Robotics: We Seem to be Getting Passed Again!

Robotics: We Seem to be Getting Passed Again!

March 3rd, 2007 | Posted in Business & Economy, Science & Tech

I noticed a CNN article about a Japanese robot serving tea today.

Bill Gates wrote an article about robotics that appeared in Scientific American in January, where he makes the case that robotics today is where PC manufacturing was about the time Bill Gates and others started Microsoft. I think he’s right, and that Japan is way ahead of us. Not just because they are serving tea: The ten-year-old in my household recently brought home an article about cool future technologies that highlighted a Japanese medical robot that can lift an 80 pound patient.

Earlier this week, I was at a meeting where the speakers were technology venture captalists. I asked them about investments in robotics here, and they said the level of investment is low even though the overall level of VC investment is fairly high right now.

I do know that the personal robotics market is supposed to grow significantly in the next few years. We need to do more investing here if we want piece of the action.

About Brenda Cooper

Brenda Cooper has been delivering keynote addresses on the future for over a decade. Her non-fiction has appeared in The Futurist Magazine and on Futurist.com. She is also currently writing a non-fiction series called Backing into Eden, which is being re-published on multiple websites. Her talks are upbeat and positive, and generally focus on how we can solve the big problems facing mankind right now.

2 Comments

  1. Luis Beck   |   Jun 18, 2008

    Hi, I am sending you a webpage about Humanoid Robots, it is a Resume about a lot of new technology. If you like new ideas to improve your research or products…
    Luis Beck

  2. Wes Duck   |   Apr 26, 2007

    Robots – I love this subject, thanks for bringing it up.

    I would say that if the future robotics industry resembles the current one for industrial robots, Japan may indeed prevail. Of course, it’s important to understand that the nature of the Japanese demographic curve is substantially ahead of us, tilted as it is with more elderly as a percentage of the total population(many of whom will eventually need care), restrictive immigration policies (limiting the pool of potential care providers for those elderly), and an unusually active cultural interest in robots, means that Japan is more strongly motivated in this area than perhaps others currently.

    However, there are grounds for hope that our current disparity with Japan in this area will not necessarily persist for the long term, for the following reasons.

    A lot depends on the nature of that future robotics industry, particularly models oriented toward consumers. Most of the predictions I’ve come across predict a multitude of specialized robots around the house, which I agree with and attribute mostly to the natural progress of whatever appliance is being roboticized, like the dishwashing machine, for example. Improvements of this kind are straightforward to predict, and actually a little boring, or at least somewhat bereft of inspiration potential for the general public.

    However capably roboticized our individual appliances become, there will still be big gaps in what consumers want done. Maybe the robot dishwasher can wash dishes you place on the counter next to it, but what if I want it to clear the table, zip up the leftovers and put them in the frig, etc? There will be gaps like this with just about every specialized robot, where the owner still has to do work to feed it, in effect. Also, the things that need doing around the house are sometimes unpredictable and/or unstructured, making the utility of a specialized robot problematic. I would say that specialized robots will not meet more than, say, 30-40% of the things that the typical consumer could use help with.

    Because of the limitations of specialized robots, I predict that there will also be a large market, maybe the eventually predominant one for robotics, for at least one multitalented droid that can perform a variety of both known and unexpected tasks, either on a schedule or ad hoc (“hey, come help me get these bricks out of the pickup”). As long as these “droids” resemble more an industrial robot than a living being, Japan could conceivably control this market for the indefinite future.
    However, even if the early versions are metal men, whirring around the kitchen and making chirping noises like r2-d2, the dynamics of consumer demand will not keep them there long. As soon as the requisite technologies are available and affordable, in maybe the next 10-25 yrs, robots will morph into forms that we are most comfortable with – living beings. The current metal doggy Aibo will over time become a dog indistinguishable from the real thing (I don’t know if it will have a variable drool setting).

    This idea may seem unsettling to some now, but those concerns will by necessity be addressed by the industry to make these devices successful in the marketplace. Even though our cars are big, powerful, and could hurt us, we know we control them, and so do not fear them. Something similar I think will occur in this area of humanoid robotics, the public’s anxieties will be addressed by taking the appropriate measures to ensure their safe behavior in every way possible. When that point is reached, everyone will want one.

    Similar to the Aibo example, where the metal dog evolves into a warm, slobbery great dane or beagle that can only be distinguished from the real thing with difficulty, these multitasking robots, although the initial versions may look like c3po, will not stay that way long. I believe a powerful competitive driver in the consumer robotics industry will be how human a manufacturer’s general-purpose robots appear and behave, much as a smooth ride or good gas mileage is with cars today – they will eventually be improved to look like people. Multi-skilled, vastly knowledgeable people. And with things like facial skin actuators, maybe anybody you want. Who wants to hang around c3po when you could have, say, Lauren Bacall in her 20s helping around the house? It is at this transition point between metal men and humanish, and I would venture not before, that robots will truly start to fire the imagination of just about everyone, and everyone who can afford one will want to get one.

    These envisioned robots will need brains as well as looks. And not just computational intelligence – a great personality, listening, understanding, and responsive to its owner’s peccadillos in complex and subtle ways – in other words, charming, patient, proactive, and sharp as a tack. One of the chief risks of having a full time droid around the house will be getting annoyed with it because its too nice, too chatty, always seems to be where you want to go, etc. It will need to have a personality that takes your quirks into account, understands the annoying behaviors and avoiding them carefully. But at the same time, can engage in intelligent and attentive conversation about those obscure interests that you may have that no one else seems to (everyone has at least one of these). Even be rude or act offended if your obviously trying to elicit those responses – in other words, act expectedly human, although an infinitely patient and attentive one. Each of us are different in terms of our quirks and the things that annoy us, so one generic approach will not work – the robot will need to be exceedingly adaptable in this way, as well as others.

    Because of this, the area of robot personality software will represent a large revenue opportunity, rivaling that of today’s PC software market, and I believe the U.S. will shine here. In addition, on the hardware side, making these robots look and move in a convincingly human way will be as much a creative process as a technological challenge, and because of this I believe we should also end up being competitive or even prevailing here as well.
    Make no mistake – one of these devices will be among the most complex things ever produced – by comparison, making a chip or building a car may look like making cookies in grandma’s kitchen.

    Even if the transition point to humanoid robot is not reached for, say, 10 or even 25 years, and even if Japan completely dominates the robotics market up to that point, after jumping in and possibly climbing a challenging learning curve the U.S. could still could take that market back because of the whole host of new technological and creative challenges that human-like robots will represent over previous models. Many of those new skills are in areas in which we historically have great strength.

    Once this market gets established, it should dwarf every other robotics sector (think adoption rates similar to the pc, car, or even the television), and probably larger than entire business sectors in other industries as well. This will be true even if they remain quite expensive. I have no idea what the eventual economics for producing these devices will be, but let’s say the price of a car – $20-30K, something like that. Even at those prices, these devices will sell like hotcakes.

    To compare this presumed future robotics industry to our current competitive situation with Japan in the auto industry (I assume that’s the main “again” referenced in your blog title), there are certainly similarities, but also vast differences. For example, for autos, one of the chief success drivers is the seamless incorporation of often miniscule quality improvements into a large, complex manufacturing process, patiently doing this year after year after year, which eventually accumulate into big differences in quality, reliability, etc. in the U.S, we often like bigger, splashier changes that sometimes are not as effective in the long run. In this area, speaking in a general sense, the Japanese are stronger, or at least more disciplined. If the robotics industry’s win themes are similar to autos in this and other ways, we may indeed a have problem here in the U.S. keeping up with Japan.

    However, it seems reasonable to expect that the rate of change in this industry will resemble less that of autos and more that of microprocessors or software, where large-scale innovation and rapid progress are key. Only this rate of change will affect a cross-section of technologies far broader than the chip industry, autos, or anything else I think of at the moment, with as much focus on, say, creating an accurate feel to the droids “flesh”, the different tasks it can safely perform, the quality of its personality, the keenness of its senses, how fast can it run, its ability to tell a joke, etc, etc. All of these will require bringing together many industry disciplines together to collaborate on a scale that has not been seen before. In all of these areas, the U.S. is quite competitive.

    There will be ethical questions and risks around these devices, of course (what if a droid robs a bank? What if you have an exact replica of yourself made, and send it to work while you have fun – if the droid does your job as well as you would, should that be a crime?). As with any new product, these concerns by definition will have to be addressed with the appropriate safety precautions and disallowed droid behaviors in order for this industry to reach critical mass.

    These robots will quickly be deemed a necessity by many, I expect, much as a car is a necessity today. And even though car wrecks kill thousands each year, because of their great usefulness in getting around no one seriously suggests getting rid of cars – the appeal of talented, humanoid robots will be on that order, I think. And other than the occasional mishap, I don’t think we’ll see anything like the killing fields that annual car mortality statistics represent today. Just the opposite could be the case, in fact – if something happens to you, the droid, always attentive, immediately does the right thing to help ensure your safety, get help, administer cpr, etc. In fact, cpr is a great example of how a human-resembling droid would be preferable to a metal man, as its “mouth” could connect with the real person’s mouth during cpr more naturally. There are undoubtedly many other cases where a utilitarian rationale can be made for humanoid robots, so not only aesthetic considerations will drive the industry in this direction, I believe.

    Even the admittedly valid concern of some unscrupulous person making the robot do bad things can most likely be managed with hardware-driven, absolutely forbidden behaviors. In any case, this risk is also imaginable if the droid was a metal man, but a humanoid robot could probably slip through more cracks, be potentially more dangerous because of its ability to blend into the crowd.

    In addition (my apologies to Detroit), the auto industry in this country has suffered from profound inertia that is primarily managerial and auto business-cultural specific, and are in some ways unique or at least unusually pronounced in that industry. Other factors also contribute, such as the immense capital investment cost to change production economics and processes significantly, at least in terms of technology on the manufacturing floor. There are many reasons for the persisitence of the challenges in this industry, but the most important are probably among the most obvious. For example, a great deal can be attributed simply to the industry’s age – 100 years, and the cultural inertia that has had a very long time to entrench itself.

    Companies and even countries are like people in that way – the young are dynamic (sometimes reckless), the older like to play it safe (which paradoxically, in today’s world is often the more reckless behavior). Even with 30 years with clear signals that they need to change, they’re still searching for the right path. This has far more to do with the U.S. automotive culture than this country’s overall competitiveness.

    However, unlike the auto industry, this future humanoid robotics industry will almost certainly not be managed by an old guard looking backward and changing reluctantly. When the industry reaches this transition, it will be one of the most exciting fields around. Unlike the staid world of industrial robots (who considers this field “exciting”?), this new type of droid industry should attract top talent from just about everywhere. With the combination of a fresh, forward looking managerial perspective (think more Google than Ford), and a talented, enthusiastic workforce, given the other advantages of our economic system, I will predict that we will be hard to beat. I am confident the U.S. will be a very competitive player in this market, if it evolves in the ways I suggest.

    I think a conservative, least-optimistic scenario is that the U.S. will be strong on the software side, the Japanese on the hardware side. In reality, it will almost certainly be some of both, with perhaps the entire world being able to contribute here, given the rich technological, creative, and integration challenges these devices will represent. For example, localizing one of these devices to behave appropriately in, say, Kenya as opposed to Uzbekistan or Phoenix is not just the current typical IT challenge of simply translating a web page to the target country’s language. The droid would need to be able to speak the local language, but that would be just a first step. Detailed info on the local customs and behaviors would be also very important, as well as being able to accurately read cue’s from human faces, even if they are complete stranger – a twinge of the eye, for example, indicating annoyance or insult.

    Come to think of it, based on these and other considerations, let me revise my estimate on the importance of the software side upward – I think it will dwarf the hardware side, at least an order of magnitude larger.

    In terms of these kinds of things, I doubt Japan is ahead of us – we’re all in the same boat at the moment, which is that it’s still a challenge that is titanically difficult, across the board in terms of just about every technology required here. But with robotic and computer intelligence being informed more and more from actual insights on the structure and function of the human brain, I have no doubt we’ll get there, although that could take 25 years or more to do just right, where you can’t really tell you’re talking to a droid. Far from being afraid or creeped out by this droid “humanity”, I think we’ll settle right in, become very comfortable with this very quickly, and soon consider them indispensable. And if this comes to pass, at times some may philosophize about how bad these robots are, make us lazy or whatever. These concerns will nevertheless have no effect on the adoption trend – similar to how we complain how TV has nothing but cruddy programming on it, but we still have one in every room.

    Of course, I could be wrong – it will be fun to observe. My feeling is that programming a convincing personality will be the biggest challenge, and I’m not sure that Moore ‘s law alone is going to ever solve that problem. New algorithms, software structures, and approaches are needed, almost certainly informed heavily from our rapidly expanding knowledge of the human brain.

    Anyway, fascinating subject, thanks for bringing it up.