Rethinking Singularity

I have concerns about Ray Kurzweil's Singularity.  The following three stories will show you where I'm coming from and give some background to what I want to say:

(1) In the '80s, Tom Axworthy, then Principal Secretary to Prime Minister Pierre Trudeau (and now with the Center for the Study of Democracy at Queens U in Kingston, Canada and the Gordon Foundation), spoke before my group, the American Association of Political Consultants, and told why Canadians and other countries distinguished themselves from Americans and American political campaign technology.  He said that Americans pursue life, liberty and the pursuit of happiness as a national credo whereas most other societies have as their goals peace, order, liberty and fraternity.  Fraternity being the sharing in the well-being of all of society.  Big difference between the individual pursuit of happiness to the altruistic sharing of the well-being of everyone.  And that difference translates into political orientation, campaign practices and social ethic.  In America elected officials have star status whereas most members of parliaments worldwide are part of the party and not well known.  They are often elected as the x-party member for the y area.  Hence there's less personality and more issue orientation.  Not Barney Frank versus Earl Sholley but instead Liberal versus Tory.  Axworthy's talk has stuck with me to this day because I strongly believe in his version of Fraternity and what it means for society and the future.  Also it was one of the many reasons I chose to sell off and quit my activities in politics.

(2) Ray Kurzweil's projections of logarithmic (exponentially accelerating) technological progress - particularly in the fields of robotics, biotechnology and nanotechnology - leading to a "singularity" or merging of these super-intelligent sciences sometime between 2040 and 2045, a merging where differentiating between a human with consciousness and a robot-like device acting as if it had consciousness, has been fascinating to me because I'm a technology enthusiast, particularly in the areas of computers, AI and robotics. I see it happening just as he says. In robots, genetics, longevity, artificial intelligence, aging, stem cells, and many more sciences, my vision of the future is similar to Kurtzweil's. And this is disturbing because his projections are leading to a conclusion that I don't want for society.

(3) While driving to and from Lake Tahoe last weekend, some friends and I listened to an audiobook entitled Death Match. Although it was a mystery, it was really about artificial intelligence. It involved a computer dating service that went beyond simple questionnaires and instead merged psychological, medical and financial data along with social data such as travel, movie and book preferences, phone call records, traffic tickets, etc. into a massive database which was then sliced and diced to provide information about the candidates well-beyond what they entered on their initial survey forms. Armed with all that data, the computer did it's match and was quite successful. A discussion occurred about individual boundaries, and computer capabilities. Coincidentally, I had recently listened to a podcast of an AI expert discussing how things were presently done (constructivist) and how they will be done shortly (software developing software). This shed light on what was fictional in the story. The discussion continued to include the fact that the story's software and manipulation of massive databases was available today but that it wasn't going to get too much better until more capable and extensive software could be developed and that was precluded because the present state of the art was constructivist (done by human programmers and limited by their time and capacity). Although software is used to create new computer chips, humans are still cranking out AI software. When AI software becomes self-generating, that's when robotics and other embedded sciences will grow - and the dangers I foresee begin.

This brings me to a long and old (2000) Wired Magazine article written by Bill Joy, co-founder and network computer scientist of Sun Microsystems, a VC at Kleiner Perkins Greentech and FOO (Friend of Obama).  In the article, Joy worked his way through his own history of thoughts about technology to an evening when he spent some time with Ray Kurzweil and learned, first-hand, what Kurzweil foresaw.

Ray was saying that the rate of improvement of technology was going to accelerate and that we were going to become robots or fuse with robots or something like that and John [Searle, also at the meeting] countering that this couldn't happen because the robots couldn't be conscious.

I had always felt sentient robots were in the realm of science fiction.  But now, from someone I respected, I was hearing a strong argument that they were a near-term possibility.  I was taken aback, especially given Ray's proven ability to imagine and create the future.  I already knew that new technologies like genetic engineering and nanotechnology were giving us the power to remake the world, but a realistic and imminent scenario for intelligent robots surprised me.

Joy wrote pages of his history in thought from then until he met scholar and author Jacques Attali who described his interpretation of Fraternity.

Jacques helped me understand... Fraternity, whose foundation is altruism. Fraternity alone associates individual happiness with the happiness of others, affording the promise of self-sustainment.

This crystallized for me my problem with Kurzweil's dream. A technological approach to Eternity - near immortality through robotics - may not be the most desirable utopia, and its pursuit brings clear dangers. Maybe we should rethink our utopian choices.

I believe we must find alternative outlets for our creative forces, beyond the culture of perpetual economic growth; this growth has largely been a blessing for several hundred years, but it has not brought us unalloyed happiness, and we must now choose between the pursuit of unrestricted and undirected growth through science and technology and the clear accompanying dangers.

We are getting a belated start on seriously addressing the issues around 21st-century technologies - the prevention of knowledge-enabled mass destruction - and further delay seems unacceptable. 

It seems to me that Joy's seriousness and concern is well-deserved and appropriate.  I share his concerns fully.  What do you think?

Conversation with Chris

I had an informative conversation with a recent LinkedIn acquaintance that I'd like to share. It began in answer to his question:

Just curious, your profile says, "Because I believe that robotics is the next big thing..." What do you envision, in terms of "next big thing"?

I responded:

Robotics, in all its interdisciplinary forms, will be everywhere very soon. In our homes, cars and appliances. At the hospital and in the workplace. Protecting us from afar. It's happening fast but not like in the movies. In America and Europe, it'll be in advanced embedded interactive systems like adaptive cruise control that now handles lane boundaries and will soon handle trucks and busses unmanned in controlled lanes; or in Kiva style warehouses (no fixed shelving; few pick & pack people; heavy computer control); or in smaller and smaller interactive medical devices.

In Japan and Korea, more humanoid-looking robots will be used for personal and factory assistants and we'll all be using exoskeletons of one type or another such as the ones being used for Japan's seniors to help them garden or Honda's factory workers who squat, climb and lift.

It's truly amazing and just beginning to get into stride. Worldwide defense spending is paying for the R&D and smart guys like Rodney Brooks are commercializing that R&D.

My interest is in finding either (or both) a basket of stocks of robotic companies that will rise in price or helping fund a select few companies in need of management, marketing and money to grow.

What do you think?

And then came this wide-ranging informative response that I found particularly illuminating:

When I was younger, I had the good fortune to witness/participate in this same dynamic in the computer industry, WANG Laboratories, Atex, Computervision (located in the same building that iRobot is in, Bedford now) and Sun Microsystems.

While I didn't realize what I was part of and witnessing at the time; it was a technological Darwinism that was unfolding with incremental technical breakthroughs that were being applied, somewhat haphazardly, achieving momentary commercial success and were quickly "leapfrogged" by a new company down the street, often with the same players. Although none of those companies and many others like them are around today, the relatively mature and stable PC we're both using has individual components (soft and hard), developed by those players and their technological derivatives; owing to materials and other advances.

Now I'm older, I realize what I was seeing then but adding a bit of experience and now seeing that in the case of robotics and many other advances; that relatively short period of success that the individual computer players enjoyed will be even shorter for this field. This leapfrog speed is and will be much faster. But, I'm reminded that to this day there are significant numbers of DEC and WANG servers still running vast arrays of applications, many for the government and others for many of the large institutions. The installed user base as it were.

So, what might this mean for robotics? If we look at the practical application of robotics, it will be interdisciplinary as you say and much of it is taking place in all the salient ways you mentioned. While all of those are viable and happening, it's the medical applications that most come to mind as immediate and timely candidates. Rodney Brooks is a great salesman but while his automated floor sweeping and gutter cleaning toys capture the public imagination, its his investment in things like MAKO Surgical that really demonstrate the potential for robotics (hardware-wise).

Take any high skilled operation in medicine, dentistry, exams, etc. study it carefully and you'll find that the operation can be mechanically fixtured and simplified via robotics which will do what they are truly best suited for, namely precision and repeatability. So envision going into a dentist for a root canal and instead of having an overpaid technician pounding in your mouth, using the same primitive tools that I use to file down a piece of aluminum; instead you sit and bite down on a universal, sterilized fixture that is ergonomically designed, and an "operator" is sitting in front of you, precisely directing the root canal "robot" via joystick/camera and or pre-canned software routines to effect all the physical force and position functions. Maybe it's mobile and can go to the patient in some remote areas, "affordable root canals for all." Perhaps this highly paid operator is making $ 60/hour so a root canal costs $600 rather than $ 2000. Further, it's repeatable every time. So the quality, repeatability, low cost, accessibility, etc. all combine as customer needs are met. Other examples exist and they will be found wherever we have a highly skilled operator operating high value equipment or carrying a highly specialized body of knowledge in his head. These are the applications that one might fund, (this paragraph long business plan, may or may not be the best example of it) but a good paradigm is to think of a skilled machinist versus a CNC milling machine. We've got very few skilled machinists today but we machine parts faster, cheaper and with much higher quality and repeatability.

So envision an installed user base of this machine across the country, in some percentage of dental clinics and pretty soon we can quantify the commercial value. That’s not to negate the commercial potential of mundane consumer applications mind you.

I think you're on the right track about the trend, as long as you focus on the substance and not the Hollywood hype, (terminator looking robots or the Kabuki Dancing dolls the Japanese trot out every year or so). The applications are happening all around us and from my perspective, one of the reasons were not moving faster is because some of the guys driving these things are coming out of University laboratories and feel it's important that they re-invent every nut and bolt, lacking practical experience in integration; rather than focusing on that one unique thing that they have developed, (perhaps a faster algorithm, a smaller/cheaper feedback device, wireless power transmission, etc. etc.).

If you visit Heartland Robotics website and go into the manufacturers’ survey link; you’ll see by the questions on that potential end user survey that the graduate student who put that question list together has never been within a mile of a factory, let alone understand what the factory floor may or may not need from a transformational concept like universal robotics. We sort of have solutions looking for problems in that regard.

To be fair to the academics and apologizing for my industrial bias here, there is some significant IP that will/is being developed as fresh eyes accidentally reinvent the “bolt”, material advances, etc. but it is taking the applications much longer to be developed. There's quite a bit of empowering technology already abundantly and cheaply available but it's not being applied, I think. That's not to say this isn't exciting stuff both technically and commercially, but much of it is still at the parlor trick stage as they agonize over things that have long ago been solved in industry.

Sorry for the long reply but this is an area I'm interested and have a bit of experience in and maybe I can help you and your group identify/evaluate some of these opportunities from a technology perspective or add value in other ways.

I'm not sure that I can help with the stock picks approach as I've long ago stopped riding up and down elevators in the Hancock Building listening to shills and I'm confident your grapevine is better than mine but if you're evaluating funding requests, perhaps I can add my two cents worth on the technology/assessment vs. what’s already available or maybe even evaluate the viability of what's proposed from a technical/architectural perspective.

What do you think?