A real-time computer-simulated cat brain? Could IBM have come up with a project more likely to trigger Internet excitement?
For the handful of you who missed the news, IBM's Almaden Research Center this week announced that it had produced a "cortical simulation" of the scale and complexity of a cat brain. This simulation ran on one of IBM's "Blue Gene" supercomputers, in this case at the Lawrence Livermore National Laboratory. (As an aside: LLNL is best known as the center for ongoing research into advanced nuclear weapons and related projects; if the lab is now turning its attention to brain simulations, I don't know whether to be happy that it's moving away from weapons or worried that it will try to weaponize AI.)
Worries about the Robopocalypse may be only partially tongue-in-cheek, but it's worth taking a moment to examine what exactly has happened here. This is what the IBM press release says about the simulation:
Scientists, at IBM Research - Almaden, in collaboration with colleagues from Lawrence Berkeley National Lab, have performed the first near real-time cortical simulation of the brain that exceeds the scale of a cat cortex and contains 1 billion spiking neurons and 10 trillion individual learning synapses.
This isn't a simulation of a cat brain; it's a simulation of a brain structure that has the scale and connection complexity of a cat brain. It doesn't include the actual structures of a cat brain, nor its actual connections; the various experiments in the project filled the memory of the cortical simulation with a bunch of data, and let the system create its own signals and connections. Put simply, it's not an artificial (feline) intelligence, but it's a platform upon which an A(F)I could conceivably be built.
Long-time readers may be having a deja vu moment here, and for good reason. The same team responsible for the cat-scale brain sim created a mouse-scale brain sim a few years ago. One of the researchers, Dharmendra Modha, runs a blog on cognitive computing, and has posted a PDF of the research paper on this project. If you want the hard-core science, not just a press release, have fun.
Ultimately, this is a very interesting development, both for the obvious reasons (Artificial Cat Brain!) and because of its associated "Blue Matter" project, which uses supercomputers and magnetic resonance to non-invasively map out brain structures and connections. The cortical sim is intended, in large part, to serve as a test-bed for the maps gleaned by the Blue Matter analysis. The combination could mean taking a reading of a brain and running the shadow mind in a box.
Science fiction writers will have a field day with this, especially if they develop a way to "write" neural connections, and not just read them. Brain back-ups? Shadow minds in a box, used to extract secret knowledge? Hypercats, with brains operating at a thousand times normal speed? The mind reels.
But the reality is that, in many ways, the IBM team has done the easy part, and still has a far greater challenge ahead of them. As I said in response to the mouse sim announcement in 2007, the brain isn't simply a haphazard mass of neural junctions; a functional structure simulation may well prove to be a far greater task than simply getting the neural connection sim working. Don't expect to be able to upload your cat's brain into your Roomba any time soon.
A real-time computer-simulated cat brain? Could IBM have come up with a project more likely to trigger Internet excitement?
For the handful of you who missed the news, IBM's Almaden Research Center announced this week that it had produced a "cortical simulation" of the scale and complexity of a cat brain. This simulation ran on one of IBM's "Blue Gene" supercomputers, in this case at the Lawrence Livermore National Laboratory (LLNL). (An aside: LLNL is best known as the center for ongoing research into advanced nuclear weapons and related projects; if the lab is now turning its attention to brain simulations, I don't know whether to be happy that it's moving away from weapons or worried that it will try to weaponize AI.)
Worries about the Robopocalypse may be only partially tongue-in-cheek, but it's worth taking a moment to examine what exactly has happened here. This is what the IBM press release says about the simulation:
Scientists, at IBM Research - Almaden, in collaboration with colleagues from Lawrence Berkeley National Lab, have performed the first near real-time cortical simulation of the brain that exceeds the scale of a cat cortex and contains 1 billion spiking neurons and 10 trillion individual learning synapses.
This isn't a simulation of a cat brain, it's a simulation of a brain structure that has the scale and connection complexity of a cat brain. It doesn't include the actual structures of a cat brain, nor its actual connections; the various experiments in the project filled the memory of the cortical simulation with a bunch of data, and let the system create its own signals and connections. Put simply, it's not an artificial (feline) intelligence, it's a platform upon which an A(F)I could conceivably be built.
Long-time readers may be having a deja vu moment here, and for good reason. The same team responsible for the cat-scale brain sim created a mouse-scale brain sim a few years ago. One of the researchers, Dharmendra Modha, runs a blog on cognitive computing, and has posted a PDF of the research paper on this project. If you want the hard-core science, not just a press release, have fun.
Ultimately, this is a very interesting development, both for the obvious reasons (an artificial cat brain!) and because of its associated "Blue Matter" project, which uses supercomputers and magnetic resonance to non-invasively map out brain structures and connections. The cortical sim is intended, in large part, to serve as a test-bed for the maps gleaned by the Blue Matter analysis. The combination could mean taking a reading of a brain and running the shadow mind in a box.
Science fiction writers will have a field day with this, especially if they develop a way to "write" neural connections, and not just read them. Brain back-ups? Shadow minds in a box, used to extract secret knowledge? Hypercats, with brains operating at a thousand times normal speed? The mind reels.
But the reality is that in many ways the IBM team has done the easy part, and still has a far greater challenge ahead of them. As I said in response to the mouse sim announcement in 2007, the brain isn't simply a haphazard mass of neural junctions; a functional structure simulation may well prove to be a far greater task than simply getting the neural connection sim working. Don't expect to be able to upload your cat's brain into your Roomba any time soon.
It's a pretty widely-accepted notion that the atmosphere is a ridiculously complex system, and the best we can do with our models is a rough approximation. The more teraflops we throw at the problem, the more granular the results--but even the best models operate at a scale of a hundred or so kilometers; we're still just seeing a shadow of the atmosphere's true complexity.
But what if that's wrong?
The atmospheric complexity idea has a lengthy provenance. Lewis Fry Richardson, the father of numerical analysis of the weather, proposed way back in 1922 that weather could be forecast using difficult math. This insight, and the work that he produced, led directly to the climate and weather models in use today. But Richardson had another insight: perhaps there's a simpler underlying system at work, something involving what would later be called fractal geometry. (He once wrote: "Big whirls have little whirls that feed on their velocity, and little whirls have lesser whirls, and so on to viscosity.") In the 1980s, when we finally had enough computational firepower to test this, the initial results weren't good, and the idea was more-or-less abandoned.
McGill University physicist Shaun Lovejoy kept coming back to the idea, though, and he and his team found suggestive indications that there was a multifractal process at work. (Standard fractal systems involve a single exponent defining the "fractal dimension" of a system; multifractal systems involve a range of exponents, given the label "singularity exponent." Seriously.) The available data weren't clear though, because the readings were muddied by the effects of the very aircraft and instruments used to gather them. So Lovejoy looked up--to satellites. And digging through data from 1,200 consecutive orbits of the Tropical Rainfall Measuring Mission, the team came up with something pretty remarkable: very strong evidence that the atmosphere follows power laws and shows fractal behavior, visible at scales from under 10km to over 20,000km.
Um, okay. Nice, I suppose. But what does that mean?
Put simply, it means that the classic "chaos theory" problem--that small variations and inaccuracies can lead to wildly divergent results, aka "the butterfly effect"--could be set aside, and we'll be able to create accurate models down to... well, here's what New Scientist says:
Now Lovejoy's team is keen to see cascades extend the reach and reliability of current models. While the existing models cannot handle structures much smaller than 100 kilometers across, the cascades may continue down to scales smaller than a millimeter. "Cascades could help fill in that missing factor of 100 million or so," says Lovejoy.
This will have a major impact on climate models--both for improving their accuracy, and (interestingly) for verification. If a given climate model's version of the atmosphere doesn't result in a system that shows power laws and multifractal behavior, then it's definitely inaccurate. Fortunately (or unfortunately, if you hope that climate science has this whole global warming thing wrong), the climate models currently in use do show power law and fractal results.
This is one of those discoveries that will undoubtedly take years to integrate into existing global circulation models, so we're not going to have ultra-accurate climate simulations overnight. But this does give a great deal of impetus to the idea that we can, in fact, generate useful insights into the functioning of global systems using simulations. I'm really curious about how well the multifractal concept could be applied to other ultra-complex systems. Psychohistory, anyone?
It may be odd to focus a political movement on a relatively obscure bit of science, but a world-wide push to limit concentration of atmospheric carbon dioxide to 350 parts-per-million made a big splash last week, with rallies and gatherings all over the planet focusing on drilling this number into the public consciousness. The number comes from work done by (among others) NASA's James Hansen, looking for potential climate "tipping points." 350ppm for CO2 is a safe limit--get too much beyond it, and the dangers multiply.
It's an audacious goal, for reasons of both communication and science.
In terms of communication, while a simple meme like "350" or "350ppm" fits nicely on protest signs and bumper stickers, it's a term without much context for the vast majority of the populace. In and of itself, that's not a problem; however, it can make a visceral connection to the concept more difficult. Activists adopting the 350 meme will need to match rhetoric with education, to make the number meaningful. Again, not impossible, but likely an ongoing challenge.
The scientific audacity with the 350 meme comes from a single, simple fact: current concentration of atmospheric CO2 is roughly 385ppm. That is, we already exceed the 350 limit, and most climate scientists say we'll be hard-pressed to keep from going over 450ppm by the middle of the century. And carbon dioxide takes centuries to cycle out of the atmosphere--even if we stopped all anthropogenic sources of CO2 right this minute, we'd still see too-high concentrations for years to come.
(Even more troubling: even if we stopped all anthropogenic carbon sources immediately, we'd still see continued warming for at least decades, possibly longer, simply from the thermal inertia of the oceans. Absent a radical step, we're guaranteed to see at least another degree or two of warming, no matter what we do.)
If this sounds like I think the 350 movement is a bad idea... I don't. I rather like the simplicity of the meme, and the target is--if difficult--smart. It's not saying "let's keep things from getting too much worse," it's saying "let's make things better." That's the kind of goal I like.
But getting back to 350ppm requires more than a rapid cessation of anthropogenic sources of atmospheric carbon. It requires an acceleration of the processes that cycle atmospheric CO2. Planting trees is an obvious step, but it's slow and actually doesn't do enough alone. We'll also need to bring in more advanced carbon sequestration techniques, such as bio-char. The combination of the two would likely bring down atmospheric carbon levels, given enough time.
Unfortunately, we may not have enough time.
If efforts to eliminate carbon emissions continue to happen at a pace most generously described as "leisurely," we will almost certainly face a situation where we approach and even pass critical tipping point concentrations. Ocean thermal inertia means that climate benefits from emission cessation won't be seen for decades. There's a very real scenario where finally get it right, both cutting out anthropogenic emissions and sequestering megatons of carbon via plants and bio-char ... and still face terrible environmental consequences, simply because we didn't act fast enough.
That's where we start to talk about much more radical, and potentially dangerous, steps. Geoengineering to hold temperatures down is one; to meet the 350ppm goal, we will likely also start looking at large-scale methods to sequester carbon, such as with triggered algae blooms.
350ppm is an audacious goal, but one worth striving for. But its challenge comes not just in the effort to eliminate anthropogenic carbon emissions around the world--a massive endeavor alone--but also in figuring out how to remove the extra carbon already there. I hope that the 350 leaders have thought through the implications of what that means.
In Futures Thinking: The Basics, I offered up an overview of how to engage in a foresight exercise. Today, as the next piece in this occasional series, I'll take a look at the first step in such a process.
"Asking the Question" is the first step in a formal futures thinking project. At first glance, it should be easy--after all, you should know what you're trying to figure out. Unfortunately, while it may be simple to ask a question, asking the right question is much more challenging It's easy to ask questions that are too vague, too narrow, or assume the answer; it's much more difficult to ask a question that can elicit both surprises and useful results.
Remember, the goal of structured futures thinking is to come up with a picture of possible futures that will help to inform strategic decisions. The answers you'll get from a futures exercise are rarely cut-and-dried, but ideally will help you make your decision more thoughtfully. Futures thinking isn't a Magic 8-Ball, a process where all you need to ask is "Should we do X?" (and getting "Ask Again Later" as a result is neither useful nor surprising).
It's a subtle point, but I tend to find it useful to talk about strategic questions in terms of dilemmas, not problems. Problem implies solution--a fix that resolves the question. Dilemmas are more difficult, typically situations where there are no clearly preferable outcomes (or where each likely outcome carries with it some difficult contingent elements). Futures thinking is less useful when trying to come up with a clear single answer to a particular problem, but can be extremely helpful when trying to determine the best response to a dilemma. The difference is that the "best response" may vary depending upon still-unresolved circumstances; futures thinking helps to illuminate possible trigger points for making a decision.
One important point about the difference between problems and dilemmas: with dilemmas, you will generally have a sense of the different possible responses, and have to make an intelligent choice between them. With problems, the solution is almost by definition hidden, and must be discovered. Futures thinking is much more robust when dealing with dilemmas.
That's because what you're doing with a futures exercise is trying to draw out the range of conditions in which your choices play out--the internal and (especially) external factors that will shape outcomes. You can then play your initial strategic choices against the different resulting futures. Bear in mind that this can sometimes have surprising results. Although a futures exercise asking "where should I build my widget factory?" may be too broad, the more narrow exercise of "should I build my widget factory in China or India?" may well lead to a determination that neither really provides the desired results.
As I noted in Futures Thinking: The Basics, another aspect of asking the question is figuring out the time frame for the exercise. This typically comes down to two key issues: how long will it take to implement the plan you're pondering; and how long into the implementation do you want to test the results. If it takes three years to set up the widget factory, a five year target for the future exercise would be useful to think through initial operating environment, while a 12 year exercise will help to think about what things will be like over time.
One trick that I find useful in determining the target date is to think about political cycles in the operating environment. For something that looks primarily at the United States, for example, eight years out works well because it guarantees that whoever has the presidency at the moment will be out of power, so narratives about the current political leadership (pro or con) have less salience. A similar process works nicely for most regions with relatively predictable political cycles.
Another technique that can be useful is to think about two or three key drivers that you're already familiar with, and look for any upcoming inflection points in their ongoing evolutions. If you're looking at mobile technologies, you might target a point after planned roll-outs of 4G wireless networks. Regulations coming into effect, changes in resource availability or pricing, and demographic shifts can all play similar roles.
One last note: even if you're satisfied with the question you're going to examine through a futures exercise, don't be afraid to reconsider it as the exercise progresses. "Am I asking the right question?" should come up repeatedly over the course of the process. You may end up discovering that a better question is out there, or that you're on the right track. Either way, the practice of reexamining your own assumptions and expectations will inevitably prove a useful part of the process.
My talk last weekend at the New York Future Salon explored the likelihood and the implications of the transformative event known as the "Singularity." I tend to part ways with many Singularity enthusiasts over two small issues: what comes before a Singularity, and what comes after.
In terms of what comes before, I'm generally in the camp that
machine-substrate intelligence is very likely possible, but is probably
a much more complex problem than some of the more enthusiastic
Singularitarians would have us think. We currently have a single model
of a mind emerging from a physical structure--the human brain--and
(as noted by one of the 2009 Singularity Summit speakers, David Chalmers)
we're not even sure how that happens. Add to that issues around
learning, around complexity, around the very definition of
intelligence, and you have the potential for a situation where--even
if there are no physical laws preventing the emergence of artificial general intelligence
-- "real" AI remains the computer science version of nuclear fusion:
perpetually just a couple of decades away (with plenty of dead-ends and
showy hoaxes along the way).
I've noted elsewhere that I suspect that "a stand-alone artificial mind will be more a tool of narrow utility than something especially apocalyptic." Part of the reason is the difficulty, but another part is the near-certainty that the technologies of human intelligence augmentation will continue apace. The technologies that may be dead-ends for efforts to construct a self-aware artificial mind could easily be of great value as non-conscious assistants to human minds.
The notion that creating "real" AI may turn out to be extraordinarily difficult, and the idea that human intelligence augmentation could itself turn out to be a more promising line of research doesn't get a lot of push-back from the more thoughtful Singularity proponents I've encountered. After all, both have been demonstrably true so far. A tougher sticking point, however, comes when I explore what could come afterwards.
If greater-than-human artificial intelligence emerges out of aggressively competitive projects, each seeking to be first, and is put to use without much thought to what might happen next, then the traditional Singularity scenario seems pretty likely. But that's not the only one:
The upper-left scenario, "Out of Control," is the more-or-less conventional Singularity story. AI gets smart, gets loose, and does as it will. Could be hell, could be heaven, but pretty much is out of our hands. In short, this is the scenario in which AIs eliminate our civilization.
Upper-right, "Taxes and Allies," is a world where competitive projects lead to real AI, but they're undertaken with a greater awareness of implications and impacts. AIs in this world are held onto as business tools--may in fact be corporations (as in Charlie Stross' novel Accelerando)--but remain embedded in human civilization. This could, by the way, be one of the pathways to a "robonomics" economy. This is the scenario in which AIs become part of our civilization.
Lower-left is "Eat Your Vegetables*." In this scenario, the greater-than-human AIs emerge not from tools of competition, but tools of collaboration--imagine, for example, AIs emerging out of software intended to help humanity manage climate disruption. The result here is a world where these systems are less about artificial intelligence and more about artificial wisdom: assisting us in doing the right things for ourselves and our planet. (Listen to the audio scenario "The Chorus" for an example of this kind of world.) This is the scenario in which AIs take care of our civilization.
Finally, "Djinni in a Bottle" offers a scenario where AIs come from collaborative tools, but in a context of management of impacts and implications. As in the classic tales of djinnis, this is a world in which beneficial and detrimental results occur based on just how wisely we use our power. This is the scenario in which AIs empower the best and worst of our civilization.
Three of the four scenarios (leaving aside "Out of Control") assume that human social intelligence, augmentation technology, and competition continue to develop. And in all three, human civilization--with its resulting conflicts and mistakes, communities and arts, and, yes, politics--remains a vital force even after a Singularity has begun.
One key aspect of the three is that they're not necessarily end states. Each could, given the right drivers, eventually evolve into one of the others. Moreover, all three could in principle exist side-by-side.
I noted earlier that I differ from many of the Singularity enthusiasts in my take on what happens before and what happens after a Singularity. I suppose I differ in my take on what happens during one, as well. I don't think that a Singularity would be visible to those going through one. Even the most disruptive changes are not universally or immediately distributed, and late-followers learn from the reactions and dilemmas of those who had initially encountered the disruptive change.
Perhaps the most notable aspect of a Singularity is that, ultimately, it's only clearly visible in the distance: off in the future, where it remains a mysterious veil; and in the past, after we've moved along far enough to see just how differently we're living our lives.
(When video from my latest talk, a lively over-capacity event, is online, I'll post a link at my home Web site, Open the Future. My talk slides are available via Slideshare.)
*For those of you unfamiliar with this bit of American idiom, it refers to being told (often forced) to do something that's actually good for us, but not necessarily pleasant.
The 2009 Singularity Summit is coming up this weekend (October 3-4) in New York City. If previous years are any guide, the next week or so will be filled with breathless articles on tech sites and magazines about the Coming Artificial Intelligence (AI) Revolution, seemingly random references to Kurzweil on Twitter, and oddly edited radio interviews with conference attendees, highlighting fringe ideas at the expense of serious conversation. I spoke at the 2007 Singularity Summit, so this is all quite familiar to me.
If you're not quite sure what this Singularity thing is all about, you're not alone. The use of the term "Singularity" comes from physics, where it describes a point of collapsed space-time, typically at the center of a black hole; the underlying claim is that all of our knowledge of how the universe works is irrelevant within a Singularity. This is the root of the metaphorical use of the term--after a Singularity event, everything we know will change in ways we can't now understand. In the early 1990s, Mathematician and science fiction writer Vernor Vinge was the first to clearly articulate this usage of the term, and begins his essay as follows:
Within thirty years, we will have the technological means to create superhuman intelligence. Shortly after, the human era will be ended.
By "superhuman intelligence," Vinge meant one of four different outcomes: We make machines that are intelligent, our computer networks "wake up" as an intelligent entity, human-computer interfaces become so intimate and powerful that the combination forms a superhuman intelligence, or we bioengineer ourselves with smarter brains. In this model, once we have an example of a greater-than-human intelligence, that in turn gives us the means to make even greater intelligence, and so forth. And because the so-much-greater-than-human intelligences will be able to figure out how to do spectacular things, the world that they will make (and remake) will very rapidly become something utterly unrecognizable to mere human minds.
The more common present-day usage of the term "Singularity" actually has a few different definitions, depending upon who you're talking to and how "serious" they are about the subject.
At the broadest end, the Singularity refers to a point in the future where technologically driven changes have hit so hard and so fast that people on the near side of the Singularity wouldn't be able to understand the lives of people living on the far side of one; the lives of the post-Singularity citizens simply wouldn't make sense to pre-Singularity folks. People who adopt this perspective tend to weave all sorts of future-y technological things into it, from radical longevity to personal robots to geoengineering, but always with the underlying point that these things are really disruptive to our lives. This is the usage that I am personally most comfortable with.
In its narrowest form, conversely, "Singularity" refers exclusively to the process described by Vinge, the creation of greater-than-human intelligence, which is then able to make itself even smarter, and so on in something that is occasionally called an "intelligence explosion"; whether or not the lives of post-Singularity citizens would make sense is an irrelevant question--the ultra-intelligent entities would be so much smarter and more powerful that human beings are little more than ants in comparison. People who adopt this perspective often get annoyed at the first group for talking about things that aren't related to intelligence, and tend to see the Singularity as something that could easily lead to the End of Everything. Many of the people associated with the Singularity Summit take this approach.
Somewhere in the middle are those who take the intelligence explosion concept, and use that as the engine for all sorts of ultra-tech fun: brain uploads, "computronium," endless digital lives lived in virtual worlds, and the like. These folks, as opposed to the former group, tend to see the Singularity as something generally desirable. They will, of course, acknowledge the potential for Bad Things to happen, but that's not the thrust of their arguments. The Singularity as presented in Ray Kurzweil's books falls into this category.
Despite the presence of the Singularity concept within various (largely online) sub-cultures, it remains on the edges of common discussion. That's hardly a surprise; the Singularity concept doesn't sit well with most people's visions of what tomorrow will hold (it's the classic "the future is weirder than I expect" scenario). Moreover, many of the loudest voices discussing the topic do so in a manner that's uncomfortably messianic. Assertions of certainty, claims of inevitability, and the dismissal of the notion that humankind has any choice in the matter--all for something that cannot be proven, and is built upon a nest of assumption--do tend to drive away people who might otherwise find the idea intriguing.
And that's a problem, as the core of the Singularity argument is actually pretty interesting, and worth thinking about. Increasing functional intelligence--whether through smarter machines or smarter people--will almost certainly disrupt how we live in pretty substantial ways, for better and for worse. And there have been periods in our history where the combination of technological change and social change has resulted in quite radical shifts in how we live our lives--so radical that the expectations, norms, and behaviors of pre-transformation societies soon become out of place in the post-transformation world.
Two examples of this kind of radical shift from our history are urbanization--the development, thousands of years ago, of large-scale, permanent cities--and the development of the printing press. Both of these technologies (and yes, urbanization is a technological development) increased the power of those who adopted them substantially over those who abstained; and both resulted in a reshaping of politics, economics, and the social order. Of course, they were both considerably slower than present-day visions of a Singularity describe. That's a function of how fast innovations--and the power shifts they produce--propagate.
Were they "singularities?" Not by the strict "intelligence explosion" definition (although a case can be made for the printing press as being a slow-motion Singularity in that regard). They definitely fit better with the broader "things get really weird" definition. But they also underscore an issue that's worth understanding in a Singularity, however defined. They weren't exclusively technological; they were technosocial. How they came about, and how they evolved, depended as much on social, cultural, and political forces as on technological capacities.
Yet few of the discussions about the Singularity -- pro or con -- move beyond the technology. Can machines think? Will IA (intelligence augmentation) beat AI (artificial intelligence)? How many teraflops does a brain run? There's too little discussion of how the social, cultural and political choices we make would shape the onset or even the possibility of a Singularity.
I hope to change that. On October 3, I'll be giving a talk entitled "If I Can't Dance, I Don't Want to be Part of Your Singularity"* for the New York Futures Salon, at 7pm. The talk is open to the public--if you're in the area, please come on by.
Next week, I'll write a bit here about the ensuing conversation.
* The title refers to the line attributed to Emma Goldman about the socialist movements of her era (the early 20th century): "If I can't dance, I don't want to be part of your revolution."
The first in an occasional series about the tools and methods for thinking about the future in a structured, useful way.
For nearly the past fifteen years, I've been working as a futurist. My job has been to provide people with insights into emerging trends and issues, to allow them to do their jobs better. I've done this work for big companies and government agencies (usually under the Very Professional sounding title of strategic foresight), and for TV writers and game companies. It's quite an enjoyable job, as it allows me to indulge my easily-distracted curiosity about the world.
Fortunately, it's also a job with some definite practical uses. Futurism as it's practiced today doesn't try to predict the future, but rather to illuminate unexpected implications of present-day issues; the emphasis isn't on what will happen, but on what could happen, given various observed drivers. It's a way of getting new perspectives and context for present-day decisions, as well as for dealing with the dilemma at the heart of all strategic thinking: the future can't be predicted, yet we have to make choices based on what is to come.
There's a bigger picture at work, too -- less practical, perhaps, but just as meaningful. A few months ago, I wrote here about the importance of futures thinking, and how I have come to view futures work:
I've sometimes called futures thinking a "wind-tunnel," a way of testing plans and ideas. Now I think that's a bit limited. Futures thinking is perhaps better understood as an immune system for our civilization. By examining and testing different possible outcomes--potential threats, emerging ideas, exciting opportunities--we strengthen our collective capacity to deal with what really does transpire. Thinking about the future, and doing so in a careful, structured, open and collaborative way, makes us a stronger civilization. Focusing only the challenges of the present may seem imperative, especially when those challenges are massive and frightening. But without a sense of what's next, a capacity for understanding connections and horizons, and a vision of what kind of world we want, our efforts to deal with today's problems will inevitably leave us weakened, vulnerable, and blind to challenges to come.
But if futures thinking is so important, why don't more people do it?
One answer is that they don't know how. So let's change that. This post, and subsequent posts in this series, will help to explain basic methods of structured futures thinking. You won't become a professional futurist overnight, but you will start to ask new questions about the world, and start to see bigger implications of events and choices.
Futures Thinking - A Process Overview
Most futures projects, whether informal or professional, follow a similar pattern: Asking the Question; Scanning the World; Mapping the Possibilities; and Asking the Next Question.
Asking the Question: Futures thinking is rarely just a free-form "what will the future look like?"--in nearly every case, the exploration into different possible futures comes from a narrow concern. Sometimes that concern is about strategic choices a company needs to make, sometimes it's about potential changes to an operating environment, and sometimes it's about gaining a better understanding of emerging markets, competitors, and/or stakeholders. Even if you're doing this for your own amusement or education, it's helpful to have a basic question in mind, simply as a framework for what follows.
Another aspect of "asking the question" is determining how far ahead you want to think. In my experience, ten years is a good target--it's far enough out that some big changes are likely, but near enough that much of the world will still be familiar. Five years is about as near-term as I like to work--sooner than that, and it becomes too tempting to over-estimate the potential for change; conversely, fifteen years is about as far out as I like to look for most projects, because beyond that the potential for disruptive change is big enough that your basic question may become irrelevant.
Scanning the World: After you have your basic question and timeline in mind, it's time to start thinking about the kinds of factors--sometimes called "drivers"--that are likely to shape how your question would be answered. This is a chance to think about a wide spectrum of issues. If you're looking at the future of mobile computing, for example, you need to take into account not just digital technologies, but also changes to mobility, to transportation, to demographics, to work patterns, to regulation, and to the wide array of new uses for mobile tools (health care, for example, or comparative shopping).
It's useful here to start gathering information. Most of us who work as professional futurists never really stop gathering information--you never know when a provocative, potentially disruptive new development might appear. News sources particular to the issue you're working on are of course useful, but I tend to rely heavily upon journals and Web sites that cover a wide array of subjects (such as New Scientist, Worldchanging, and Fast Company). Web sites and organizations devoted explicitly to thinking about future possibilities can also be of great value.
It can also be useful to bring in perspectives and viewpoints beyond your own. Friends and colleagues can offer new perspectives, and most good futures work is done collaboratively. You should also consider checking out sources that you disagree with, because you can sometimes still find useful insights--and the very act of thinking through why you disagree with their perspective can trigger new ideas, too.
Mapping the Possibilities: Now that you've done the ground-work, here's the heart of the process. Books can be (and have been) written about the various professional methodologies out there for doing futures thinking, but they all boil down to the same core idea: there is no one future. Trying to figure out "the" future is always a mistake; it's much more productive to think about an array of possible outcomes. Remember that the futures you come up with will almost certainly be wrong--the goal is to be wrong in a way that offers insights into present choices.
One technique that's good to start with is to use what some professionals call "futures archetypes"--generic headlines that offer platforms upon which to build more specific stories. Four that can be very easy to use are expectations:
The future is what I expect.
The future is better than I expect.
The future is worse than I expect.
The future is weirder than I expect.
The first three are fairly self-explanatory, but the last may be a surprise. The goal with the fourth archetype is to explore possibilities that completely shake things up (a big earthquake, perhaps, or a war, or a revolution in computing power). This doesn't mean fantasy--alien invasions and robot uprisings are probably best left to the movies--but it does mean something outside of your expectations. The phrase I love to use for this is "plausibly surreal."
Now, write up a short essay about each of these futures, combining the drivers and ideas you came across in your scanning; it's often useful to write a page or two from the perspective of someone living in the year you picked as your future date. The essay need not talk about the subject of your question directly--this is more of a chance to describe the world in which that answer has emerged. This doesn't have to be a science fiction story, by the way. It's better to think of it as a newspaper article, or even a blog entry.
Asking the Next Question: Now that you have four different futures, it's time to return to your original question. Ask yourself how you would answer that question in each of these four worlds. Is there a way to answer that question that can lead to happy (or at least acceptable) results in each, even in the "worse" or "weird" futures? If not, is there a way to minimize the risks in those unacceptable scenarios?
Now ask the next question: what happens then? You've made your choice in each of these futures, but that's not the end of the story. What kind of pressures are there on the success of your decision? What do your friends, colleagues, and competitors say about it?
Thinking it Through: Finally, ask yourself how you get from today to the futures you've laid out. What kinds of choices, what kinds of changes do you need to make now to lead to the outcomes you'd prefer? What can you do to avoid the futures you don't want to see? Often one of the key insights from many futures projects is the simple realization that the future is in our hands--that our choices matter.
But this isn't an easy task. Futures thinking is hard work. Fortunately, you do get better at it with practice. It's worth the effort.
I've been doing quite a bit of work on the impacts of the emerging tools allowing us to manipulate our perceptions of the world (e.g., neurotechnology), our physical environment (e.g., geoengineering), and the building blocks of the material world itself (e.g., synthetic biology and molecular nanotechnology). There's a theme that recurs across all of these arenas: what happens when someone does something careless or malicious with the technology? It's bad enough when the technology in question is an automobile or computer network. These emerging disciplines fall into a category I sometimes call "catalytic innovations," and one characteristic is that the worst-case misuse scenarios can be truly terrifying.
For some, the knee-jerk response is a desire to prohibit the development of these technologies. As appealing as that might sound, it suffers from a fundamental flaw: these technologies do not require a massive industrial base, so surreptitious development would be far harder to detect than (say) nuclear weapons development. Ultimately, the only way to enforce the ban would likely be with constant, unrelenting, global surveillance. Few of us, even those afraid of the potential of these catalytic technologies, would be willing to take that path.
A more nuanced response, and one that I see frequently from the proponents of these various technologies, is that well-designed systems could make catastrophic misuse difficult, even impossible. A synthetic biology lab-in-a-box, for example, might be pre-programmed with a variety of forbidden combinations of bio-components, perhaps with limiting and tracking components built into every synthbio design. A molecular nanofactory could have similar restrictions. Whatever the system, if there's a programming interface, there's the potential for automatic limits on output.
This is a manifestation of a philosophy I see quite often online across a wide array of subjects, that of "tools, not rules"--don't try to get people to change their behavior, alter systems to shape the results of their behavior.
There's certainly a great deal of sense to this notion. Accidents can happen, so tools that stop people from doing something destructive can be of enormous value, even when they're as simple as a dialogue box popping up saying "Are you sure you really want to delete your life's work?"
But this model does little to prevent misbehavior arising from novel approaches, nor from abuses that fall within the system rules, but are still harmful. Some of us might understand the latter as "griefing"--taking advantage of system functions to harm the play of other players. We might also understand it as what happens when people follow the letter of the law, but not the "spirit of the law." This kind of behavior is certainly bad when it comes to elements of society such as finance, but when combined with techno-social developments that quite literally have the power to reshape the planet, this kind of behavior is potentially deadly.
If the tools are insufficient, then, we're left with rules.
The best kind of rules are those we apply to ourselves, those we believe in. Ethics--sometimes thought of as "how you behave when no-one is looking"--have the advantage of being readily applied to novel situations, and able to guide responses fitting the spirit of the law. People in positions of social power (such as doctors and lawyers) often receive training in ethics as part of their educations. What I'd like to see is the introduction of ethics training in these new catalytic disciplines.
Computer programmers, biotechnologists, environmental scientists, neuroscientists, nanotech engineers--all of these fields, and more, should have at least a course in ethics as part of their degree requirements. Ideally, it should be a recurring element in every class, so that it's not seen as just another hoop to jump through (check off the "is this ethical? Y/N" box), but instead as a consideration woven into every professional decision.
This is one reason, by the way, that I was so frustrated with the proposed curriculum for the "Singularity University"--the study of ethics was shoehorned in with policy and law in a way that appeared to be something of an "oh, by the way" add-on.
But my larger point is this: as tempting as it is to rely on well-structured tools to prevent disastrous outcomes, even the best tools are ultimately insufficient. Good interfaces need to be accompanied by strong ethics. It's not just a matter of right and wrong; increasingly, it's a matter of survival.
As our various electronic devices gain more and more sensory awareness, we open up the potential for entirely new forms of interaction.
As our various electronic devices gain more and more sensory awareness, we open up the potential for entirely new forms of interaction. Not just new interfaces--tapping and shaking and whatnot--but a shift in presence. With few exceptions, we use these new technologies in rather familiar ways. We might speak instead of type, or tap instead of click, or wave a control wand instead of mash a control pad, but these are essentially the same kinds of direct input processes we've done for years, just dressed up in a new look.
The real shift comes when we move away from direct interaction and input, towards a world of ambient interaction and awareness.
Our laptops, mobile phones, and sometimes desktop computers increasingly come with built-in microphones, cameras, accelerometers, and even GPS. For the most part, these sensory technologies only come into play when we call upon them directly by launching a related application (to take a picture, or find something on a map, etc.). The rest of the time, these senses are turned off. Battery life probably plays a role in keeping the senses off, but I suspect a bigger reason is that we're simply not accustomed to thinking about our tools as always "paying attention."
Sudden Motion Sensor," but Lenovo, Acer, and HP all have similar systems. But think about this in the abstract--in the laptop I'm typing on right now, there's an environmental sensor paying constant attention, ready to act if certain conditions are met. Now, imagine that same concept holding true for other kinds of sensors.
Imagine a desktop with a camera that knows to shut down the screen and eventually go to sleep when you walk away (but stays awake when you're sitting there reading something or thinking), and will wake up when you sit down in front of it (no mouse-jiggling required).
Or a system with a microphone that listens for the combination of a phone ringing (sudden loud noise) followed by a nearby voice saying "hello" (or similar greeting), and will mute the system automatically.
Perhaps a "sudden motion sensor" for phones, not to detect when the phone is dropped, but to detect when the phone has too-quickly gone from freeway speed to zero (perhaps with the microphone picking up collision noises, or sounds of distress), and auto-dialing a 911-like service.
These are just a few simple examples, relying on some fairly basic rules. But imagine if you combine the sensory awareness with a more complex Bayesian-style learning system. What if your digital device could learn your habits, and adjust accordingly?
Imagine a phone that pays attention to what kinds of lighting and noise conditions typically cause the user to turn off the ringer (or perhaps turn it up), in order to eventually do so automatically.
Or a mobile device that could keep track of the user's location, changing settings (network, mail servers, desktop image, even available applications) automatically.
What prompted this line of thought for me was the story about the Outbreaks Near Me application for the iPhone. It struck me that a system that provided near-real-time weather, pollution, pollen, and flu (etc.) information based on watching where you are--and learning where you typically go, to give you early warnings--was well within our capabilities.
Or a system that listened for coughing--how many different voices, how often, how intense, where--to add to health maps used by epidemiologists (and other mobile apps).
And, of course, there are the misuses and abuses, whether by malicious hackers (listening for social security codes and credit card numbers) or by government agencies.
Most of these are technically possible today, although they would probably be too much of a drain on the batteries of smaller devices. Nonetheless, the question isn't "can this happen?," it's "will we want it?" Are you ready for your phone, your laptop, your digital environment to be paying attention to everything you do?
Remember, the goal of structured futures thinking is to come up with a picture of possible futures that will help to inform strategic decisions. The answers you'll get from a futures exercise are rarely cut-and-dried, but ideally will help you make your decision more thoughtfully. Futures thinking isn't a Magic 8-Ball, a process where all you need to ask is "Should we do X?" (and getting "Ask Again Later" as a result is neither useful nor surprising). 
The 2009 
Two examples of this kind of radical shift from our history are urbanization--the development, thousands of years ago, of large-scale, permanent cities--and the development of the printing press. Both of these technologies (and yes, urbanization is a technological development) increased the power of those who adopted them substantially over those who abstained; and both resulted in a reshaping of politics, economics, and the social order. Of course, they were both considerably slower than present-day visions of a Singularity describe. That's a function of how fast innovations--and the power shifts they produce--propagate.
The first in an occasional series about the tools and methods for thinking about the future in a structured, useful way.
Scanning the World: After you have your basic question and timeline in mind, it's time to start thinking about the kinds of factors--sometimes called "drivers"--that are likely to shape how your question would be answered. This is a chance to think about a wide spectrum of issues. If you're looking at the future of mobile computing, for example, you need to take into account not just digital technologies, but also changes to mobility, to transportation, to demographics, to work patterns, to regulation, and to the wide array of new uses for mobile tools (health care, for example, or comparative shopping).
For some, the knee-jerk response is a desire to prohibit the development of these technologies. As appealing as that might sound, it suffers from a fundamental flaw: these technologies do not require a massive industrial base, so surreptitious development would be far harder to detect than (say) nuclear weapons development. Ultimately, the only way to enforce the ban would likely be with constant, unrelenting, global surveillance. Few of us, even those afraid of the potential of these catalytic technologies, would be willing to take that path.
But this model does little to prevent misbehavior arising from novel approaches, nor from abuses that fall within the system rules, but are still harmful. Some of us might understand the latter as "griefing"--taking advantage of system functions to harm the play of other players. We might also understand it as what happens when people follow the letter of the law, but not the "
As our various electronic devices gain more and more sensory awareness, we open up the potential for entirely new forms of interaction. Not just new interfaces--tapping and shaking and whatnot--but a shift in presence. With few exceptions, we use these new technologies in rather familiar ways. We might speak instead of type, or tap instead of click, or wave a control wand instead of mash a control pad, but these are essentially the same kinds of direct input processes we've done for years, just dressed up in a new look.
What prompted this line of thought for me was the story about the 
