagget described how historically humans had in many instances enhanced ecosystem productivity. But in the development of modern high input, energy and resource intensive industrial factory farming, this relationship became inverted, causing ecosystem degradation and contributing to global warming. Precision agriculture holds the promise of picking up where we left off, bringing ecosystems to greater health with greater crop diversity and expanded farming methods. Norman Borlaug inaugurated the Green Revolution that brought rises in productivity, but the overall effects on the environment were mixed. The future of our sustainable relationship with nature is probably just beginning; let's put precision farming to work building on the research of Albert Howard and utilizing the insights of agroecology. This will create augmented ecosystems where the lines between artificial and biotic systems blur and begin to merge. Perhaps someday farms operated by solar "agribots" will mimic the ant-fungal mutualism in efficiency and productivity. We may already be seeing the beginning stages.

Precision farming is an important field and will only become more so as arable land is degraded, lost to erosion, inundated by sea level rise, and as marginal lands or areas with shorter growing seasons (Alaska) are increasingly utilized. Furthermore, as we look for negative emissions technologies, regenerative agriculture continues to be one of the most promising options for increasing soil carbon sequestration and enhancing biomass growth. All this suggests that agricultural robotics, drones, and the computing power, sensors, and software for sophisticated automation will increase. And this technology is scalable - there's as much potential to improve small farms as large commercial operations. In agriculture a lot of time and energy is spent preparing a plot of land for planting, weeding, and eventually harvest. These are all tasks involving moving and contouring the land, spreading fertilizer, irrigating, etc. Many of these tasks can be completed using biomimetic autonomous robots capable of merging infrastructure with agriculture. While it is possible for humans to do much of this work, we have limitations in terms of time, energy, precision, and the conditions in which we operate. The application of AI related technologies to the field of precision agriculture could transform monoculture crop fields to farms that more closely resemble species diverse natural habitats while reducing large farm machinery and increasing production levels.

Salah Sukkarieh, Professor of Robotics and Intelligent Systems at The Australian Centre For Field Robotics at the University of Sydney, is leading the vanguard of making robots to work on farms. “Robotics and automation technology provides the grower with greater knowledge of their farm state, and the capability for acting in real-time, thus increasing efficiency, reliability and productivity whilst minimising environmental impact." “In addition we need to be thinking beyond the robotic devices themselves, and focus on how future farms will be structured and operate as a whole with autonomous systems. This will necessitate the need to investigate ways to support agribusiness start-up companies, as well as educate and manage the transition of current agricultural knowledge into the era of digitisation.” "There's been a lot of discussion around STEM education and Australia's urgent need to prepare future generations for a new-look work environment. Introducing rural and regional students to hands-on robotic technologies and activities would give them access to exciting new career options.” See a video of RIPPA (Robot for Intelligent Perception and Precision Application) in action.

DeepMind: Optimize Energy Use.

Statement

Feed DeepMind raw data (economic, demographic, geographic, resources, infrastructure, etc.) and/or connect it directly to sensors, and it will undertake the tasks of feature specification and optimization to provide a roadmap with the greatest chance of success for reaching predetermined values. It is the most powerful tool yet created for scenarios modeling, or anything else for that matter. And as such it has already demonstrated an ability to reveal possibilities and generate designs that no one has ever considered before.

The closest human analogues to this kind of analysis have all been performed by energy specialists at the forefront of their fields, like Bill Powers (BASE 2020), Mark Jacobson (Stanford), Richard Heinberg (Our Renewable Future), Sadhu Johnston (Vancouver's Renewable City Strategy), Amory Lovins (RMI), and Christian Breyer (Finland).

Here's the thing. Jacobson has estimated a 100% transition to carbon neutral energy for all sectors by 2050 is entirely possible. I suspect that tapping into advanced tools like DeepMind could slash that timeframe and the associated costs, while providing a blueprint that is far more detailed, specific and flexible where needed. Demis Hassabis, the co-founder of DeepMind Technologies, wants to solve complex global problems, he has his chance. We're one step away from David Grinspoon's mature anthropocene.

Two examples of the power of this machine learning algorithm approach already exist. At Google "DeepMind developed a reinforcement-learning algorithm that experimented with different data-center configurations (in simulation), including cooling systems and windows, until it lowered overall power consumption.51"

Extrapolating the potential impact of AI from a single use case.

DeepMind learned how to become better than any human player at Go, a game with more possible moves than atoms in the universe. John McPhee suggests that "When AlphaGo made move 37 in Match 2 against Lee Sedol, one of the world's human Go champions, it was not just following what humans had programmed it to do, it appeared to "understand" and come up with a move no human player would have made. Let's say we put an AI computer in charge of our power grid and it decides to shut down portions of it because it "understands" those coal-fired power plants are harming the environment. Move 37.

George Zarkadakis: "Arguably, AlphaGo, with its Move 37, triggered Sedol’s human Move 78. In other words, the human-machine interaction during the game enabled - or liberated - the human to become more intuitive, and to discover a part of reality that was hitherto unknown. The transcendent experience of Sedol’s Move 78 may be a harbinger of our future relation with human-like intelligent machines. They will allow us to explore and discover new areas of intelligence, and thus expand our intelligence too. ...The intelligent machines of the future will become our vehicles, and partners, in pushing the boundaries of our own intelligence even further, much further, more often; and by many more."

Cade Metz: "Following the game, in the control room, Silver could revisit the precise calculations AlphaGo made in choosing Move 37. Drawing on its extensive training with millions upon millions of human moves, the machine actually calculates the probability that a human will make a particular play in the midst of a game. “That’s how it guides the moves it considers,” Silver says. For Move 37, the probability was one in ten thousand. In other words, AlphaGo knew this was not a move that a professional Go player would make. But, drawing on all its other training with millions of moves generated by games with itself, it came to view Move 37 in a different way. It came to realize that, although no professional would play it, the move would likely prove quite successful. “It discovered this for itself,” Silver says, “through its own process of introspection and analysis.”

Artificial General Intelligence, Existentialism, and Poker: The deep questions.

Life includes many things that are imperfect, impermanent, and incomplete (the definition of wabi sabi). One consequnce of that is imperfect information. The development of AI has also had to deal with this problem. Deepmind's AlphaGo and Carnegie Mellon University's Libratus are two AI programs that learned how to defeat the best players in games involving imperfect information, like poker, where there's no way to know the cards your opponent is holding. Combining logic with a more intuitive approach, as psychologist Daniel Kahneman²⁵ suggests, is part of how this was possible. My question is (okay, I have a lot of questions, but here's the first): At what point would an AGI program have too little information to win? How far can the information handicap go? What is the mininal amount of information needed to succeed in any endeavor? Related to this, Craig Venter sought to learn "What is the smallest number of genes a cell could possess?"

For the existentialists, "our existence precedes our essence." Sartre said that to avoid anguish and vertigo it is necessary "that I apprehend in myself a strict psychological determinism." ...It's arguable that psychological determinism has indeed already been apprehended, more or less - as Craig Venter said recently, “It is becoming clear that all living cells that we know of on this planet are DNA-software-driven biological machines.” With this insight Venter moved us one step closer to the materialist understanding of life, a step closer to psychological determinism, and perhaps a step farther from the vertigo of Kierkegaard's abyss. Is this the motive that drives us - to be able to stand at the edge of the abyss and not feel dizzy?

Anyone who wades into the deep end of AI discussions encounters questions about the nature of consciousness. While many AI researchers supposed creating an AGI with human capabilities is theoretically possible, not all have. Notable among them has been Roger Penrose's objections to computational algorithms as they relate to AI. It's an interesting question, but essentially immaterial to the current article's main topic of addressing climate change and planetary boundaries. Nonetheless, let's review the main points of that debate.

Penrose might say "a computer working within a fixed formal system can never prove that system’s consistency, but we, “looking in from the outside,” can see that it’s consistent. So the brain must be taking advantage of a non-computational feature of the world, but no such feature exists in modern physics. Where is there room for such a feature at a level the brain could take advantage of? Maybe in some kind of quantum-mechanically enhanced, "non-algorithmic" process." (As a mathematician and physicist I think Penrose found this appealing.) ...Even so, Turing considered and rejected the “Gödelian case against AI” in his 1950 paper, although Penrose was not sufficiently convinced by it.

Biologist PZ Myers also disagrees with Penrose: "Quantum effects matter in that they’re fundamental to how all matter behaves, but cells are big — any counter-intuitive weird quantum effects are going to be negligible in the large-scale bulk activity of a synapse. This is a world where the laws of thermodynamics and electromagnetism rule." Similarly Marvin Minsky in "Conscious Machines" (1991) dismissed Penrose's argument, suggesting the explanation of consciousness lies in the details of these interactions, rather than in new basic principles. Also compare Penrose with Daniel Dennett (Consciousness Explained, Freedom Evolves), Andy Clark (Supersizing the Mind), or see Bringsjord and Xiao. As an aside, Minsky also took a philosophically positive view. He believed that AI might eventually offer a way to solve some of humanity’s biggest problems.

From "Consciousness as a state of matter" by Max Tegmark, 2015:

"Generations of physicists and chemists have studied what happens when you group together vast numbers of atoms, finding that their collective behavior depends on the pattern in which they are arranged: the key difference between a solid, a liquid and a gas lies not in the types of atoms, but in their arrangement. In this paper, I conjecture that consciousness can be understood as yet another state of matter. Just as there are many types of liquids, there are many types of consciousness. However, this should not preclude us from identifying, quantifying, modeling and ultimately understanding the characteristic properties that all liquid forms of matter (or all conscious forms of matter) share.

To classify the traditionally studied states of matter, we need to measure only a small number of physical parameters: viscosity, compressibility, electrical conductivity and (optionally) diffusivity. We call a substance a solid if its viscosity is effectively infinite (producing structural stiffness), and call it a fluid otherwise. We call a fluid a liquid if its compressibility and diffusivity are small and otherwise call it either a gas or a plasma, depending on its electrical conductivity.

What are the corresponding physical parameters that can help us identify conscious matter, and what are the key physical features that characterize it? If such parameters can be identified, understood and measured, this will help us identify (or at least rule out) consciousness “from the outside”, without access to subjective introspection. This could be important for reaching consensus on many currently controversial topics, ranging from the future of artificial intelligence to determining when an animal, fetus or unresponsive patient can feel pain. It would also be important for fundamental theoretical physics, by allowing us to identify conscious observers in our universe by using the equations of physics and thereby answer thorny observation-related questions such as those mentioned in the introductory paragraph."

Emotional intelligence, Symbiosis, and Neuralink:

Tim Urban has written several lengthy articles on AI and Elon Musk's various projects. In one of his most recent articles he speculated at length on the future impact of brain machine interface (BMI) technology, so I'll quote/paraphrase from that article here:

"I’m pretty sure that gaining control over your limbic system is both the definition of maturity and the core human struggle. It’s not that we would be better off without our limbic systems—limbic systems are half of what makes us distinctly human, and most of the fun of life is related to emotions and/or fulfilling your animal needs—it’s just that your limbic system doesn’t get that you live in a civilization, and if you let it run your life too much, it’ll quickly ruin your life.

"Often, the battle in our heads between our prefrontal cortex and limbic system comes down to the fact that both parties are trying to do what’s best for us—it’s just that our limbic system is wrong about what it thinks is best for us because it thinks we live in a tribe 50,000 years ago. Ramez Naam talks about how a brain interface could also help us win the discipline battle when it comes to time: "We know that stimulating the right centers in the brain can induce sleep or alertness, hunger or satiation, ease or stimulation, as quick as the flip of a switch. Or, if you’re running code, on a schedule. (Siri: Put me to sleep until 7:30, high priority interruptions only. And let’s get hungry for lunch around noon. Turn down the sugar cravings, though.)"

"Ramez also emphasized that a great deal of scientific evidence suggests that moods and disorders are tied to what the chemicals in your brain are doing. Right now, we take drugs to alter those chemicals, and Ramez explains why direct neural stimulation is a far better option: "Pharmaceuticals enter the brain and then spread out randomly, hitting whatever receptor they work on all across your brain. Neural interfaces, by contrast, can stimulate just one area at a time, can be tuned in real-time, and can carry information out about what’s happening." Depression, anxiety, OCD, and other disorders may be easy to eradicate once we can take better control of what goes on in our brain.

"If we can just use engineering to get neurons to talk to computers, we’ll have done our job, and machine learning can do much of the rest. Which then, ironically, will teach us about the brain. When Elon refers to a “digital tertiary layer,” he’s considering our existing brain having two layers—our animal limbic system (which could be called our primary layer) and our advanced cortex (which could be called our secondary layer). The neuralink interface, then, would be our tertiary layer—a new physical brain part to complement the other two. ...In a future world made up of AI and everyone else, he thinks we have only one good option: To be AI."

Insofar as the goal of a brain machine interface is to allow greater control over the limbic system, it essentially would help to realize the central goal of Buddhist transhumanism and the "middle path" of Buddhism in general. It is interesting to reflect on how meditative practices do in fact tend to focus on controlling our limbic system.

Artificial intelligence benefits agriculture

Quotable lines: Hyperbole or Reality?

• “After humanity spent thousands of years improving our tactics, computers tell us that humans are completely wrong. I would go as far as to say not a single human has touched the edge of the truth of Go.” (Ke Jie, 2017)
• “AI is the new electricity.”
• "AI is intended to be a learning machine. It is this capacity to learn that marks the newness of the current technological era, and this capacity of learning that makes it possible to even speak of AI agency.⁸" (Urvashi Aneja, 2017)
• “All life is problem solving.” (Karl Popper)
• "And as more devices and sensors become connected, we will learn even more about the world around us. This ability to make sense of all this data could help us cure disease, tackle climate change, grow food more efficiently and generally run our lives in a much smarter, more sustainable way, proponents believe.⁹"
• “Any sufficiently advanced technology is indistinguishable from magic.” (Arthur C. Clarke's third law)
• “Because your refrigerator draws power in a different way than your coffee pot, over time Sense can detect and track individual appliances in your home.”64 (Sense home energy monitor)
• Buddhist transhumanism is “a movement that seeks to attain the traditional Buddhist goals of reducing suffering and realizing Awakening, but with the assistance of scientific knowledge and technological means.32” (Michael LaTorra, James Hughes)
• "Closed-loop systems are when you take an action, you measure the results, and you change your action accordingly. Systems with closed loops have feedback loops; they self-adjust and quickly stabilize in optimal conditions. ...As we get better and better at measuring the world we are going to be able to close the loop in industry, agriculture, and the environment. We now need to extend the tendrils of the Internet out into the physical world and start measuring things, act on that information, and then make smarter choices about our planet. ...It goes both ways: The tendrils of the Internet reach out through sensors, and then these sensors feed back to the Internet. The sensors get smarter because they're connected to the Internet, and the Internet gets smarter because it's connected to the sensors. This feedback loop extends beyond the industry that's feeding back to the meta-industry, which is the Internet and the planet.¹⁰" (Chris Anderson, 2017)
• "Consciousness is the way information feels when being processed.75" (Max Tegmark)
• Cyber-physical systems play an important role in the transition from fossil fuels to renewable energy sources, with wind and solar power increasingly covering energy demand. However, the amount of electricity contributed by these energy sources naturally fluctuates. Thus, in order to ensure that the supply still satisfies the demand, the electricity must be transmitted using a clever mechanism. This approach is based on a vast energy information network that combines the regulation of the power grid with consumers, electricity producers and energy storage devices. The important components of this gigantic system include sensors in the form of smart meters located in households, along with information and communication technology, as well as adaptive arithmetic techniques.67
• “Data is the new coal.”
• “Dataism is most firmly entrenched in its two mother disciplines: computer science and biology. Of the two biology is the more important. It was biology's embrace of Dataism that turned a limited breakthrough in computer science into a world shattering cataclysm that may completely transform the very nature of life. You may not agree with the idea that organisms are algorithms, and that giraffes, tomatoes and human beings are just different methods for processing data. But you should know that this is current scientific dogma, and it is changing our world beyond recognition. Not only individual organisms are seen today as data processing systems, but also entire societies such as beehives, bacteria colonies, forests and human cities.” (Yuval Harari, “Homo Deus,” p. 373)
• Ecological rationality seeks to create structural coupling [between agent and domain] and to bring forth a world that allows structural coupling to be maintained. (Gateway to The Global Garden: Beta/Gamma Science for Dealing with Ecological Rationality, Niels Röling; see also David Grinspoon's Sapiezoic¹⁴)
• Expanding circle of empathy (Peter Singer)²³
• "Fairness Through Awareness70"
• "Far back in human history, natural selection discovered that, given the particular problems humans faced, there were practical advantages to having a brain capable of introspection. Likewise machine programmers may well discover that, when and if machines face similar problems, the software trick that works for humans will work for them as well."40 (Nicholas Humphrey)
• “Games are the fruit fly of AI research.²²” (Chess as the Drosphila of AI, John McCarthy)
• “I firmly believe that the important things about humans are social in character and that relief by machines from many of our present demanding intellectual functions will finally give the human race time and incentive to learn how to live well together.¹⁸" (Merrill Flood)
• "I suspect, as always, that the most interesting questions are the ones we haven't yet thought of. What questions will machines focus on when they get to choose the questions as well as the answers?"72 (Lawrence Krauss)
• “I would advocate prioritizing vital social questions over data availability.” (Hanna Wallach)
• “If the operating system was the first run time, the second run time you could say was the browser, and the third run time can actually be the agent.¹¹” (Satya Nadella, Microsoft CEO, 2017)
• “In short, I think the application of iterative algorithms (e.g., machine learning, directed evolution, generative design) to build complex systems is the most powerful advance in engineering since the Scientific Method.29” (Steve Jurvetson)
• “In some sense, the humanism of AI will eventually be what brings us together.”65 (Manuela Veloso)
• “It can also be maintained that it is best to provide the machine with the best sense organs that money can buy, and then teach it to understand...” (Alan Turing, 1950)
• “It is becoming clear that all living cells that we know of on this planet are DNA-software-driven biological machines.” (Craig Venter) ["Software is eating the world." (Marc Andreessen)]
• It’s all about the algorithm. Flat mirrors won’t focus the sun, so eSolar’s software creates a “dynamic parabola” out of the entire heliostat field to concentrate the sunlight on the power tower.38
• “Just as animals dream of positively or negatively rewarding events more frequently, our agents preferentially replay sequences containing rewarding events.”49 (“Hybrid computing using a neural network with dynamic external memory,” 2016 paper)
• “Let us then conclude boldly that man is a machine and that the whole universe contains only one substance variously modified.” (L'homme Machine, 1748, Julien Offray de La Mettrie)
• Living organisms are closed-loop systems that act to keep perceptual variables in pre-specified states, protected from disturbances caused by variations in environmental circumstances. (Perceptual Control Theory model of behavioral organization, 1973, William Powers)
• "Many AI systems are built with a framework that maximizes expected utility. Such an AI system estimates the current state of the world, considers all the possible actions it can take, simulates the possible outcomes of those actions, and then chooses the action that leads to the best possible distribution of outcomes."71 (Peter Norvig)
• Neural nets are essentially “spreadsheets on steroids.”41 (John Launchbury)
• “One of the unintended consequences of big data and the internet of things is that some things will become visible and compel us to confront them.”63 (Genevieve Bell)
• "One very simple difference between human judgement and algorithms is that algorithms are noise-free. If you present an algorithm with the same problem twice you're going to get the same output. There's a huge amount of noise in human decision making at all levels. The advantage of algorithms over humans is primarily because algorithms are noise free. And now in addition they are based on very large databases. ...There is a lot of potential.26"(Daniel Kahneman)
• "Over the past decades, computers have broadly automated tasks that programmers could describe with clear rules and algorithms. Modern machine learning techniques now allow us to do the same for tasks where describing the precise rules is much harder." AI lets a computer automate tasks that you can’t really put into clear “if, then” rules, such as, how do you drive a car? Or which photos contain a cat? That’s what AI can already do. (Jeff Bezos)
• Robots are “always polite, they always up-sell, they never take a vacation, they never show up late, there’s never a slip-and-fall, or an age, sex, or race discrimination case.”60 (Andrew Puzder, Trump’s nominee for Secretary of Labor)
• “Some economists argue this combination of fast-expanding data sets, machine learning and ever-increasing computing power should be classified as an entirely new factor of production, alongside capital and labour.²¹"
• “Technology is made by humans. If we modify our body with human creations we become more human.” (Neil Harbisson)
• Technology of Perspective taking (Steven Pinker)
• "The Department of Defense is facing challenges that demand more intelligent, autonomous, and symbiotic systems. Explainable AI—especially explainable machine learning—will be essential if future warfighters are to understand, appropriately trust, and effectively manage an emerging generation of artificially intelligent machine partners.34" (David Gunning)
• “The last 10 years have been about building a world that is mobile-first. In the next 10 years, we will shift to a world that is AI-first.”56 (Sundar Pichai, CEO of Google, October 2016)
• "The paradox is that it has taken computers and machines to do what spiders and pine cones have been doing since long before there were architects.27" (Jonathan Glancey)
• “There is no AI without robotics.” (Refers to the “embodiment problem.”)
• Ultimately, new connected and intelligent capacities allow us to, in Jeff Immelt’s words, to “find meaning where it did not exist before”. And not only meaning: value.39
• “Use cases for machine learning and deep learning are only limited by our imaginations.”58 (Kazunori Sato)
• “Vertical AI startups solve full-stack industry problems that require subject matter expertise, unique data, and a product that uses AI to deliver its core value proposition. Massive and timely problems are lurking in every major industry outside tech: financial services, life sciences and healthcare, energy, transportation, heavy industry, agriculture, and materials. These startups will solve high level domain problems.¹²” (Bradford Cross, 2017)
• "We might gain energy independence by developing much smarter infrastructure, as Google subsidiary DeepMind did for its parent company’s power usage. The opportunities are too great to ignore."55 (Sandra Upson)
• "We need to get really smart about our energy use and really smart about how we create energy. The only way we can do that is by creating an ecosystem where there are a lot of ideas working together. And we need to start valuing that kind of community of ideas instead of the one hero that is going to save us all."69 (Leila Madrone)
• “When wireless is perfectly applied the whole earth will be converted into a huge brain...” (Nikola Tesla, 1926)
• “With technology like this, you can reduce the risk of the unknowns - the unknowns become part of the equation... I haven't before seen a system that's better than a human at handling the unknown.31” (Abdul Razack, 2017; describing Libratus)
• “You could work with a computer to solve a problem that we've never faced before. For instance, climate change. We're not doing a very good job on our own, we could certainly use all the help we can get. That's what I'm talking about, technology amplifying our cognitive abilities so we can imagine and design things that were simply out of our reach. ...If this is the future, the Augmented Age, what will that look like? I think we're going to see a world where we're moving from things that are fabricated to things that are farmed. Where we're moving from things that are constructed to that which is grown. We're going to move from being isolated to being connected. And we'll move away from extraction to embrace aggregation. I also think we'll shift from craving obedience from our things to valuing autonomy.¹³” (Maurice Conti, Feb 2017)

Quotable lines from the 2015 Edge question: What do you think about machines that can think?

• Alex Pentland: "Creation of an effective Global Artificial Intelligence (GAI) is critical because today the entire human race faces many extremely serious problems. The GAI we have developed over the last four thousand years, mostly made up of politicians and lawyers executing algorithms and programs developed centuries ago, is not only failing to address these serious problems, it is threatening to extinguish us.
We must figure out how to build broadly democratic systems that include both humans and computer intelligences. In my opinion, it is critical that we start building and testing GAIs that both solve humanity's existential problems and which ensure equality of control and access. Otherwise we may be doomed to a future full of environmental disasters, wars, and needless suffering."
• Giulio Boccaletti: "One area where the convergence of need, urgency, and opportunity is great is in the monitoring and management of our planetary resources. Despite the dramatic increase in cognitive and labor productivity, we have not fundamentally changed our relationship to Earth: we are still stripping it of its resources to manufacture goods that turn to waste relatively quickly, with essentially zero end-of-life value to us. A linear economy on a finite planet, with seven billion people aspiring to become consumers—our relationship to the planet is arguably more productive, but not much more intelligent than it was a hundred years ago.
Understanding what the planet is doing in response, and managing our behavior accordingly, is a complicated problem, whose solution is hindered by colossal amounts of imperfect information. From climate change, to water availability, to the management of ocean resources, to the interactions between ecosystems and working landscapes, our computational approaches are often inadequate to conduct the exploratory analyses required to understand what is happening, to process the exponentially growing amount of data about the world we inhabit, and to generate and test theories of how we might do things differently."
• Peter Norvig: "Many AI systems are built with a framework that maximizes expected utility. Such an AI system estimates the current state of the world, considers all the possible actions it can take, simulates the possible outcomes of those actions, and then chooses the action that leads to the best possible distribution of outcomes."
• Luca De Biase: "If only profit counts, then externalities don't count: cultural, social, environmental externalities are not the concern of financial institutions... [but] if new narratives emerge from an open, ecological approach, if they can grow in a neutral network, they will shape the next generation of artificial intelligences in a plural, diverse way, helping humans understand externalities. Artificial intelligence won't challenge humans as a species, it will challenge their civilizations."
• Joichi Ito: "Animist religions may have less trouble dealing with the idea that maybe we're not really in charge. If nature is a complex system in which all things—humans, trees, stones, rivers and homes—are all animated in some way and all have their own spirits, then maybe it's okay that God doesn't really look like us or think like us or think that we're really that special.
Human beings are part of a massively complex system—complex beyond our comprehension. Like the animate trees, stones, rivers and homes, maybe algorithms running on computers are just another part of this complex ecosystem. ...it may never be clear who's in charge - us, or our machines. But maybe we've done more damage by believing that humans are special than we possibly could by embracing a more humble relationship with the other creatures, objects, and machines around us."
• Dirk Helbing: "Intelligent machines would probably learn that it is good to network and cooperate, to decide in other-regarding ways, and to pay attention to systemic outcomes. They would soon learn that diversity is important for innovation, systemic resilience, and collective intelligence. Humans would become nodes in a global network of intelligences and a huge ecosystem of ideas."
• Molly Crockett: Human brains are incapable of solving the interpersonal utility comparison problem. But if we create machines that are better empathizers than us, we could solve it and create better social contracts, improve self-control. Achieving this feat could be essential to our survival.
• Marti Hearst: "We will find ourselves in a world of omniscient instrumentation and automation... Let's call this world "eGaia" for lack of a better word."
• Nicholas A. Christakis: "Culture is the earliest sort of intelligence outside our own minds that we humans created... Culture fosters collective action or makes life easier by positing assumptions on which we can base our lives. Moreover, we typically take culture for granted, just as we take nascent forms of AI for granted and just as we will likely take advanced forms of AI for granted. Gene/culture coevolution might even provide a model for how we and thinking machines will get along over many centuries - mutually affecting each other and coevolving."
• John Tooby: "Because we evolved with certain adaptive problems, our imaginations project primate dominance dramas onto AIs, dramas that are alien to their nature... We could transform them from Buddhas—brilliant teachers passively contemplating without desire, free from suffering—into motivated intelligences capable of taking action (MICTAs), seething with desire and able to act. That would be insane—we are already bowed under the conflicting demands of people.
The foreseeable danger comes not from AIs but from those humans in which predatory programs for dominance have been triggered, and who are deploying ever-growing arsenals of technological (including computational) tools for winning conflicts by inflicting destruction."
• Steve Fuller: Historic struggles for social justice are likely to be on the side of the insubordinate machines.
• Timothy Taylor: "If we tacitly assume that a machine is something produced by humans, we underestimate the degree to which machines produce us, and the fact that thought has long emerged from this interaction, properly belonging to neither side (and thinking there are sides may be wrong too)." [This reminds me of a paper titled “The Advantages and Disadvantages of Being Domesticated” by Colin Groves. He explained how domestication mutually shapes the evolution of both species. Technological innovations, such as fire, tools, and culture may have had a similar effect.]
• Nicholas Humphrey: "Far back in human history, natural selection discovered that, given the particular problems humans faced, there were practical advantages to having a brain capable of introspection. Likewise machine programmers may well discover that, when and if machines face similar problems, the software trick that works for humans will work for them as well. But what are these problems, and why is the theatre of consciousness the answer?"
• Stanislas Dehaene: "Global workspace and theory-of-mind... I predict that, once a machine pays attention to what it knows and what the user knows, we will immediately call it a "thinking machine", because it will closely approximate what we do."

The TL;DR version:

What is preventing us from effectively addressing societal problems?
We are part of a massively complex system—complex beyond our comprehension. Human brains are incapable of solving the interpersonal utility comparison problem. We are struggling to strategically make, evaluate, and prioritize long-term plans. We struggle to maintain sustained focus on multiple things at the same time and their large scale (and small scale) interactions. We struggle to synchronize interactions at the individual, social, and environmental levels. These are complicated problems hindered by colossal amounts of imperfect information.

What can we do to address that?
Build broadly democratic systems that include both humans and computer intelligences to A) process the exponentially growing amount of data about the world we inhabit, and B) generate and test theories of how we might do things differently. The application of iterative algorithms (e.g., machine learning, directed evolution, generative design) to build complex systems is the most powerful advance in engineering since the Scientific Method. In fact, the three essential components of the scientific method (observation, hypothesis generation, and experimentation) can be automated.

How might we respond differently?
The first step is to create closed-loop systems. You take an action, you measure the results, and you change your action accordingly. Systems with closed loops have feedback loops; they self-adjust and quickly stabilize in optimal conditions. As we get better at measuring the world we are going to be able to close the loop in industry, agriculture, and the environment, extending beyond to the planet as a whole. A closed-loop system doesn't necessarily require iterative algorithms, or even democracy to ensure equality of control and access, but if we want to maintain our current quality of life, we'll need both.

Glossary of a few AI terms:

• Adaptive learning (individualized learning)
• Affective computing30 (emotional intelligence)
• Agricultural robotics (agrobot, farmbot)
• Ambient intelligence
• Artificial general intelligence (AGI)
• Artificial/machine intuition (Daniel Kahneman)
• Augmented ecosystems
• Automation (home, vehicle)
• Biobot (cyborg stingray, robotics and tissue engineering, synthetic lifeform)
• Biohybrids43 (plant robot biohybrids, Flora Robotica; symbots, ecobot)
• Biological computers (Dan Nicolau)
• Brain-Machine Interface (BMI)
• Church-Turing thesis
• Closed-loop systems
• Cognitive computing (IBM’s AI)
• Collaborative robots (cobots, Franka Emika48)
• Computational chemistry61
• Convolutional neural networks
• Cyber-physical systems
• Cyber security
• Cybernetics (the art of ensuring the efficacy of action, Louis Couffignal)
• Cyberwarfare
• Deep Learning (figuring out how, given a massive enough dataset, to capture interactions and latent variables)
• Demand response
• Embodiment problem (intelligence and embodiment are tightly coupled)
• Evolutionary computing and algorithms
• Explainable AI33 (XAI, the ability of machines to explain their rationale, characterize the strengths and weaknesses of their decision-making process, and convey a sense of how they will behave in the future; transparency in AI)
• Farm hacking^{15, 17}
• Feature specification and optimization
• Fourth Industrial Revolution
• Generative design and modeling (a collaboration between human and machine)44
• Imperfect information
• Internet of Things
• Machine ethics¹
• Machine learning (Tom M. Mitchell’s 1997 definition of machine learning: “A computer program is said to learn from experience E with respect to some class of tasks T and performance measure P, if its performance at tasks in T, as measured by P, improves with experience E.”)
• Meta-learning
• Mixed/augmented reality66
• Neural networks
• Positive feedback/reinforcement (Recursive self programming; AI designing new AI software; Robots that teach each other62; the automation of automation; code that writes itself¹⁶)
• Precision agriculture, field robotics, & agroecology
• Predictive analytics
• Reinforcement learning algorithm
• Robot milking parlours (robotic milking)45
• Robotic waste management74
• Robotics
• Robotics and automation in meat and poultry plants47
• Scenarios-based energy planning
• Smart technology (cities37, buildings)
• Substrate-independence75
• Swarm intelligence/robotics
• Technological unemployment
• Technogaianism
• Technorealism

The film Robot and Frank (2012)

AI related organizations:

• Agility Robotics (Cassie, bipedal robot) • Aktivhaus² (uses a predictive, self-learning energy management system that allows the house to react to its environment and continuously optimize its performance)
• Australian Centre For Field Robotics (Salah Sukkarieh, James Underwood)
• Boston Dynamics (SpotMini, Handle)
• Carnegie Mellon University (Libratus²⁴, Manuela Veloso)
• DARPA (Perdix mini-drones, swarm robotics)
• DeepMind (Demis Hassabis; AlphaGo52, LipNet lipreading)
• Ekotrope³ (developed an algorithm that optimizes all the capital investments in a house that have to do with energy utilization)
• Ethics and Governance of AI Fund (MIT Media Lab)36
• FarmBot (open-source CNC farming machine)
• Flora Robotica42 (replicating the ant-fungal mutualism)
• GardenBot⁴ (open source garden monitoring system)
• Ghost Robotics (Minitaur)
• Google (Multilingual Neural Machine Translation System50, aka “Google Translate” interlingua; TensorFlow; SpatialOS57)
• IBM (Deep Blue, Watson, “Green Horizons” Initiative⁵)
• Institution for Computational Design (Stuttgart University)
• Institute for Ethics and Emerging Technologies (Cyborg Buddha Project)
• Lego Robotics (FIRST Lego League with Lego Mindstorms)
• Monsanto (bought digital agriculture company "The Climate Corporation" in 2013; "Green data revolution")
• Poverty and Technology Lab28 (Stanford)
• Siemens68 (City Performance Tool, CyPT)

Selected papers, articles, presentations, etc.:

• "AlphaGo: A Documentary" (2017)
• “And everyone gets a robot pony!”54 by PZ Myers (obstacles to AGI)
• “Brain activity is too complicated for humans to decipher. Machines can decode it for us,”53 by Brian Resnick
• Consciousness as a State of Matter77, by Max Tegmark, 2015
• Consciousness: here, there and everywhere?76 by Giulio Tononi, Christof Koch, 2015
• “Deep Thinking: Where Machine Intelligence Ends and Human Creativity Begins,” by Garry Kasparov
• “Explainable Artificial Intelligence (XAI)35" DARPA-BAA-16-53, August 10, 2016
• “Gateway to The Global Garden: Beta/Gamma Science for Dealing with Ecological Rationality⁶” by Niels Röling
• “Homo Deus: A Brief History of Tomorrow,” by Yuval Harari
• “How Blockchains could transform Artificial Intelligience,” Trent McConaghy59
• “Intelligent Biosphere⁷,” by Drew Purves, Neural Information Processing Systems (NIPS) Conference, 2016
• “Parallel computation with molecular-motor-propelled agents in nanofabricated networks” by Dan Nicolau, 2015
• “Preparing for the Future of Artificial Intelligence²⁰" White House report, 2016
• “The Fourth Industrial Revolution: a primer on Artificial Intelligence (AI),” by David Kelnar
• “What constraints are needed to prevent AI from becoming a dystopian threat to humanity?”46 by David Brin