Weekly Editorials Page
6/14 to 6/20, 2001

    Tonight's robotics articles describe very expensive robots that might, perhaps, be suitable for commercial applications. They're steps in a progression that won't reach Walmart for years, but they're inching their way forward. "Adavanced Imaging" magazine this month has an article about robotic vision. The article "Robotics Vision: Applications & Developments" by Len Yencharis, pg.22 says,
    "According to new market estimates--not mine--mobile robot sales are expected to soar from $665 million to more than $17 billion in five years. Technologies are such that decades of labor in artificial intelligence, sensing, navigation, communications and response are beginning to bear fruit in the form of practical mobile robots. While fixed robotics tools have become increasingly flexible and productive, the self-directed mobile robot and its relatives---various freestanding robots now appearing on the world stage---are expected to transform business and industry across all sectors. The major segments include:

    "In 2001-2005, robots and robotic components more closely resembling humanoids are prepared to revolutionize the way humankind explores, monitors, inspects, controls, builds, repairs, transports, and plays."
    "In the next three to five years, intelligent networked mobile platforms and manipulators will permeate our society. Where computers enhance the processing power of human brains, and networks extend okur senses and nervous systems out into the world around us, robots are poised to serve as extensions of our feet and hands, performing real-world tasks at our command.".

    The next three articles deal with the resurgency of mysticism, magic, vampires, fantasy, astrology, palmistry, and other retreats from reason. For untold thousands of years, magic was the only science we had. Kingdoms rose and fell on the unexpected appearances of comets, eclipses, and other unpredictable manifestations of the divine. Empires cast their lots on the readings of the entrails, the pronunciamentos of the oracles, the dreams of the prophets, and other signs and portents. In all that time, if anyone succeeded in devising a magic incantation that would cause a besom to sweep a room, the secret died with her. Magic mirrors that would reveal to a seer what was hapening in a distant land didn't survive to the 21st century. If they had, we'd be using them instead of TV systems. (Of course, with TV, stories about magic mirrors sound passé now that a reasonable approximation is available to everyone.) As I've mentioned before, we are moving into the age of mental telepathy, only we're calling it a "cellphone". When it becomes embedded inside us, and is linkable to virtually every other person on the planet, then people will no longe fantasize about telepathy.
    As the second article, Magical thinking, points out,
    "Science, the ultimate product of rational thought, is on a roll: One ancient scourge after another is being eliminated, hunger has been reduced to a political problem, life spans have doubled, all the knowledge of the world has been put at the fingertips of ordinary citizens, and the deepest mysteries of the cosmos are unfolding before our eyes. We are even beginning to get a handle on how to keep our planet healthy.
    "Why then is the public turning away from science? Nay, not just turning away, but fleeing in the opposite direction. My bookcase overflows with wonderful, reductionist accounts of how the world works, written by brilliant scientists for nontechnical audiences--Gould, Dawkins, Sagan, Goodenough--but I look in vain for their names on bestseller lists. Instead, I find such pathetic drivel as Deepak Chopra's Ageless Body, Timeless Mind: The Quantum Alternative to Growing Old and Harvard psychiatrist John Mack's Abduction. There is James Van Praagh, Talking to Heaven, while Neale Walsch is having Conversations with God.
    "How are we to account for such widespread nuttiness? Is it indelibly coded into our DNA? Perhaps. In her latest book, Sleeping with Extra-Terrestrials: The Rise of Irrationalism and Perils of Piety, Wendy Kaminer turns to psychologist James Alcock, who points out that all of us, scientists and professional skeptics included, engage at times in magical thinking. It's not surprising. Evolution is a slow business. All of recorded history covers a mere 4,000 years, the space age a mere four decades--far too little time to influence our genes. We are all saddled with genes selected for life in a Pleistocene wilderness. In a time before science, there was no gene for scientific thinking, and there still isn't. It must be learned."
    There is an article in the "Neurology" file, Scientist says gene mutation is key to genius and despair, that postulates that an early mutation gave us genius and madness. It's perhaps because of this that we began burying our dead with "grave goods", and began to draw pictures on cave walls. This may also be the origin of our affinity for the mystic, and for peopling the dark with imaginary gods and spirits.

    I have begun adding news sources to the science news articles.
    The article, Antibiotic Prevents Lyme Disease if Given Quickly -Reuters, explains that just two doxycyclin antibiotic tablets given within 72 hours after a tick bite has an 87% chance of warding off Lyme disease.
    Flashes have been observed on Mars since the 50's. They've been pooh-poohed in the past, but this time, a network of amateurs captured them on film. They're thought to be reflections of sunlight.
    Monster-hunters set sights on Ireland - MSNBC... well, fayth and begorrah! What better place to hunt monsters than on the Auld Sod, home of Darby O'Gill and the Little People.
    The article, Geothermal Power: If It's So Good, Why Isn't it More Popular?, mentions that the U. S. is already drawig about 2,880,000 kilowatts... enough power for 1,500,000 homes... from geothermal sources. We're also harnessing about 2,500,000 kw. of wind-derived power.
    I've upped the number of news items each day from 25 to 40 because I'm building a huge backlog of news releases. When I first started this sinecure, there didn't seem to be enough news releases to carry us through the weekends. (I keep nattering on through weekends, when most other news services shut down.) I hope this doesn't overwhelm you. Hm-m-m, let's see. Maybe we can call this, "Uncle Bob's One-Stop News Shop." (Grin) How does that sound?
    I'm retaining last night's discussions because tonight's comments will update and emend them.
    After re-reading last night's news releases, including one from Intel, "Intel develops 20-nm transistor for 20-GHz processors by 2007", I realized that the 20 nanometer figure quoted in some of these press releases is the gate length of the transistors rather than their size. These transistors will switch at 1.33 trillion times per second (1.33 terahertz), and the entire billion-transistor chip will allegedly dissipate only about 1 watt(!?) at 20 GHz. This generation of chips, dubbed the P1266 chip family is expected to debut in 2007, and will feature 45-nanometer design features.
    In 2009, Intel is planning the next (P1268) chip series, with 0.15 nanometer gate lengths, and 0.035 circuit features (Intel Researchers Build World's Fastest Silicon Transistors).
   "Intel managers here also said the 20-nm transistor proves that silicon ICs will be viable well into the so-called "nanotechnology" era. They said the development debunks the belief that silicon will run out of gas in the next 10 years or so."
   "There's been a lot of talk and concern about the end of Moore's Law," Gerald Marcyk, the director of components research for Intel's technology and manufacturing group, told Reuters this week. "So far, we haven't hit any fundamental limits with respect to our transistor technology." One of the things Andy Grove keeps asking me is, when do they stop  working?" Marcyk said. "And I say I don't know yet. I keep shrinking them, and they keep working."

    "IBM scientists said they believe this development (of carbon nanotube transistors - see April 27 story) could help clear the way for production of ICs once silicon become no longer capable of handling device shrinks in the next 10 to 20 years."
   Note the "next 10 to 20 years" for silicon, with, perhaps, some other concept to follow. See also
Intel unveils world's smallest transistors, and Intel claims world's smallest, fastest transistor.
    Ten years ago, when I wrote the HATS/TABES 40-year computer technology forecast, the word was that computer technology was good until the year 2000, when circuit features would drop to 0.2 microns (200 nanometers). Then we'd hit a brick wall.
    So here we are in 2001, and good to go until 2010, at which point we'll have to fall back and punt. In other words, it looks just as promising today as it did in 1991. I
    t looks as though the state of the art in 2010 will be based upon 35-nanometer design features, down from 180-nanometer features in 2000. We will have gone through 130-nanometer, 100-nanometer, 65-anometer, 45-nanometer and 35-nanometer generations to get there. I would have looked for 25-nanometer and 18-nanometer features by 2010, but 35-nanometers will do nicely, thank you.
    At that rate, we might look for 7 to 8 nanometers by 2020, and that might still be realizable in silicon. At 7 nanometers, features would be 20 to 30 atoms across. But if not, there may be other options by then.
    To anyone looking ahead from 1980 (like me), it would have been utterly mind-boggling that 35-nanometer-feature chips could be built at all. A typical living cell is 10,000 to 20,000 nanometers in length.
    But we're good at least through 2010. And things look as bright as they did in 1991

    Tonight's most interesting news release for me is the Discovery that Magnetic Fields Flatten Space-Time. The idea that someone can unite electromagnetism and general relativity is startling news to me. It brings back questions about Dr. Ning Li's experiments with mmagnetogravitic interactions (NASA Funds Controversial Gravity Shield). The paper is available from the Physical Review of Letters for $15.

Intel Discusses Its Plans Its 1,000,000,000-Transistor, 20MHz PC chips in 2007
    Tonight's lead articles, Intel develops 0.07-micron transistor for 10-GHz processors by 2005, and, Intel develops 20-nm transistor for 20-GHz processors by 2007, may, perhaps, be Intel's response to IBM's announcement that IBM, Sony, and Toshiba plan to introduce a teraflops personal computer chip by 2004, (Sony, IBM, Toshiba chips to bring supercomputers home ), and IBM's announcement yesterday, IBM touts breakthough in faster chip speed, of a 35% increase in computer speed-power products by using "strained silicon". It seems as though IBM has been pulling ahead in the War of the Words lately. Now Intel has responded with a peek at its own product-introduction agenda, planning to deliver computers 10 times faster than the upcoming 2 GHz Pentium 4, with 1,000,000,000 transistors on the chip instead of the Pentium 4's 42,000,000.
What Are the Practical Consequences of This?
    Some of the practical consequences have to do with much better speech recognition, virtual reality, telepresence, and computer intelligence. If IBM's, Sony's, and Toshiba's teraflops computer really arrives on our desktops by, say, 2004, then by 2010, it might be showing up in some pretty smart household and industrial robots.
Reason for Multiple Articles
    I'm including several articles concerning both IBM's articles and Intel's articles because I think these announcements are important, and different articles give different slants on these subjects.
Implications of, and Questions About Intel's Announcement
    Intel is basically announcing what one would logically expect by 2007, but its annoucement acknowledges that such technological goals are feasible. Intel is saying that these goals can be met using more-or-less-conventional silicon designs and manufacturing techniques. To me, Intel's announcement is significant for two reasons.
    First, Intel is planning 0.065-micron (65-nanometer) circuit features by 2005, rather than the 70-nanometer features that have preciously been described for that time frame, and 45-nanometer design features in 2007 instead of 50-nanometer features.
    Second, Intel is quoted as planning to utilize 30-nanometer transistors in their 65-nanometer circuitry in 2005, and 20-nanometer transistors.
    How does one implement 30-nanometer-wide transistors in circuitry whose smallest features are 65-nanometers wide, and 20-nanometer-wide transistors in circuits with 45-nanometer minimum elements? A shrinkage from today's 180-nanometer circuitry to 65-nanometer circuits should permit, perhaps, a 7- or 8-folding of processor speeds, and to 45-nanometer circuits, should allow a 16-folding of speeds, if my notions about speeds versus sizes is correct. I'm wondering if something has been lost in the translation, with 20- and 30-nanometer transistors slated for later than 2007. The logical next steps beyond 45-nanometer circuits might be to 30-nanometer features in 2009, and then to 20-nanometer circuits in 2011.
Speed-Power Product, and Speculations about 2007, 2010
    A couple of other points that sound interesting are the facts that Intel feels it can mutiply the number of transistors on its chips by 24 while simultaneously multiplying their clock speed by a factor of 10... increasing the speed-power product by a factor of 240... and all this without increasing the total power output. Also, IBM's teraflops computer is predicated upon 100-nanometer circuit features in 2004-2005. Circuit features of 45-50 nanometers in 2007 might afford 4 to 5 teraflops performances on the basis of clock speed alone, and possibly 10 to 15 teraflops, given additional transistors on the chip supporting a higher degree of parallelism. (Of course, this is highly fanciful on my part.)
    In the realm of flights of fancy, I might predict 50 GHz computer clocks by 2010, 250 GHz clocks by 2015, and terahertz+ clocks by 2020. But time will tell.
Computer-to-Computer Software Translator
    The Software switches code on-the-fly describes a translator that will convert instructions for one computer to another during execution. This would permit software written for an IBM computer to run on an IBM or Motorola G4. (Of course, something that would rapidly convert software before it began to run would seem to be a very valuable tool.
Who Needs Faster Computers?
    The article, Who needs more than a gigahertz?, is a classic "who-needs-it" kind of discourse. Similar articles have dogged the footsteps of computer development since the computer's inception in the1940's. Thomas  Watson, the son of the founder of IBM, is reputed to have said in the 1950's that there wouldn't be a market for than a dozen 704 computers in the whole country. Paul Furmeister, the Director of the Computing Center at NASA's Langley Research Center, told me in 1967, as he was driving me to the airport, that in the 50's no one could foresee a market for more than a dozen 1000-bit (128-byte) Magnetic RAM memories in the country. And there were similar arguments in the early 90's about the frivolity of CD-ROM's and "multimedia".
Spaceships Made Of Concrete?
    For more than a decade now, UAH (the University of Alabama in Huntsville) has been competing in a contest to build a concrete canoe. The latest pictures in the Huntsville Times show this year's UAH entry looking like an ordinary "Cyclolac" canoe. Apparently, UAH has found ways to render concrete flexible and very strong.
Baldness pill 'passing early tests' Listen up, guys! Sounds as though this is for us! (Wonder if it beats minoxidil?)