I really liked this article because it was a concise demonstration of the 'known' solution problem. A while ago[1] there was a discussion of what folks from the 19th century thought the world would be like in the 20th century. The common theme is that the futurists could never anticipate a change to a technology they didn't have any experience with. The rise of communications as a means of linking people together was completely missed, even by the Dick Tracy types who had wristwatch phones. Why? Because not only did it communicate like everyone had experienced (telegraph, telephone) it <i>persisted</i> with bulletin boards, MUDS, Facebook, and Twitter.<p>That is why I always try to ask a question my Dad suggested I ask when looking a future predictions, "What is so common place that they can't see it changing?" That is so difficult to do. Things like "What would the future be like if there literally was no disease and any physical injury could be patched up quickly?" Or "Energy is suddenly free?" or "Lie detection becomes infallible?" etc. Very tough to do.<p>[1] <a href="https://news.ycombinator.com/item?id=13101643" rel="nofollow">https://news.ycombinator.com/item?id=13101643</a>
Two problems in the article analysis<p>One is the classic hard vs soft sci fi. A classic example in book form is Stranger in a Strange Land reads like a story about 60s boomer hippies from India who meditate and get high and die in an allegory of Jesus, because it is. There's a very thin wrap of search and replace science fiction over that hippie story most famously a giant plot hole where the author wrote in video telephone video conferencing without following up on the hole in the plot generated by being able to see the other party. In comparison, hard sci fi takes into account and weaves a story about the human effects of technology...<p>The other problem is in the "The authors polled a range of experts". Asking my plumber about theology or my lawyer about physics is obviously dumb, and in pop culture we accept that asking a 50s dude about the 70s is pretty much a waste of time. However in all fields of human activity if you want an intelligent commentary on contemporary computer programming you have to either hatch or grow a contemporary computer programmer. The opinions of a dude in another field from decades ago are only accidentally going to be correct. Regardless who was an expert in automation in the 60s, I'm pretty sure they weren't being asked because they almost certainly didn't have cool authoritarian credentials because the topic wasn't cool enough at that time, so by definition they were talking to the wrong people. The history of many present day trends and concepts shows they were not very cool before they became cool, therefore asking contemporary experts on the topic of cool will mostly not work.
This is the premise of 'disrupt', isn't it? That the existing business models or technological solutions are so thoroughly shaken by the appearance of something new that doesn't pass for a mere iteration of the same. Electricity was not directly comparable to steam, despite having some of the same uses at first, cars were quite different from horses, and the packet-switched internet behaves differently from circuit-switched telephone lines.<p>Part of this phenomenon of even experts unable to predict the new is that often our current technology is capped by our societal understanding of physics. Before the Wrights and other early pioneers, we knew that powered heavier-than-air flying was possible for light things like birds, but weren't sure whether the same applied to large, heavy frames, or how we'd be able to deliver enough power to the machine while keeping the weight down. But after a couple of early success, a vibrant field of haphazard experimentation and serious scientific research opened up, and we learned a lot about aerodynamic lift.<p>Before the transistor, it was difficult to seriously posit that electricity-powered, non-mechanical computers would be ubiquitous not just to the point of every government and large enterprise having them, but also every single person possessing multiple miniaturized versions on their person. The invention of integrated circuits was a transformational achievement: it enabled us to go from having very scarce computing capacity to having a slight excess, and so new programs emerged to harness that ability to do calculation. VR and AR in turn aren't that transformative yet: they're simply video games where a portion of the input comes from the 'real world'.<p>Today, we're at a UI/UX/AI/ML crisis, where we have immense computing resources at our disposal and still lack an effective ability to communicate our intent with the machine. We dip down to the ancient metaphor of manipulating a pre-set UI with a pointer like a mouse or our fingers or our eyes, or we have to speak audibly to those within earshot so a microphone can capture our command, ascertain some meaning, then map it to one of many predefined actions. These seem like the dark ages. It's not hard to guess that something revolutionary is going to happen in this arena, and will require actual scientific discoveries to make it possible.
Interesting aside: One of the RAND's predictions (No. 24) is for "<i>International agreements which guarantee certain economic minima to the world's population as a result of high production from automation</i>", which sounds pretty much like Basic Income, along with some of the same arguments we've heard recently about why it will be increasingly needed. (Not saying I agree nor disagree)
One way to detect suspect predictions is to look at which ones suggest improvements to things we have today without changing them significantly.<p>For example, people envision self-driving cars that are exactly like the cars we have today, only automatic. They predict that they will still be big, expensive, complicated, multi-person vehicles. I'm sure that category will continue exist, but why will that remain the norm? The size and complexity of cars are major drawbacks (e.g. storage, traffic, cost). With bicycles, we can see that small individual vehicles have significant advantages due to their size, cost, and simplicity. With self-driving cars, I think there will be room to borrow some of those advantages. Today, we have to build cars with safety as a major concern, but when self-driving vehicles are common, we will not have to. If we are free to relax during a car journey, the ability to travel at high speed may not be such a major concern, either. The typical self-driving car could easily be something inexpensive, small, and simple like the PodRide instead: <a href="https://www.youtube.com/watch?v=4lKq1fGtXFM" rel="nofollow">https://www.youtube.com/watch?v=4lKq1fGtXFM</a>
When I see things like the RAND 'long-range forecasting' study, I always look to see if they specify ubiquity -- it's very different to assert that we'll have ultrasonic implants for the blind in a laboratory setting vs. being commonplace and woven into the fabric of society.<p>As William Gibson is fond of saying, "the future is already here — it's just not very evenly distributed." Predicting what will become integral to our daily life is much more interesting/difficult than determining how long it would take to develop a specific technology's proof-of-concept.
Thanks for this piece, it made me recall the joy of spending my afternoons in the library and the vivid dreams I got from science fiction books - I remember reading Verne's "20,000 Leagues under the Sea" as a kid in the 70's and picturing the Nautilus as a nuclear submarine in Victorian clothes.<p>Back then the cold war was rampant and Brazil was under military dictatorship. Now I'm living in the future, and it is half familiar and half surprising.
Applied AI will be the next "nobody knew how fundamentally things would change."<p>I don't believe in the singularity (at least as currently envisioned) but I do believe we need to figure out what a society must look like when most human beings are, almost literally, not useful for commercial employment.<p>Like, the whole tenets of a market economy and capitalist ownership will smash into the tenets of human worth and the function of society, with violence and consequences much greater than we've seen until now.<p>We can see it coming. It's obviously clear. And nobody in charge is taking about it, much less doing anything about it. A totally avoidable tragedy, walked into on purpose.
For anyone interested: There's an enjoyable radio adaptation of 'A Logic Named Joe' from X Minus One in 1955 <a href="https://archive.org/details/XMinusOne55122831ALogicNamedJoe" rel="nofollow">https://archive.org/details/XMinusOne55122831ALogicNamedJoe</a>
This is a bit banal. Nobody can predict second and third order effects reliably because they don't know how to weight one predicted effect vs another to know how they'll intersect and interact.<p>Extrapolation for any one technology usually works well for a short distance into the future (a decade or so), but the chances of a technology being blindsided by something else entirely rapidly rise over time.
I will relate to my observation in Harry Potter books. Harry Potter universe has time machine which we know scientifically are impossible while in the same HP universe, Hermoine has to physically go to the forbidden library to read books. JK Rowling couldn't think of wikipedia or internet which both exist in real world but not time travel.
The philosopher Alasdair MacIntyre writes about why the future is inherently unpredictable in his book <i>After Virtue</i>. It's in the context of social sciences but it applies just as well here. There's a decent summary here[0]<p>[0]: <a href="http://www.thenewatlantis.com/doclib/20120203_aftervirtuechaptersummary.pdf" rel="nofollow">http://www.thenewatlantis.com/doclib/20120203_aftervirtuecha...</a><p>> MacIntyre argues that "there are four sources of systematic unpredictability in human affairs" which preclude social science from being like natural science (93). The first is radical conceptual innovation, which can be explained in retrospect, but inherently can only be predicted when the innovation has already occurred. MacIntyre notes that this also means that the future of scientific innovation cannot be predicted, invoking the Church-Turing thesis as further proof. The second source is the fact that "the unpredictability of certain of his own future actions by each agent individually" implies the unpredictability of that agent by any other agent, and hence an aggregate unpredictability to the social world (95). The third source "arises from the game-theoretic character of social life" (97). Social life in fact embodies multiple games, players, and transactions and thus cannot be studied as a single instance, reducing the predictive power of game theory. The fourth source is "pure contingency", the way in which "trivial contingencies can powerfully influence the outcome of great events", such as the length of Cleopatra's nose, or Napoleon's cold at Waterloo (99-0).
I have to say I find this piece pretty disappointing. Slightly smug throughout, which is easy in retrospect, with approximately zero to add to the conversation. The thesis seems to be: The future is hard to predict, and even if you can, it's hard to get the timeline right. No shit. Is this what VCs do with their spare time? Remember granddad and contemplate uncertainty?
One of my favorite examples of SF writers underestimating the pace of certain changes is a scene early in Isaac Asimov's Second Foundation -- which is set 1000s of years in our future -- showing the life of a teenage girl before she starts adventuring.<p>Sitting in the bedroom of her suburban house, she uses the new technology of a voice-activated typewriter to write a paper for history class. Looking over this homework assignment, her father reprimands her for one of her writing choices, and she sullenly revises the paper accordingly.<p>(The choice, as you will likely recall if you read the Foundation Trilogy, was to brag about her grandmother, the heroine of the previous book.)
It's a little gloomy to see how early most of the predictions are, they were overly-optimistic by 10-20 years in most cases.<p>Has anyone ever tried aggregating popular predictions over time? It would be interesting to see a history of predictions for self-driving cars up to now.
My main take-away from this article was that people often use the underlying paradigms of today (perhaps unknowingly) to predict surface level details of tomorrow. "Asking the wrong questions", as the article byline states.<p>Instead, we should try to predict how the fundamental paradigms will change, and then deduce visible, commonplace details from that.<p>I did just started reading "The Structure of Scientific Revolutions" by Kuhn, which is all about paradigms (of a specific sort). It's good stuff.
I enjoyed the photo, the family history, and the premise of "asking the right/wrong question". I think that all technologists - or perhaps everyone - can and should take a step back from their daily grind and ask some questions.<p>I think the wisdom of the crowd can be leveraged in asking questions about the future. There are sites like longbets.org where you can weight in.<p>It is one of the great gifts to the human species to be able to contemplate a future beyond tomorrow.
I am going to choose to believe that the wrap-up paragraph, asking "who will have the right kind of driving data for autonomy?" is a subtle wink, because the current conventional wisdom about autonomous cars -- that they are coming soon, and that machine learning on massive datasets is the key to their success -- is very likely going to end up being one of those things we talk about jokingly in 50 years. Ha! Remember when they thought we'd have a colony on the moon in 1990, that machines would have genius IQs by 2000, and that cars could drive themselves by 2025? Oh, the people of the past were so cute.
I thought this would be about how we have prioritized the advancement of technology for the few over the welfare of the rest.<p>Maybe it should be titled "Still asking the wrong questions"
I am not surprised that a telco centric forcasts would be locked into a switched circuit mind-set.<p>eve in 96/97 interviewing at a british telecom board the 3 (all mainframe guys) just did not get the internet at all - ironically the board was at 207 old street