Wednesday, 28 December 2016

Researching the Far Future

Where is Humanity Going and How Will We Get There?

N.B. Holiday post, off topic and longer than usual. The construction industry does not get a mention, although one assumes there will be a construction industry in the future.

While the idea of researching the future may seem like an oxymoron, thinking about the implications of current issues and trends in technology, and applying the limited amount of information we have about potential new technologies, is not a pointless exercise. While it is true that many, if not most, of these books date quickly, that is because they focus on relatively short-term effects or are overly speculative. There have, however, been a surprising number of notably prescient writers who have tried to think realistically about technological development over time, Moore’s Law and Metcalf’s Law for transistors and networks are good examples

As it happens, 2016 saw a number of particularly interesting books on the future, Although the four reviewed here are very different they share certain methodological similarities. In each case they use what we already know from both history and the present in a sensible way, by not relying on some hypothetical breakthrough like fusion power or conscious machines (as opposed to machine intelligence). Secondly, they avoid specific predictions and forecasts, rather presenting their view of what is possible based on what is understood to be technologically feasible. And third, they do not propose some fundamental restructuring of human nature or the organisation of society and the economy, although they are generally aware of the feedback between them and technological change, and the importance of that link.

The similarities and differences between these four authors are interesting in their own right. As researchers they seriously address the topics and issues chosen, and each has singular insights that should be considered if the far future is of interest to you. The books from Peter Frase, Kevin Kelly, Robin Hanson and Yuval Noah Harari are discussed below.

Peter Frase describes himself as a lapsed academic sociologist. His Four Futures: Life After Capitalism presents scenarios for the next few decades. This is the most conventional of these books, in that his scenarios (two good, two bad) are constructed following a widely accepted approach and reflect his Marxist perspective. He describes his approach as “social science fiction”, and his futures are interesting and plausible because they are magnified versions of the present.

Communism - a society that is so productive and egalitarian that nobody has to work. As a result of unlimited clean energy, robots and automation we have the material basis for a post-work, post-scarcity and post-carbon world. In fact, this is the sort of Communism originally expounded by Karl Marx, where we get “from each according to his ability, to each according to his need”.

Rentism - abundance exists, but in a wold where a small elite retains dominance by owning and monopolising the technology used to produce it. This monopoly is maintained by owning not just the robots, but the data that tells them how to do their job. In a world where you can automate anything, you can encode any task as information. In this world the software that robots run on can be copyrighted as intellectual property, and to use them you have to pay a fee. That means people need a job, however there aren’t enough jobs because all the work is done by machines. Therefore, it is only workers required to sustain the regime that get paid, and most people will be unemployed or under employed. The economy will be stagnant because it requires consumers but the jobless masses won’t be able to afford to consume. This scenario also assumes abundant clean energy.

Socialism - in this third future automation exists but the breakthrough to carbonless energy doesn’t happen, which means we deal with climate change through a massive, state-lead campaign to reshape infrastructure and consumption. Because climate change affects different groups of people differently, some will do much better than others, particularly those in less vulnerable geographies or can afford to insulate themselves from the effects of climate change. Through policies like a universal basic income and market planning a democratic, egalitarian outcome is produced.

Exterminism - here we have the robots and scarcity of socialism without the egalitarianism. The result is a neo-feudal dystopia, where the rich retreat to fortified enclaves and everyone else is left outside in a hot, damp hell of a rapidly warming planet. Without needing labour, because of automation, the rich may have no reason to try and save the masses.

One of the strengths of this book is the insistence that creating the future we want is ultimately as much about politics as technology. In these four scenarios automation and machine learning are the drivers, what changes is the political and ecological context. Who owns the robots and how climate change affects the outcomes will still be issues after the robots arrive. Frase argues that the future is (can be?) what we make it, and between nihilism and utopianism lies politics, with its patterns of long, slow struggle punctuated by an occasional social explosion. A sophisticated argument informed by history that is intuitively appealing, despite the ideological baggage.

Where the weakness of a class based analysis really shows is in the missing parts of his scenarios. Although automation and robots feature strongly in his future, the wider impact of massive amounts of information and huge data flows does not. It’s a strangely limited view of technology, tailored to suit a particular argument. The second missing element is the potential of technologies like genetic engineering, nanotechnology and computer interfaces to upgrade human abilities.

Kevin Kelly was a founder of Wired magazine in 1992, which documented the rise of new tech over the next two decades, and before that was editor of the Whole Earth Review. His 2010 book What Technology Wants argued tech evolves in a fitness landscape with a recognisable ecology like biology in a natural system, and suggested 12 trajectories. In this year’s book The Inevitable: Understanding the 12 Technological Forces That Will Shape Our Future he pushes that view to another level, envisioning a world utterly mediated by data that constantly streams and flows around us while ubiquitous technology reshapes society.

The chapter titles are verbs and capture his ideas about this future, which are, in order: becoming, cognifying, flowing, screening, accessing, sharing, filtering, remixing, tracking, questioning and beginning. The 12 forces interact and reinforce each other. They overlap, so that sharing encourages flowing and depends on it, cognifying requires screening, screening is inseparable from interacting, everything is remixed, and so on.

He argues that there are broad historical trends we can observe over the last few decades that will continue over the next few decades, and that the dynamics of technology tend to favour certain outcomes. He also believes that these trends and forces have accelerating momentum, and he wants to “expose these roots of digital technology because from them will issue the enduring trends of the next three decades”. He is interested in what he calls aggregate forces, rather than specifics. The fundamental change will be turning products into services and processes, so for example cars become transportation services, regardless of the type of car it is. This is his idea of becoming.

The book does not make predictions, these forces are “trajectories not destinies”, and he argues that understanding these will lessen the social and economic displacement they cause while increasing positive outcomes. Many will disagree, and the sometimes hyperbolic language and techno-optimism can seem strangely naive in a world beset today with so many issues to deal with.On the other hand, there is a global middle class of about one billion people whose jobs increasingly involved moving bits of information around and the first generation, the millennials, that have grown up in s connected enviroment.

Kelly believes that the combination of people, connectivity and silicon is creating a global super organism, and we are at the beginning of a century long process of creating “a new mind for an old species”. This is not a utopia, there will be firewalls, corporate monopolies, unequal access, and the rich will be privileged, as always. Nevertheless, he is optimistic that centralised authority and uniformity will diminish, while the techno-cultural forces he discusses come to dominate institutions and peoples’ lives. Maybe, but he heavily discounts the forces of reaction to his trends, because he sees them as inevitable and irresistible.

Robin Hanson is known for his work in creating prediction markets in the 1990s, but he is a genuine polymath with a background in physics, early AI and economics. His book The Age of Em: Work, Love and Life When Robots Rule the Earth is one of the strangest, most challenging books I have ever read. It is an extended analysis of a single ‘baseline’ scenario of brain emulations, or ‘ems’, a technology he sees as feasible within the next century, with implications so great they should be seriously considered.

Whole brain emulation is one of the possible pathways to artificial intelligence. Unlike alternatives such as building an artificial brain neuron by neuron or designing a conscious machine, a brain emulation comes with a personality and previous life, and will go on to have many more experiences. Hanson’s scenario is actually about what a society made up of many artificial intelligences might be like, so there is little discussion of what humans do and how they interact with ems.

These brain emulations are, of course, digital copies of an existing brain, so they can be easily created, copied and shared. They will live in a virtual world, but can inhabit robotic bodies when required. They will think and feel like humans, remembering their past and be capable of learning. They do, however, come at a cost in terms of energy, hardware and resources, which means they live in “a few tall hot densely packed cities where the volume is about equally split between racks of computer hardware and pipes for cooling and transport”. The interplay between cost and performance underpins much of his, highly speculative, analysis.

This is because em minds can run at different speeds, from perhaps one million times slower than humans to one million times faster, and the cost to run an em is proportional to its speed. Faster ems have higher status, and for those with robotic bodies their bodies are proportionally smaller. A typical em running a thousand times faster than human speed would be 2mm tall. The implications of this for em subjective time and capacity for work are analysed at length. At one thousand times faster than human speed, one human year is a millennium to an em. They will, however, require rest, sex and relaxation, and eventually retire at some very slow speed. All this is analysed at length.

The basic characteristic of ems is that they are a scan of a specific brain, thus carry in them the experience, talent and traits of an individual. Therefore, only the world’s best are likely to be copied, perhaps less than a thousand people, and those ems will them copy themselves to form clans: “Strong competitive pressures result in most ems being copies of the thousand humans best suited for em jobs. Ems are mostly very able focused workaholics at the level of Olympic medallists, billionaires or heads of state.” They are mostly happy, but can be moody, just like people. They have hierarchies, partners and long-term relationships.

The em economy will grow very quickly once started, perhaps doubling every month or so, driven mainly by population growth. Their clans specialise, have big financial and reputational investments in their subsidiary firms, and ems have less variation in wages and productivity. With many copies, em wages are at subsistence levels (i.e. running costs) and well below human wages. Humans will have to live off their investments in the em economy, which has high returns from its rapid growth, and some of the em economy does projects and production useful to humans.

Managing the transition to an em economy will be challenging, particularly as it has the rapid expansion characteristic found in other AI scenarios.  This is an important issue, so Hanson's scenario allows the question of where the first em city might be built. and his view that if only “a few suitable enough places give the new em economy sufficient support, opposition in other places would then be quickly overwhelmed by very rapid growth in those few areas.” Once started, the transition will happen one way or another. In fact what this depicts is an entirely digital economy, with ems managing all production and distribution. Exactly what and why they are producing anything is not clear.

This future is one of an em dominated economy, because they are so efficient, coordinated and well organised they take over. Hanson finishes his book with suggestions on how humans can prepare for the coming of an em economy, the policy options that could be considered, and strategies to use. He sees “many ways for individuals to work to help deal with the possible coming of an em world”, but “to succeed in this new world, prepare to become what it needs”. Whatever that may be.

Hanson seems to see the em era as the next stage in the evolution of humanity, a continuation of the progression from foraging to farming to the industrial age of the present. He points out that lives in the future may be as different from our lives today as we are from foragers and farmers: “If you understood just how different your ancestors were, you’d realise that you should expect your descendants to seem quite strange”. This view is one he shares with Yuval Noah Harari, who also starts with a similar long view of stages of development.

Harari is the author of Sapiens: A Brief History of Mankind, which explored the development of society and economy as humans’ abilities increased over millennia. From the prehistoric Cognitive Revolution to the Industrial Revolution and the discovery of the scientific method he draws a picture of progress and challenges. His follow-up is Homo Deus: A Brief History of Tomorrow, an equally bold and expansive take on the big issues in the evolution of humanity, technology and society. He starts with the observation that of the perhaps 70 billion humans to have existed so far, the great majority died from famine, plague and war. That all changed in the 20th century, and now humanity needs new challenges.

The first is aging and dying. The science of regenerative medicine is advancing quickly, with the unravelling of the molecular structure of cells, genetic engineering and nanotechnology well underway. Over the coming decades life expectancy will continue to rise and at some point death will become much less common, often due to misadventure. This is Harari’s technique, take a topic like death, designer babies, animal minds, consciousness or free will, and investigate it using the latest science. Many people would find his conclusions disturbing, as humanity reaches for his goals of immortality, bliss and divinity. Bliss is happiness and divinity is power, essentially power over nature. Both largely come from the combination of upgraded humans with machine intelligence, producing Homo Deus. His website leads with "History began when humans invented gods, and will end when humans become gods".

The decoupling of intelligence from consciousness is an important part of the scenarios Harari investigates. An intelligent machine is not a conscious one, we have lots of the former and none of the latter. Meaning and purpose are human qualities: “Yet over decades and centuries the web of meaning unravels and a new web is spun in its place. To study history means to watch this spinning and unravelling of these webs, and to realise that what seems to people in one age the most important thing in life becomes utterly meaningless to their descendants.”

The key to humanity’s success as a species is large-scale cooperation. This allows institutions to create meaningful stories about their purpose, culture, nation or religion. Harari calls these stories ‘fictions’. However, powerful institutions can evolve into entrenched bureaucracies: “Corporations, money and nations exist only in our imaginations. We invented them to serve us; how come we find ourselves sacrificing our lives in their service?”

The answer is what he calls humanism, a broad-based shift that started in the 17th century when we gave up believing in a cosmic plan to gain power over the physical world. We exercise that power through scientific progress and economic growth, and have thus solved the problems of plague and famine, and haven’t had a large war for some time. Humanism sees life as a gradual process of inner change, developing knowledge through a variety of experiences, and puts people at the centre. Because those experiences belong to an individual, they should be given as much freedom as possible to experience the world. Thus, individual liberty within a democratic system is the ideal state.

A humanist civilisation will invest in science to increase lifespans, improve peoples’ cognitive abilities and research new, powerful technologies. If genetic engineering and artificial intelligence are the means to achieve those goals, they will be pursued. The irony is that these technologies will also allow us to design the human experiences, the emotions and desires of people, that humanism places at the centre of the universe. At the end of chapter eight we come to the crux of the matter: “We are about to face a flood of extremely useful devices, tools and structures that make no allowance for the free will of individual humans. Can democracy, the free market and human rights survive the flood?”.

The 20th century provided mass education and mass health, to provide the manpower needed in factories and armies. In the 21st century there may be much less demand for the lower skilled workers that system produced. Therefore, elites maybe more interested in upgrading a few million rich people, rather than investing in fixing the problems of hundreds of millions of poor people. This would result in a small elite of upgraded superhumans with exceptional physical, emotional and intellectual abilities, and they would have quite different experiences to the rest of humankind. The alternative would be large scale upgrading, with cultures like the Amish being the exception. There is already an active biohacking movement.

Harari thinks upgraded humans could develop two new techno-religions to replace humanism, in the same way as humanism replaced belief systems and theology. The first, techno-humanism, believes we should use technology to upgrade to a superior human model, from Homo sapien to Homo deus. With upgrades like genetic engineering, nanotechnology and brain-computer interfaces, humans will be able to hold their own against intelligent but not conscious machines, and their combined abilities will spark a new cognitive revolution as important as the agricultural revolution 7,000 years ago. It’s a recognisable version of evolutionary humanism that is reasonably comforting.

Unlike the alternative. He asks “What might replace desires and experiences as the source of all meaning and authority?” The answer is information, and the religion that worships information is Dataism. The universe is reduced to data flows and we understand life as data processing and decision-making. Organisms are algorithms, markets and politics are information flows, markets are flexible and adaptive, politics is not. Freedom of information becomes the new greatest good, replacing liberty and equality.

In this world, the generating and sharing of data is the best contribution we can make. Whether it’s answering emails, going to work, or posting on Facebook, it all contributes to the data flow. One bit of data is as good as another, but only humans share their experiences as data. However, Dataism sees those experiences as biochemical algorithms, purely physical processes that can be known and understood, and uploaded into a universal Internet-of-All-Things.

Using the data stored in the Internet-of-All-Things, non-conscious but highly intelligent algorithms get to know us better than we know ourselves. How this might affect society, politics and daily life is discussed at length. Harari sees Dataism as the scientific paradigm of the 21st century, as everything gets reduced to algorithms and data processing. Over the century everything and everybody could eventually get connected to the Internet-of-All-Things and the data it collects will be used to help us gain immortality, bliss and power. On the other hand, once the thing is up and running we may dissolve in the data flow “like a clump of earth within a gushing river”.

These four writers are contributing to a debate about a future living with automation and machine intelligence that’s been going for a few years now. They are not making predictions, but extrapolating from trends that are already visible. They consider the implications of those trends and develop possible futures. It is mostly very analytical, with each writer laying a logical trail through their ideas to the main points and conclusions. They build their scenarios with care. Like them or not, those scenarios shine a light on an uncertain future and help frame questions and issues that need to be addressed. In all these futures there is an expanding digital world running parallel to the physical world we humans have lived in, and on, so far.

Just where the dividing line between the digital and physical worlds will end up in another hundred years is, I think, the great unanswered question. A wide range of different degrees of integration in the medium-term seems plausible, similar to the digital divide that exists today with access to broadband. Over the longer run an extreme version of the digital economy seems possible, maybe like one run by ems. How long is that? No-one knows, but it’s worth thinking about.

Frase, P. 2016. Four Futures: Life After Capitalism. London: Verso.

Hanson, R. 2016. The Age of Em: Work, Love and Life When Robots Rule the Earth. Oxford: Oxford University Press.

Harari, Y. N. 2016. Homo Deus: A Brief History of Tomorrow. London: Harvill Secker.

Kelly K. 2016. The Inevitable: Understanding the 12 Technological Forces That Will Shape Our Future, New York: Viking.

The holiday post a year after this one is called Rethinking the Deep Past.

No comments:

Post a Comment