N.B. This is a rather long post, motivated by my recent discussions about the art of predicting the future.
Baseball player and everyman philosopher, Yogi Berra, is justifiably famous for his malapropisms (e.g., “Nobody goes there anymore, it’s too crowded.”), some actual, some simply attributed. Now enshrined in the popular vernacular is his amusingly true and philosophically vacuous observation, “It's tough to make predictions, especially about the future.”
Alas, navigating the darkling plain of history is as murky as predicting the pristine future, for variants of Yogi’s bon mot on the uncertain future have been attributed to such diverse figures as physicist and Nobel Laureate, Niels Bohr, and filmmaker, Samuel Goldwyn. I speak from personal experience on uncertain exegesis.
At various times, in op-eds, columns, articles, interviews, debates, and speeches, I have pontificated on the murky future, speculating on possible and likely outcomes in education, innovation, science, medicine, technology, and society. While doing so, I have also been privileged to serve on a diverse set of advisory groups, among them, the U.S. President’s Council of Advisors on Science and Technology (PCAST), the FCC’s Technical Advisory Council, the Department of Energy’s Advanced Scientific Computing Advisory Committee, and the National Science Board (NSB). I have also participated in or led various World Economic Forum and B20 events.
None of these experiences made me an oracle, far from it, but they have given me a sense that, as Morpheus promised in The Matrix, the rabbit hole of the future goes very deep indeed. Have a cookie and free your mind.
From musings on Jevon’s paradox though innovation friction and educational futures to the effects of accelerating change, I have plunged in fearlessly – and some might say foolishly – to reflect on the future, with the uncertain chronicle of my dubious musings documented at http://www.hpcdan.org.
In one memorable exchange, I once debated the European Union’s science adviser, Anne Glover, before EU policymakers. It quickly became clear that she (a biologist) and I (a computer geek) were in complete agreement, though both of us puzzled at the audience’s repeated and insistent questioning about slowing the pace of change in microbiology and computing. Whether concerning digital privacy, genetically modified organisms, cloud computing, or genetic engineering, we emphasized that scientific insights could only be managed wisely, not forestalled.
I remain convinced this was and is the correct answer, but it was not the desired one. It is also why I believe passionately that an educated populace is an essential prerequisite to thoughtful and informed debate on the appropriate use of new technologies. Absent that, we are but on a darkling plain, swept with confused alarms of struggle and flight.
In a less serious but equally thoughtful vein, I interviewed and debated futurist and science fiction writer, Bruce Sterling. To the amusement of our live audience, we sought to establish the major consensus narrative (aka the truth) about future technology and society, as we rambled across post-genomic biology and medicine, nanotechnology and personalized manufacturing, cognitive radio and the Internet of Things, autonomous vehicles and grand AI, and privacy and information flow in the digital age, while discussing the cultural aspects of Clayton Christensen's innovator's dilemma. (See A Conversation on Designing the Future.)
Several times, I have been mostly right (e.g., the criticality of digital privacy) about the future. Other times, I have been right about the technology and wrong about the socioeconomics (e.g., online higher education and MOOCs) or wrong about the technology but right about the socioeconomics. In many cases, I have just been unambiguously and laughably wrong. (Bought any Itanium computers lately?) However, embarrassment has never stopped me from thinking or asking others what the future might become. As writer Neal Stephenson provocatively noted, “… if we want to create a better future, we need to start with better dreams.”
Those waking dreams begin by identifying which aspects of the present are malleable – and the associated costs (economic, political, social, and technical) and timescales (months, years, or decades) – as well as those aspects that are truly immutable. The goal, however, is not just to understand and adapt to the future, but to shape and reify it, for dreams without plans are simply illusions.
Recognize, though, that each envisioned change must be pursued thoughtfully and wisely, for change has both expected and unexpected benefits and consequences. Foresight is always uncertain, as Ursula K. Le Guin so poignantly portrayed in the Lathe of Heaven.
Futurist Imagining
Technical innovation – predicting the qualitative effects of quantitative change – is often easier than social and legal/policy change, though I recognize my geek bias may be evident in so saying. Even within a specific technical domain, the locus of innovation can shift with time and technology. Consider commercial aviation as one salutatory illustration.
Though the cruising speed of commercial aircraft rose throughout the first half of the 20th century, today’s Boeing and Airbus commercial airplanes fly no faster than the vaunted Boeing 707 of the late 1950s. Higher speeds were and are technically possible, as military aircraft and the Concorde demonstrated, but supersonic commercial transport remains economically non-viable. Nevertheless, aircraft innovation continues apace, with more reliable, fuel efficient jet engines, carbon fiber designs, glass cockpits and fly by wire. (See Nothing Lasts Forever.)
For those playing the futurist game at home, consider just a few other disruptive examples, past and present. What if long distance communications were effectively free, rather than costing dollars per minute as in the days of Ma Bell? (Answer: the Internet, the web, and e-commerce). What if computers were small and cheap and could store the digital equivalent of the Library of Congress in the palm of your hand, rather than costing millions of dollars and filling massive rooms? (Answer: laptops, smartphones, and the Internet of Things)
What if the DNA you inherited at birth were just a starting blueprint, rather than a lifelong constant? (Answer: CRISPR and personalized medicine). What if each decade the summers got just a little hotter and the winters got just a little colder? (Answer: climate change, economic loss, and global migration) What if autonomous AI systems were empowered to identify individuals and make life and death decisions? (Answer: stay tuned, this experiment is still underway)
Now slip on your futurist goggles and engage in a brief, though humorous, gedanken experiment. Might the triumvirate of 5G cellular, talking shoes (no, not the ones with flapping soles), and drone pizza delivery be in our near future? (Stop, right now! How dare you laugh at talking shoes; they’re just a lateral from talking thermostats. Where’s your George Jetson ethos?) In a world where a dystopian Max Headroom is a quaint memory, all three depend on advanced computing, deep learning, and edge AI, but each has divergent social and economic drivers and probabilities. (See Come to the Supercomputing Big Data Edge.)
Despite trade wars, deployment costs now measured in the hundreds of billions of dollars, and massive picocell requirements, the early rollout of 5G cellular – or at least its marketing twin – is already underway, driven by revenue-salivating carriers and insatiable consumer demand for wireless connectivity and cat videos. Conversely, marketing talking shoes would be cheap, a dubious fashion statement realizable by any wannabe teenaged designer of chic couture, no more or less likely than any of a dozen other bizarre trends. Finally, pizza delivery by autonomous drone is now technically feasible, and Dominos has already delivered pizza by drone in New Zealand. Drone delivery is currently illegal in the United States, though legislation to satisfy a national craving for pepperoni might be in our future.
Multivariate Uncertainty
Make no mistake, predicting the future is not for the faint of heart, even with a 5G smartphone, a full belly, and talking shoes; the future is not a simple convolution of linearly independent possibilities. Each possibility depends on a host of uncertainties, with different constraints and interdependences – some are technical, others are cultural; some are economic, others are social; and still others are legal or ethical. More tellingly, each of these uncertainties evolves with varying speeds and scales. The future is many things, but it is not an unconstrained n-dimensional random walk; even if it does occasionally feel like a Wiener process; it is far more Bayesian, as any inveterate Let’s Make a Deal knows all too well. (The Monty Hall problem is a classic veridical paradox; choose wisely lest you lose the Big Deal behind door number three.)
A few of my geekier friends might say the future is explained by Everett’s many worlds interpretation of quantum mechanics, where all possible outcomes attain, and we live in but one of many possible universes. Yes, this might mean there is a world where polyester leisure suits and platform shoes are part of a man’s professional wardrobe, disco reigns as a national art form, Inktomi dominates web search and digital advertising, high-speed maglev trains crisscross the U.S. hourly, and adult obesity and diabetes are rare. This is not our world, though I might reluctantly accept disco, but not leisure suits, in exchange for high-speed trains and diabetes’ elimination.
Despair not, there is hope in understanding the Copenhagen interpretation of the wave function’s collapse – when the superposition of eigenstates reduces to a single state consistent with observation. Put another way, observation tells us the status of Schrödinger's cat, the future of flying cars, and talking shoes. Flying cars, though technically possible, are economically and statistically unlikely. However, the jury is still out on talking shoes. (For an anecdote about uncertainty and faculty recruiting, see Schrödinger’s Gershwin.)
Lessons for Futurists
Enough philosophy and raconteuring – herewith are five things I have learned about predicting and shaping the future.
- n the near term, we tend to overestimate change, but in the long term, we underestimate it.
The Gartner hype cycle is real. Shiny objects readily capture our attention, mesmerizing and beguiling us with pregnant possibilities. We see a world of inflated options, infected with go fever, ignoring the challenges posed by a still nascent idea. After the trough of disillusionment, most ideas descend quickly into the valley of death. In a few cases, after innovation and much hard work, the slope of enlightenment appears, with subsequent success. By then, we have forgotten how much change the innovation has wrought, now comfortable in an Orwellian sense that the past is invariant, as we willingly and enthusiastically embrace double think. Yes, we have always had smartphones, and Google has always been the go to source for homework questions.
- Linear extrapolation is easy; exponential extrapolation is extraordinarily difficult.
Tomorrow is generally an incremental variant of today, and small changes rarely have dramatic effects. Yet there are situations where change truly is exponential. The cadence of transistor shrinkage, popularly known as Moore’s law, along with its more subtle technical enabler, Dennard scaling, has birthed exponentially more powerful and inexpensive computing systems over the past thirty years. That cadence is now ending, with profound implications for a future where an exponential has reached its fuzzy terminus. The cost of whole genome sequencing has plunged similarly, with concomitant disruptions in the science and finances of health care. (DNA sequencing at near consumer prices is here, albeit for smaller sequences – look at the nanopore-based SmidgION.) Neither of these exponential changes was guaranteed; rather, they were the product of continued, incremental innovation.
- Quantitative change leads to qualitative change, but this but one type of disruptive change.
Sufficient change in speed, size, capacity, capability, and price almost always lead to qualitative change. If unit prices drop dramatically and volumes rise commensurately, new opportunities and markets arise. Contrarily, continuing design improvements in gas lamps would have never led to the electric light bulb, whose disruptive technical innovation triggered an ecosystem revolution, as did the automobile and the airplane. Anticipating these disruptive innovations is by definition, difficult, but understanding context and tipping points can lessen surprise; these are the trigger points of the hype cycle.
- Disciplinary experts are often blinded by knowledge of certain constraints; outsiders see differently.
If you have ever worked on any new idea or approach, you are intimately aware of its limitations and the associated risks, and occasionally terrified when others embrace it with blithe obliviousness. At one level, this cautionary awareness is a critical damper on unbridled enthusiasm. (After all, few things really are better than sliced bread, though some might opine that canned beer has been a close competitor.) Building an inclusive and thoughtful circle can lessen surprise; non-domain experts, including many science fiction writers, often better anticipate the broader implications and uptake (or rejection) of new ideas.
- The interplay of technology evolution and social dynamics is both complex and subtle.
Predicting social reaction to technical change is challenging. I have personally seen a single technical idea met with unbridled enthusiasm and unqualified horror when applied to slightly different human domains. None of the scientists and technologists, and I count myself among that community, would have predicted that response divergence. Again, diverse teams with disparate experiences and knowledge can lessen surprise and increase predictive accuracy.
Coda
The future is a tabula rasa, albeit with social, economic, political, and technical constraints. We are not simply passive voyagers on an uncertain sea; we can and should shape our future. Doing so requires both contextual awareness and thoughtful shaping of alternatives.
Remember, the best way to predict the future is to create it.
Yogi was right; the futurist mantra is pity and cautionary.
Predict, reify; rinse, repeat.
Comments