New exotic material could revolutionize electronics

Announcements of "breakthroughs" like this often don't pan out. This one might be a little more promising. It involves confirmation of a new, fairly simple material – bismuth telluride (Bi₂Te₃), a compound of bismuth and tellurium – that exhibits what is known as the "quantum spin Hall effect" and could have a dramatic impact on electronics, computing, and other significant areas of technology.

One reason for the optimism is that the technology for working with bismuth telluride is similar to that for commonplace semiconductors like silicon.

New exotic material could revolutionize electronics (6/15/09)

SLAC National Accelerator Laboratory and Stanford University have confirmed the existence of a type of material that could one day provide dramatically faster, more efficient computer chips.

Recently-predicted and much-sought, the material allows electrons on its surface to travel with no loss of energy at room temperatures and can be fabricated using existing semiconductor technologies. Such material could provide a leap in microchip speeds, and even become the bedrock of an entirely new kind of computing industry based on spintronics, the next evolution of electronics.

Materials with the properties of bismuth telluride have been predicted theoretically, and in this case the predictions have been pretty accurate. In other words, physicists already have a good understanding of the material's properties.

This magic is possible thanks to surprisingly well-behaved electrons. The quantum spin of each electron is aligned with the electron's motion—a phenomenon called the quantum spin Hall effect. This alignment is a key component in creating spintronics devices, new kinds of devices that go beyond standard electronics. "When you hit something, there's usually scattering, some possibility of bouncing back," explained theorist Xiaoliang Qi. "But the quantum spin Hall effect means that you can't reflect to exactly the reverse path." As a dramatic consequence, electrons flow without resistance. Put a voltage on a topological insulator, and this special spin current will flow without heating the material or dissipating.

Practical implementations of spintronics have been avidly sought, because the technolgy takes advantage for the first time of an electron's spin, as opposed to its electric charge. This may enable the manufacture of much faster and denser forms of digital information storage devices, and even more exotic things like quantum computers.

The quantum spin Hall effect mentioned above is a quantum version of a non-quantum effect, the spin Hall effect, known for about ten years. That, in turn, is an analog of the classical Hall effect, which has been known since 1879.

In the classical Hall effect, a voltage difference is produced in an electrical conductor transverse to an electrical current in the conductor. A magnetic field is also produced perpendicular to the current.

In the quantum spin Hall effect, there is also an electric current, and in fact electrons flow without dissipating heat. Consequently, for example, transistors that take advantage of the effect could be much more efficient than existing semiconductor transistors.

Importantly, in bismuth telluride, the effect occurs at much higher temperatures than those at which known superconducting materials work. This will make practical applications much easier.

The predictions on which the discovery is based are quite recent, having been published only in May of this year – see Super-efficient Transistor Material Predicted.

Research abstract:

Experimental Realization of a Three-Dimensional Topological Insulator, Bi₂Te₃

Three-dimensional topological insulators are a new state of quantum matter with a bulk gap and odd number of relativistic Dirac fermions on the surface. By investigating the surface state of Bi₂Te₃ with angle-resolved photoemission spectroscopy, we demonstrate that the surface state consists of a single nondegenerate Dirac cone. Furthermore, with appropriate hole doping, the Fermi level can be tuned to intersect only the surface states, indicating a full energy gap for the bulk states. Our results establish that Bi₂Te₃ is a simple model system for the three-dimensional topological insulator with a single Dirac cone on the surface. The large bulk gap of Bi₂Te₃ also points to promising potential for high-temperature spintronics applications.

Further reading:

Topological insulators in Bi₂Se₃, Bi₂Te₃ and Sb₂Te₃ with a single Dirac cone on the surface – abstract of May 2009 research paper in Nature Physics reporting predictions of topological insulators

Hydrogen economy?

Here's a little more discussion of the prospects for a "hydrogen economy". I'd welcome it if it were practical, but I'm still skeptical. (Some previous discussion of hydrogen here).

Hydrogen economy sustainable in 15 years (7/17/08)

Hydrogen would be most efficient when used in fuel cells, which extract energy via a chemical reaction rather than by combustion. But fuel cells are still very expensive and distributing hydrogen to consumers would require new infrastructure. Consequently, a large-scale transition to hydrogen will require help from the federal government.

More on this report: Hydrogen Vehicles Coming Soon? Two Million Could Be On Roads By 2020 (7/17/08), Fuel cell cars still 15 years away at best: study (7/17/08)

Looking at Hydrogen to Replace Gasoline in Our Cars (7/3/08)

The jury is still out on whether hydrogen will ultimately be our environmental savior, replacing the fossil fuels responsible for global warming and various nagging forms of pollution. Two main hurdles stand in the way of mass production and widespread consumer adoption of hydrogen “fuel cell” vehicles: the still high cost of producing fuel cells, and the lack of a hydrogen refueling network.

Reining in manufacturing costs of fuel cell vehicles is the first major issue the automakers are addressing. While several have fuel cell prototype vehicles on the road—Toyota and Honda are even leasing them to the public in Japan and California—they are spending upwards of $1 million to produce each one due to the advanced technology involved and low production runs. ...

Another problem is the lack of hydrogen refueling stations. Major oil companies have been loathe to set up hydrogen tanks at existing gas stations for many reasons ranging from safety to cost to lack of demand. But obviously the oil companies are also trying to keep customers interested in their highly profitable bread-and-butter, gasoline.

More problems with alternative energy

It's certainly not encouraging to find research that casts doubt on the economic or technological viability of energy sources other than fossil fuels. There are two compelling reasons the whole world needs to transition away from fossil fuels: (1) To derive energy from fossil fuels requires combustion, which releases large amounts of the greenhouse gas CO₂, and other pollutants, into the atmosphere. (2) Resources of the most convenient (for use in land vehicles and aircraft) fossil fuel – petroleum – are rapidly dwindling, especially in easily accessible locations, so the price will necessarily increase over time, as will the risk of armed conflict between countries to protect access.

On the other hand, it's necessary to be as realistic as possible about the alternatives, in order to avoid heading down a path that could be (quite expensively) wrong.

So in a couple of recent posts here and here I discussed various studies that raised cautions about the potential for solar energy, nuclear energy, hydrogen (for fuel cells), and biofuels. New cautionary reports about these various alternatives keep appearing.

Let's begin with solar. There's more than one way to take advantage of energy from the Sun. Photovoltaic production of electricity directly from sunlight receives the most attention. But there are other approaches, such as solar thermal energy and perhaps (some day) even the use of artificial photosynthesis:

Artificial Photosynthesis Moves A Step Closer (4/28/08)

Imagine a technology that would not only provide a green and renewable source of electrical energy, but could also help scrub the atmosphere of excessive carbon dioxide resulting from the burning of fossil fuels. That’s the promise of artificial versions of photosynthesis, the process by which green plants have been converting solar energy into electrochemical energy for millions of years. To get there, however, scientists need a far better understanding of how Nature does it, starting with the harvesting of sunlight and the transporting of this energy to electrochemical reaction centers.

But most commonly when "solar energy" is discussed, photovoltaic technology is what's actually meant. The technology has existed for many years. The problem has always been cost. Even today it's estimated to be almost ten times as expensive per kilowatt-hour to generate electricity with photovoltaics as it is from fossil fuel:

Expert Foresees 10 More Years Of Research & Development To Make Solar Energy Competitive (4/7/08)

The single biggest challenge, Gray said, is reducing costs so that a large-scale shift away from coal, natural gas and other non-renewable sources of electricity makes economic sense. Gray estimated the average cost of photovoltaic energy at 35 to 50 cents per kilowatt-hour. By comparison, other sources are considerably less expensive, with coal and natural gas hovering around 5-6 cents per kilowatt-hour.

Because of its other advantages -- being clean and renewable, for instance -- solar energy need not match the cost of conventional energy sources, Gray indicated. The breakthrough for solar energy probably will come when scientists reduce the costs of photovoltaic energy to about 10 cents per kilowatt-hour, he added. "Once it reaches that level, large numbers of consumers will start to buy in, driving the per-kilowatt price down even further. I believe we are at least ten years away from photovoltaics being competitive with more traditional forms of energy."

Major challenges include developing cheap solar cells that work without deterioration and reducing the amounts of toxic materials used in the manufacture of these cells. But producing low cost photovoltaics is only a step in the right direction. Chemists also need to focus on the generation of clean fuels at costs that can compete with oil and coal.

Nuclear power continues to be quite problematical too. Many of its problems have been known for a long time, such as the difficulty of safely disposing of spent fuel and the dangers of diversion of nuclear fuel to weapons. But more recently attention has begun to focus on the increasing cost of extracting and processing uranium, and the greenhouse gases generated in that process:

Questioning Nuclear Power's Ability To Forestall Global Warming (4/30/08)

Rising energy and environmental costs may prevent nuclear power from being a sustainable alternative energy source in the fight against global warming, according to a new study.

In the article, Gavin M. Mudd and Mark Diesendorf investigate the "eco-efficiency" of mining and milling uranium for use as fuel in nuclear power plants. ...

The study points out that supplies of high-grade uranium ore are declining, which may boost nuclear fuel's environmental and economic costs, including increases in energy use, water consumption and greenhouse gas emissions. In addition, newly discovered uranium deposits may be more difficult to extract in the future -- a further drain on economic and environmental resources.

Here are two somewhat longer articles about this report, with some rebuttal from the uranium mining industry: Nuclear may lose green tag if fuel costs rise and Nuclear energy becoming less sustainable.

Finally, it's often overlooked that there's one additional increasingly scarce resource that's usually needed to produce energy by many different technologies: fresh water.

Water Needed To Produce Various Types Of Energy (4/17/08)

It is easy to overlook that most of the energy we consume daily, such as electricity or natural gas, is produced with the help of a dwindling resource – fresh water. Virginia Tech professor Tamim Younos and undergraduate student Rachelle Hill are researching the water-efficiency of some of the most common energy sources and power generating methods.

When the requirements for fresh water are considered, new disadvantages appear for many alternative energy sources, especially biofuels and nuclear energy:

According to the study, the most water-efficient energy sources are natural gas and synthetic fuels produced by coal gasification. The least water-efficient energy sources are fuel ethanol and biodiesel.

In terms of power generation, Younos and Hill have found that geothermal and hydroelectric energy types use the least amount of water, while nuclear plants use the most.

Update, 5/14/08:

More headaches of nuclear energy. There's a new, longish Scientific American article on the problem of nuclear fuel recycling: Nuclear Fuel Recycling: More Trouble Than It's Worth


Tags: nuclear power, solar power, photovoltaic power, ethanol, biofuel

More about alternative energy

About a month ago, I wrote about the shortcomings of various alternative energy sources. That was mainly about a variety of problems with nuclear energy, solar energy (photovoltaics), and hydrogen.

I didn't even get into the subject of biofuels, but I should have, because the problems in that area are becoming painfully obvious.

Ordinarily I would not expect to find much significant reporting on a scientific/technical subject in Time magazine, especially something that challenges "conventional wisdom". But via DarkSyde at Kos I see there's an interesting article on the problems of "biofuel": The Clean Energy Scam

Several new studies show the biofuel boom is doing exactly the opposite of what its proponents intended: it's dramatically accelerating global warming, imperiling the planet in the name of saving it. Corn ethanol, always environmentally suspect, turns out to be environmentally disastrous. Even cellulosic ethanol made from switchgrass, which has been promoted by eco-activists and eco-investors as well as by President Bush as the fuel of the future, looks less green than oil-derived gasoline.

Meanwhile, by diverting grain and oilseed crops from dinner plates to fuel tanks, biofuels are jacking up world food prices and endangering the hungry.

The Time article focuses on the loss of rainforest, and consequently the loss of its ability to soak up and sequester CO₂. When the forest is gone, CO₂ will still be incorporated in biomass (crops of some sort). But then that is converted to biofuel, and released back into the atmosphere when it's burned. (To say nothing of the energy that's just wasted along with release of CO₂ when the forest biomass is burned to clear it away.) Given all the energy that has to be expended to grow and harvest biofuel crops, with resulting additional release of CO₂, we are worse off in terms of greenhouse gas emissions than if we just burned oil (or even coal).

But that's not the only serious problem. Crops that are grown to make fuel (from sugar cane, corn, switchgrass, or whatever) use land where food crops (for people and animals) could be grown instead. Driving up the cost of food for everyone on the planet. (Have you checked the price of bread or eggs at the market recently?)

Economists have spoken out about this problem for several years, when the hype for biofuels and ethanol was just beginning to build. For instance, we have from Howard Simons in early 2006: Making Our Food Fuel Isn't the Answer

If high prices strengthen energy's claim on food supplies, governments everywhere will intervene on behalf of their hungry citizens. If low prices torpedo biofuels' economics, governments everywhere will respond with subsidies for these industries. Only an elimination of current mandates and subsidies today will avoid these problems tomorrow, but the likelihood of this happening is near zero. Somehow I believe we will rue the day when we decided to make food and fuel substitutes at the margin.

In early 2007 Paul Krugman picked up the story: The Sum of All Ears

There is a place for ethanol in the world’s energy future — but that place is in the tropics. Brazil has managed to replace a lot of its gasoline consumption with ethanol. But Brazil’s ethanol comes from sugar cane.

In the United States, ethanol comes overwhelmingly from corn, a much less suitable raw material. In fact, corn is such a poor source of ethanol that researchers at the University of Minnesota estimate that converting the entire U.S. corn crop — the sum of all our ears — into ethanol would replace only 12 percent of our gasoline consumption.

So ethanol doesn't even help the U. S. all that much in terms of dependence on foreign oil. And this February Krugman returned to the subject here, linking to this: Ethanol Demand in U.S. Adds to Food, Fertilizer Costs

About 33 percent of U.S. corn will be used for fuel during the next decade, up from 11 percent in 2002, the Agriculture Department estimates. Corn rose 20 percent to a record on the Chicago Board of Trade since Dec. 19, the day President George W. Bush signed a law requiring a fivefold jump in renewable fuels by 2022.

Increased demand for the grain helped boost food prices by 4.9 percent last year, the most since 1990, and will reduce global inventories of corn to the lowest in 24 years, government data show. While advocates say ethanol is cleaner than gasoline, a Princeton University study this month said it causes more environmental harm than fossil fuels.

And then last week Krugman had even more: Grains Gone Wild

The subsidized conversion of crops into fuel was supposed to promote energy independence and help limit global warming. But this promise was, as Time magazine bluntly put it, a “scam.”

This is especially true of corn ethanol: even on optimistic estimates, producing a gallon of ethanol from corn uses most of the energy the gallon contains. But it turns out that even seemingly “good” biofuel policies, like Brazil’s use of ethanol from sugar cane, accelerate the pace of climate change by promoting deforestation.

And meanwhile, land used to grow biofuel feedstock is land not available to grow food, so subsidies to biofuels are a major factor in the food crisis. You might put it this way: people are starving in Africa so that American politicians can court votes in farm states.

Here's a report of a scientific study on the issue: Some Biofuels Risk Biodiversity And Could End Up Harming Environment

Corn-based ethanol is currently the most widely used biofuel in the United States, but it is also the most environmentally damaging among crop-based energy sources.

Finally, to bring this back to a solid scientific foundation, Sean at Cosmic Variance reminds us that Energy Doesn’t Grow on Trees

In particular, biofuels (such as ethanol) and hydrogen are not actually sources of energy — given the vagaries of thermodynamics, it costs more energy to create them than we can get by actually using them, as there will inevitably be some waste heat and entropy produced

.

Although all this bad news about just about every prospective near-term form of alternative energy is discouraging, there are a few other options that may become available in the slightly more distant future. There's the old perennial, controlled nuclear fusion. Even though work on that is even more active than ever, it's still at least several decades away.

But there's another significant option that's often overlooked: solar power satellites. This technology uses very large arrays of photovoltaic panels high in orbit around the earth. The energy is beamed back to the ground in the form of microwaves. (So this should not be confused with simply using mirrors to redirect additional sunlight, which presents serious problems of its own.)

Solar power satellites also have many uncertainties and potential problems, but the largest is simply boosting enough of them into orbit, and maintaining them. A possible approach to those problems involves space elevators. But those, again, present a whole additional set of challenges.

For now, here are a couple of articles from last fall with more details:

Pentagon backs plan to beam solar power from space

New Space Solar Power Report from DoD NSSO

Tags: ethanol, biofuel, alternative energy, solar power satellite

Alternative energy sources

The outlook on energy alternatives to fossil fuels is looking a little bleak.

There have been several recent studies or reports casting significant doubt on the economic and/or environmental viability, at least for the near and intermediate future, of some of the leading contenders to supplant fossil fuels.

First up: nuclear power. Of course, environmentalists and others have had grave doubts about nuclear for decades, because of problems with safe disposal of spent nuclear fuel and the dangers of diversion of enriched uranium to manufacture of weapons. On top of that, there is the argument that replacing generation of power from burning fossil fuels with generation from nuclear sources may well contribute more to release of CO₂ into the atmosphere than continuing to use fossil fuels. This comes about because so much power (generated from burning of fossil fuels) will need to be expended simply to build from scratch many new nuclear power plants and sharply increase the mining and purification of uranium:

Nuclear Power Not Efficient Enough To Replace Fossil Fuels, Study Finds

Nuclear energy production must increase by more than 10 percent each year from 2010 to 2050 to meet all future energy demands and replace fossil fuels, but this is an unsustainable prospect. According to a report published in Inderscience's International Journal of Nuclear Governance, Economy and Ecology such a large growth rate will require a major improvement in nuclear power efficiency otherwise each new power plant will simply cannibalize the energy produced by earlier nuclear power plants.

Here's another way to look at this. If you consider just the marginal costs of producing a kW of energy from nuclear fuel vs. fossil fuel – counting (if you can) both direct economic costs and costs due to release of CO₂ into the atmosphere – nuclear energy might be superior. However, if you also consider the capital expense (both direct and indirect) required to build enough new nuclear facilities to replace existing conventional facilities and also meet increased demand, then (according to the study) nuclear loses.

So what about using other energy sources as alternatives to fossil fuels, in order to significantly reduce dependency on fossil fuels and release of CO₂? Like hydrogen, for example. Of course, this depends on further developing a lot of technology that's either not cost-competitive yet (fuel cells) or not even available yet (practical and safe means of storing and transporting hydrogen). To say nothing of the capital costs (as above) needed to build hydrogen infrastructure if and when the technology is available.

Even if technology can solve the difficult problems of storing and transporting hydrogen, there's another fundamental problem. Hydrogen itself is more of a form of energy suitable for transport and storage than it is a readily available source of energy (like sunlight or fossil fuels) that can be acquired or extracted (relatively) cheaply. There's no hydrogen just sitting around (like natural gas) waiting to be mined and distributed. Energy has to be consumed in order to separate hydrogen from oxygen, which together make up H₂O. This energy has to come from some other source, as input to the electrical/chemical process that separates out hydrogen (or recombines it to make another fuel such as methane). This energy is regained later – but always with some percentage loss – when hydrogen is chemically recombined with oxygen (as in a fuel cell).

There really isn't any energy advantage to hydrogen at all, except for the (presumed) advantage over batteries in storage and transport. Of course, energy in a storable form is required for use in vehicles like cars and airplanes, in spite of the unavoidable losses along the way. The following essay goes into all of this in more detail.

The Hydrogen Economy

Skeptics scoff at perpetual motion, free energy, and cold fusion, but what about energy from hydrogen? Before we invest trillions of dollars in a hydrogen economy, we should examine the science and pseudoscience behind the hydrogen hype.

There are some problems with the essay. First, one does not "make" hydrogen. It is extracted from chemical compounds like water, hydrocarbons (fossil fuels except coal), or biomass (carbohydrates, cellulose, etc.). Energy has to be input to the process in order to break the chemical bonds between hydrogen and other elements (carbon or oxygen). You get the energy back out when hydrogen recombines with oxygen or carbon (in a fuel cell, combustion chamber, etc.) – but always at some loss.

Second, the essay mostly assumes hydrogen will be stored and transported in liquid form, which is difficult and expensive, since liquid hydrogen boils at an ultracold -253°C. There is some hope that technology can be developed to store gaseous hydrogen in exotic solid materials at reasonable temperatures and pressures. (Recent examples: here, here.) However, at this point that's still conjectural. The larger point is that a practical "hydrogen economy" is still, at best, not in the near future.

So hydrogen is not an energy source, and it is even very problematical as a way to store energy in a portable form for use in cars and airplanes. Fortunately, there are other ways to make energy portable, such as batteries. A Toyota Prius uses nickel metal hydride batteries to store energy from the regenerative braking system, and it seems to be an economically successful product. Lithium ion batteries, such as are used in laptop computers, have a higher energy density than the nickel metal hydride type. They have problems of their own, but significant improvements are being made. (See here, here, here.)

That still leaves the problem of developing additional actual sources of energy, that are alternatives to fossil fuels. Ethanol (grain alcohol) is getting a lot of publicity these days. It's politically popular with the agricultural industry, for obvious reasons. Ethanol partially solves one problem with fossil hydrocarbon fuels – by removing some dependence on politically unstable areas as a fuel source. But ethanol does nothing for the problem of CO₂ emissions.

And it creates serious problems of its own, such as driving up the cost of agricultural products needed to feed people. Further, as with hydrogen, it takes a lot of energy to extract ethanol (or other energy carriers such as other biofuels or methane) from agricultural crops or biomass. Critiques of ethanol and other biofuels are not new, though they don't seem to get the attention they deserve. (See here, here, here, for example.)

Other alternatives? There's always solar (photovoltaic) energy. Of all new but currently available alternative energy sources to fossil fuels (whether oil, natural gas, or coal), solar seems to be the most economical, especially taking reduced CO₂ emissions into account.

But of course, solar also has its problems too. These include capital costs for building infrastructure to capture solar energy and to store it (for peak or nighttime use) or transmit it from the sunniest areas with low land prices. It's these capital costs (initial construction and eventual replacement) that hurt, since the marginal cost of each kWh is almost nil.

However, making detailed economic comparisons with traditional energy sources is rather difficult, as this study argues: Cloudy Outlook For Solar Panels: Costs Substantially Eclipse Benefits.

It would seem that the real difficulty of economic analysis lies in predicting the future costs of conventional energy sources – fossil fuels, especially oil. Some of the problems:

• How to estimate costs associated with CO₂ emissions, given that the idea of global warming itself is so controversial (especially in the minds of economists and political officials, if not atmospheric scientists). To say nothing of estimating social costs of conjectural side effects, such as sea level rise, serious water shortages, detrimental impact on human and animal health, impact on agricultural production, etc.
  
• How to estimate the foreseeable rise in price of fossil fuels (especially oil) due to political instability, rising extraction costs (deep ocean sources), depletion of supplies, and rapid increase in demand from developing parts of the world. (There are large uncertainties in all of these factors, and some cost has to be allocated to this uncertainty itself.)
  
• How to handle the issue of proper pricing for energy at times of peak demand, as opposed to off-hours. (The report just mentioned discusses this.)
  

At present, the cost of solar energy, taking into account such things as installation costs, depreciation, etc., might well be two to four times the cost of energy from fossil fuels. But at least the cost of solar is pretty certain to decline, while the cost of energy from fossil fuels can only increase – and at a worrisomely unpredictable rate, in view of the uncertainties just listed.

Tags: nuclear power, solar power, photovoltaic power, hydrogen economy, ethanol, biofuel

Exascale computing

I don't have a whole lot new to say on this topic right now. Earlier comments are here. To recap, "petascale" refers to computers capable of doing 10¹⁵ arithmetic operations ("flops") per second. Such computers already exist. The next main step is "exascale" – computers 1000 times as fast, capable of doing 10¹⁸ operations per second. Such computers don't exist yet – but people are already starting to work towards this goal.

One Million Trillion 'Flops' Per Second Targeted

Preparing groundwork for an exascale computer is the mission of the new Institute for Advanced Architectures, launched jointly at Sandia and Oak Ridge national laboratories. ...

The idea behind the institute —under consideration for a year and a half prior to its opening — is “to close critical gaps between theoretical peak performance and actual performance on current supercomputers,” says Sandia project lead Sudip Dosanjh. “We believe this can be done by developing novel and innovative computer architectures.”

Ultrafast supercomputers improve detection of real-world conditions by helping researchers more closely examine the interactions of larger numbers of particles over time periods divided into smaller segments.

Tags: petascale computing, exascale computing, supercomputers, Moore's Law

Japanese robot not quite ready for prime time

No further comment...

New Hitachi Robot Rolls Around, Crashes

Hitachi's new toddler-like robot rolled around and waved for reporters Wednesday, only to crash into a desk and demonstrate the challenge of turning automatons into everyday helpers.

The red and white robot, designed to run errands in offices, wasn't prepared for the jam of lunch-break wireless network traffic at the company's research center. Unable to communicate with its handler's laptop, it smashed into the office furniture as reporters gasped.

Petascale computing and beyond

National Science Board Approves Funds For Petascale Computing Systems

Today [Aug. 14, 2007] the National Science Board (NSB) approved a resolution authorizing the National Science Foundation (NSF) to fund the acquisition and deployment of the world's most powerful "leadership-class" supercomputer, proposed in response to NSF's "Track 1" supercomputing solicitation. This "petascale" system is expected to be able to make arithmetic calculations at a sustained rate in excess of a sizzling 1,000-trillion operations per second (a "petaflop" per second) to help investigators solve some of the world's most challenging science and engineering research problems.

This would be a significant milestone – "petascale" supercomputers that can process 1000 trillion floating point operations per second.

But there's something puzzling about this announcement. A little further on it describes the "Track 1" system:

In the first award, the University of Illinois at Urbana-Champaign (UIUC) will receive $208 million over 4.5 years to acquire and make available a petascale computer it calls "Blue Waters," which is 500 times more powerful than today's typical supercomputers. The system is expected to go online in 2011.

This system is described as being 500 times as powerful as a "typical" supercomputer today. It's not clear what is being assumed as "typical", but according to the Wikipedia article, as of August 2007 the fastest supercomputer currently installed and operational is an IBM Blue Gene/L at Lawrence Livermore National Laboratory, which is rated at 280 teraflops, or .28 petaflop. Of course, that's not a "typical" supercomputer, but one wonders exactly what is expected for this proposed "Track 1" system.

However, it appears that an actual petaflop supercomputer will be available this year, not 2011, if IBM meets the objectives with "Blue Gene/P" described here:

IBM Triples Performance of World's Fastest, Most Energy-Efficient Supercomputer

ARMONK, NY - 26 Jun 2007: IBM (NYSE: IBM) today announced Blue Gene/P, the second generation of the world's most powerful supercomputer. Blue Gene/P nearly triples the performance of its predecessor, Blue Gene/L -- currently the world's fastest computer. ...

The IBM® System Blue Gene®/P Solution scales to operate continuously at speeds exceeding one "petaflop" -- or one-quadrillion operations per second. ...

The U.S. Dept. of Energy's Argonne National Laboratory, Argonne, Ill., will deploy the first Blue Gene/P supercomputer in the U.S. beginning later this year. In Germany, the Max Planck Society and Forschungszentrum Julich also plan to begin installing Blue Gene/P systems in late 2007.

If the schedule and performance objectives are met, then the petascale threshold will be crossed little more than 10 years after the terascale threshold was crossed in 1997 – by ASCI Red. So there would be a factor of 1000 top speed growth in a little over 10 years – which represents nearly a doubling of top speed every year.

It will be interesting to observe whether this keeps up. That would mean we could see exascale computing (an exaflop, which is one million teraflops) in 2017. This will depend to a large extent on whether Moore's Law holds up 10 more years, and that's not by any means a done deal.

Between 1997 and 2007, smallest silicon integrated circuit feature sizes decreased from about 500nm to 45nm, so a 100-fold improvement in areal density was achieved together with 1000-fold improvement in top computer speed. The 100-fold increase in circuit density translates to a similar increase in speed, assuming that results from being able to fit 100 times as many instruction processors in roughly the same size package. The additional factor of 10 in performance presumably results from some combination of faster processor "clock speed", internal parallelism, and larger total system size.

Somehow I don't see general purpose quantum computers available within 10 years, yet some technology beyond current silicon chips will most likely be needed to keep speed doubling going until 2017. State of the art feature sizes this year are at 45nm. Even a 10-fold improvement to 4.5nm – which may not be feasible – would yield only 100-fold improvement in number of circuits per unit area. As in the previous decade, additional performance improvements will be required to add a further factor of 10 in total performance. What techniques might allow this? For example, will it be possible to fabricate practicable 3D multilayer devices by 2017?

Here are just some of the engineering challenges that must be overcome in order to push feature sizes towards 5nm with anything like current silicon chip designs:

• Preventing current leakage between adjacent features
  
• Dissipation of heat from more densely packed circuit elements
  
• Development of new fabrication technologies if current lithographic techniques don't scale down as far as necessary
  
• Testing and error detection in nanoscale circuits, and achieving acceptable manufacturing "yield"
  

Perhaps exascale computing power will be within reach by 2017. But it won't be easy. To be accomplished with anything like current silicon technology assumes that trends of the past 10 years continue in both decreasing feature size and the ability to squeeze another factor of 10 in speed from additional parallelism, clock speed, or other means.

Or perhaps it will be necessary to successfully implement radical new technologies, such as circuits fabricated with carbon nanotubes.

Tags: petascale computing, exascale computing, supercomputers, Moore's Law

Virtual reality to get its own network?

This could be very interesting if it's not, as some suggest, a scam:

Virtual reality to get its own network?

A nonprofit group says it plans to build a network called Neuronet purely to support virtual-reality game and business applications.

Neuronet, which is planned to be separate from the Internet, "will evolve into the world's first public network capable of meeting the data transmission requirements of emerging cinematic and immersive virtual-reality technologies," according to a Thursday announcement from the Vancouver-based International Association of Virtual Reality Technologies.

For more, see the home page of the group that's promting this: International Association of Virtual Reality Technologies

For the skeptical appraisal, see Group promises dedicated VR "Neuronet," skepticism ensues and Is Neuronet A Scam?

If this thing isn't a real project, it should be. If you look at the success of Second Life, you can perhaps imagine where this could go with high-quality video data and user-side equipment to create a "virtual reality" experience.

This particular project may not be for real. But just wait 10 years or so. The applications won't be just game playing. This is the future of business teleconferencing (big bucks there), and eventually virtual gatherings of families and friends.

Tags: virtual reality, Neuronet

Recent technology news, 8/6/06

If your main source of information is the traditional media, then "technology news" is all about PCs, large-screen TVs, high-density DVDs, digital cameras, cell phones, MP3 players, and other electronic gadgets.

Stuff for consumers to gorge themselves on, in other words. Buy, buy, buy. Then buy some more, when the new models come out next year. Gotta keep the economy wheezing along, y'know.

Bletch.

In fact, there's plenty of interesting technology news that comes out every week. Only it's about stuff that won't be available on the shelves at Best Buy or Circuit City for maybe 10 years, or not oriented towards consumers at all. Stuff that may affect our lives just as significantly as cell phones, if only indirectly -- in much better health care and medicine, for example.

The following is just one week's worth of such technology news. I wish I had time to report on this kind of thing every week, but at least this is a sample.

Bayesian Mathematics Breathes Perception Into Robots

Robots have been around for decades. But you still can't buy one to prepare and serve dinner or help Grandma clean her house and do her wash, because existing robots are too stupid. They can "see" just fine, like a digital camera. But they have only the barest understanding of what they see. Robotic vision is a big problem. It's an "artificial intelligence" problem in which progress is slow, in spite of orders of magnitude improvement in computer speeds over the past several decades. What's holding things back isn't speed, but lack of adequate algorithms -- understanding of how humans comprehend their sensory inputs. The research discussed in the present article concerns a mathematical technique called "Bayesian reasoning" -- "a model for rational judgment when only uncertain and incomplete information is available."

AI Reaches the Golden Years

Here's another perspective on the AI problem. The occasion for the article is the 50^th anniversary of a famous workshop at Dartmouth where the term "artificial intelligence" was coined. You'd think that there should have been a significant amount of progress in 50 years, something commensurate with the progress in computers that makes the comparison between 1950s computers and contemporary ones like the relation of cuneiform tablets to high-density DVDs. Sadly, no. One of the main participants in the Dartmouth workshop, Marvin Minsky, said in 2003, "AI has been brain-dead since the 1970s." (I took an intoductory course in AI from Minsky a mere 40 years ago.) Perhaps that's too harsh a judgment, but the main problem is still "commonsense reasoning". However, we may be closer to the advent of "real" AI than to the founding of the subject. Stanley, the self-driving "autonomous vehicle" developed at the Stanford AI Lab which won the Darpa Challenge in 2005, may indicate that vehicles with at least some such capabilities may be on our roads in, oh, just another decade or two.

The Quest for the $1,000 Human Genome

AI will arrive eventually, but a bet I'd be a lot more confident making for the nearer term is the "$1000 genome" described in this article. The idea is that for the (arbitrarily chosen) price of a mere $1000 the genome of any particular individual could be sequenced. This means, in particular, that all of the potential disease-related genes of the individual would be known, along with genes that affect (for better or worse) individual reactions to therapeutic drugs. The first complete human genome sequence was produced by the government's Human Genome Project at the cost of only $3 billion (and almost simultaneously by the Celera Genomics company for only $300 million). This was essentially complete just 3 years ago, in 2003. Now it is estimated a complete sequence would cost only $10 million -- a factor of 30 improvement over the Celera work in only 3 years. But a cost reduction by a factor of 10,000 is still needed. How long will that take? 15 years, maybe, 20 at most? Together with the evolving knowledge of what genes make us more susceptible to cancer, Alzheimer's disease, diabetes, etc., this will make a really significant difference in our lives. Heck, it might be worth the price even at $10,000. After all, if you're diagnosed with cancer, you're probably talking medical bills above $100,000 or more. Knowing your actual genome might help dodge the cancer entirely. How much is your life worth?

Team Invents Fast, Flexible Computer Chips On Plastic

The range of small, hand-held electronic devices now available, like MP3 players, advanced cell phones, and digital cameras, is certainly impressive. Such devices now have the computational ability and memory capacity of computers that occupied entire rooms only 30 years ago -- a thousand times the memory capacity in some cases. This is an aspect of "ubiquitous computing", where computers disappear into the innards of common objects like coffee makers or telephones. But this trend has a lot farther to go, and it is due to be accelerated when electronic circuits no longer need to be fabricated on silicon chips but instead can be "printed" onto more convenient materials such as glass, plastic, or other flexible substances. The research described in this article is aimed at this goal.

A New BEC Magnetometer

A Bose-Einstein condensate is a state of matter that exists only at temperatures very near absolute zero. It forms when a large number of individual atoms (which must be "bosons") all fall to the lowest quantum state. BECs have been created in the laboratory only since 1995, and have remained largely of theoretical interest -- until now. The article reports on "the first application for Bose-Einstein condensates (BECs) outside the realm of atomic physics." In this case, the measurement of very small magnetic fields. There will certainly be other applications, such as in nanoscale optics and, quite possibly, in quantum computing.

Rice Scientists Unveil 'Nanoegg'

The article is about "asymmetric specks of matter whose striking optical properties can be harnessed for molecular imaging, medical diagnostics, chemical sensing and more." This is just one of a very large number of recent developments in the field of nanotechnology. Traditional media have carried many alarmist articles about the potential health dangers of nanoparticles that come in contact with human bodies. Much of that alarmism is rather beside the point, since (as here) the potential applications use nanotechnology safely encapsulated within high-tech medical or industrial devices that the pubic has no actual interaction with.

Connect the Quantum Dots

A quantum dot is another product of nanotechnology which already has significant applications, such as biomedical sensing devices and the blue lasers used in high-density DVD recording. They may very well be used, also, in quantum computing. A quantum dot is an aggregate of 100 to 100,000 atoms confined to a region with a diameter of 2 to 10 nanometers. Such an aggregate behaves like a single atom, which can assume discrete energy levels. This means that, among other things, quantum dots can absorb and emit distinct wavelengths of light, and hence are capable of flourescing with specific colors. The article here is about the use of quantum dots to replace conventional organic dyes in biomedical applications.

Add Nanotubes And Stir

Yet another, and well-publicized, product of nanotechnology is found in carbon nanotubes. A vast amount of research and development work is going on in this area. One application is nanoelectronics, where nanotubes are used to construct electronic devices at nanometer scales. They are also being used to make composite materials by mixing them with organic polymers. Such composites can have exceptional strength, toughness and electrical conductivity. The problem is that the composites have to be made with great care. The research here is about using just the right amount of force in mixing the composite in order to achieve the desired properties.

----------------------------------

Tags: technology, nanotechnology, robotics, artificial intelligence, genome sequencing, flexible electronics, quantum dots, carbon nanotubes, Bose-Einstein condensates

Is the singularity near?

Just about two months ago, on a (typically) gorgeous spring day in May, I found myself back on the Stanford campus to attend a day's worth of talks by a distinguished panel of speakers at a symposium entitled, modestly enough, The Singularity Summit.

What on Earth is that about?

Here's what the Introduction to the day's program says:

The singularity scenario is a hypothesized "event horizon" in human technological development beyond which our models of the future cease to give reliable answers. The hypothesis rests on the creation of "superintelligence": any future intellect, possibly strong artificial intelligence, that tremendously eclipses the best human minds in practically every field, including scientific creativity, general wisdom, and social skills.

A little more detail: here.


In the past few years, Ray Kurzweil (at right), the lead speaker at the "summit", has been perhaps the most visible exponent (that's kind of a pun) of the singularity idea, through a series of books, beginning with (as editor) The Age of Intelligent Machines (1990), followed by (as author) The Age of Spiritual Machines (1999) and The Singularity is Near (2005).

Kurzweil writes:

What, then, is the singularity? It's a future period during which the pace of technological change will be so rapid, its impact so deep, that human life will be irreversibly transformed. Although neither utopian or dystopian, this epoch will transform the concepts that we rely on to give meaning to our lives, from our business models to the cycle of human life, including death itself. Understanding the singularity will alter our perspective on the significance of our past and the ramifications for our future. To truly understand it inherently changes one's view of life in general and one's own particular life.

But Kurzweil is hardly the originator of the idea. It has been foreshadowed by people such as the mathematician and major contributor to early computer science John von Neumann (1950s) and statistician I. J. Good.

A more organized presentation of the singularity idea came from mathematician and (naturally) science fiction writer Vernor Vinge (also here) in the 1980s, culminating in his 1993 paper The Coming Technological Singularity. (Updated a little in 2003 here.) At the beginning of his 1993 paper Vinge wrote "Within thirty years [by 2023], we will have the technological means to create superhuman intelligence. Shortly after, the human era will be ended."

Of course, there's plenty of skepticism that something like this singularity will actually occur in the foreseeable future, let alone by 2023. Many people are skeptical (to say the least) not only of the possibility, but also of allowing such a thing to happen even if possible. I won't go into that here. But one of the less hostile skeptics is Douglas Hofstadter, who took an interest in the idea around the time that Kurzweil's Spiritual Machines book appeared in 1999. Shortly thereafter he came out with a paper discussing his philosophical reservations, Moore's Law, Artificial Evolution, and the Fate of Humanity.

Hofstadter was the second speaker at the Summit and his presentation was (in my opinion) as interesting as Kurzweil's. The third speaker whose presentation I found valuable (i. e., added much to the discussion) was Sebastian Thrun, current director of the fabled Stanford AI Lab. Thrun mostly described his laboratory's work on Stanley, the winning autonomous robotic vehicle, of the 2005 DARPA "Grand Challenge". He had little to say about the Singularity per se, and instead used Stanley to represent, in his humble (and possibly correct) opinion, the current state of the art in artificial intelligence.

I was attending the Summit very much with the idea in mind of writing about it here. But after all was said and done, I couldn't think of much very useful to say. In a private discussion group I summarized my impressions thusly:

Maybe Hofstadter had the right attitude. ... I think his best point was that this sort of thing needs much more rigorous discussion is very apt. Science fictional blue sky thinking is all well and good, but it will not be what enables this kind of thing to actually be implemented (and, one hopes, in a benign way).

Kurzweil's ideas are interesting, once you get beyond just extrapolation of exponential growth curves. And he may well be the most rigorous thinker (as least on the relevant topics) of all the speakers, excepting the SAIL guy (Thrun). (Drexler appeared very bored, and didn't even bother to come back after lunch.) But so far I can't see much "science" in these ideas...

Oh, yeah. I'd also like to have had someone talk rigorously and concretely about near-term prospects for "life extension". (Prospects more than, say, 50 years out won't help most of us, much.) But even Kurzweil didn't get into this. Looks like most non-biologists don't care to explore the subject much beyond the high-level generalities.

I guess the point to be made here is that this discussion is all about the possible future of artificial intelligence, and not its current state of the art. And as Yogi Berra said, prediction is very difficult, especially about the future. It's possible to expostulate easily and at great length about whether the Singularity that Kurzweil and others foresee is good or bad. That, after all, is the province of mere philosophers and op ed writers. Trying to assess what science and technology is actually capable of doing in the near future is much harder, to say nothing of the farther future.

But it's irresistably interesting, and writing a lot more about this is on my to-do list.

------------------------------------

Additional references:

Singularity Summit Coverage - includes links to some press and blog articles, plus audio of all presentations and powerpoints of most

Singularity Institute for Artificial Intelligence - one of the Summit's primary sponsors

Singularity Summit LIVE! - a series of live blog posts at the Responsible Nanotechnology site (this item is only the first)

Singularity Summit Opens - first of another series of blog posts (by Kurzweil's publicist)

The Singularity Summit - a Daily Kos diary by one Summit attendee

Technological Singularity - Wikipedia article

Stanford conference ponders a brave new world with machines more powerful than their creators - San Francisco Chronicle article (before the event)

The age of Ray Kurzweil - friendly bio in Kurzweil's hometown paper

KurzweilAI.net - huge collection of essays and news articles about AI, the Singularity, and related topics

The Singularity is Near - promotional site for Kurzweil's book

Selected Annotated Bibliography of Douglas R. Hofstadter - in case you aren't familiar with his writing

Critical Discussion of Vinge's Singularity Concept - collection of 13 essays

The Singularity - brief list of references

------------------------------------

Tags: the singularity, artificial intelligence, Singularity Summit, Vernor Vinge, Raymond Kurzweil, Douglas Hofstadter

Surveillance technology

Let's talk about surveillance technology. You know, stuff like wiretapping, data mining, spyware, pervasive surveillance cameras. That sort of thing.

What's raised the issue in my mind is stories like this:

If You're Cheating on Your Taxes...

But a few states, such as Texas, are building more sophisticated data mining programs that will predict taxpayer behavior, much as credit-card companies try to estimate how much consumers will spend over the course of a year. "The capability is there to figure out which taxpayers have the highest probability of becoming noncompliant," says Steven E. Taylor, director of the revenue and compliance team of the data warehousing firm Teradata, a unit of NCR (NCR). Iowa, Massachusetts, and Virginia are also in the data mining vanguard.

MANY DATABASES. Typically, each state revenue agency will work with one data management company and its subcontractors, drawing from a list that includes Teradata, Revenue Solutions of Pembroke, Mass., and CGI Group of Montreal. The data miners can construct powerful programs that assign each taxpayer the equivalent of a credit score, flagging those who should be targeted for an audit. They can project who is likely to file on time, who won't pay until they get a visit from a collection agent, and even who is likely to declare bankruptcy before paying their taxes.

(Also available here).

Doesn't sound all that unreasonable -- if everyone pays their "fair share", honest people will not have to make up the difference by paying more. But is it OK for local, state, and federal governments to scrutinize every detail of our lives and our lifestyles in order to catch "cheaters", just because, with the latest current and future technologies, they can?

Or how about this:

Wi-Fi plan stirs Big Brother concerns

Privacy advocates are raising concerns about Google Inc.'s plans to cover San Francisco with free wireless Internet access, calling the company's proposal to track users' locations a potential gold mine of information for law enforcement and private litigators.

The Mountain View search engine intends to use the geographic data to match users with advertising so that they would see marketing messages from neighborhood businesses such as pizza parlors, cafes and book stores.

Google plans to use technology that would allow it to track users' whereabouts within a few hundred feet. The company said in its bid that it would retain the data for up to 180 days before deleting it, as part of an effort to "maintain the Google Wi-Fi network and deliver the best possible service."

Privacy advocates fear the information could by used by government officials to place users under surveillance and are more generally concerned that this new power raises the specter of "Big Brother" run amok.

"The greatest concern is that once you have that treasure trove of information, will people start to come looking for it?" said Kurt Opsahl, staff attorney for the Electronic Frontier Foundation, a privacy watchdog group.

Is this relatively harmless, if in return for being tracked we get free WiFi access and the "benefit" of (possibly) relevant advertising from businesses just around the corner from wherever we happen to be? Should we just not worry that big companies like Google and their advertisers -- to say nothing about governments at all levels who will be looking over the businesses' shoulders -- can know, within a few hundred feet, where we are at all times our computers or cell phones are turned on? Is it just a paranoid fantasy that some day governments will have real-time access to such Internet traffic, so that we find ourselves arrested at Fisherman's Wharf because some government snoop is on the lookout for a "terrorist" who just happens to have the same name as ours?

What makes such things so scary is the liklihood that today, right now, or if not now, very soon, the government can and does watch every last bit of traffic on the Internet, with the cheerful cooperation of telephone and cellular companies. At least, that's the implication of what I'm reading, like this:

Whistle-Blower Outs NSA Spy Room

According to a statement released by Klein's attorney, an NSA agent showed up at the San Francisco switching center in 2002 to interview a management-level technician for a special job. In January 2003, Klein observed a new room being built adjacent to the room housing AT&T's #4ESS switching equipment, which is responsible for routing long distance and international calls.

"I learned that the person whom the NSA interviewed for the secret job was the person working to install equipment in this room," Klein wrote. "The regular technician work force was not allowed in the room."

Klein's job eventually included connecting internet circuits to a splitting cabinet that led to the secret room. During the course of that work, he learned from a co-worker that similar cabinets were being installed in other cities, including Seattle, San Jose, Los Angeles and San Diego.

More on this story: here, here, here.

Of course, since the New York Times last December broke the story of illegal NSA wiretapping of U. S. citizens, vague information about this has been widely reported. But the latest revelations are rather alarming. There's a lot more to this than ordinary, old-fashioned wiretaps of a few suspected "terrorists". It's looking more and more as though what this is about is the systematic surveillance of (potentially) every last telephone and Internet communication that the NSA can get its hands on.

And that is just about everything that flows over communications lines in U. S. territory. The technology to do this exists and is very real, as this long and technical diary at Daily Kos explains: All About NSA's and AT&T's Big Brother Machine, the Narus 6400.

I won't quote from there, since it's long and geeky. But if you have any doubts about the technical ability of the government to perform such surveillance, maybe you should read it. The bottom line is that the government can monitor, in real time, every phone call, every email, every file download, every Web page access, and every electronic financial transaction that involves anybody located in the U. S.

And that's only the beginning. The Internet is no longer just for computers. All sorts of everyday things are, or before long will be, connecting to the Internet. Many new cars come equipped with "black boxes", like those on aircraft, to record speed, braking action, stops and starts, etc. (See here and here.) They could record location as well, if a GPS device (or similar technology based on the cellular phone system) is on board. With this, and with wireless Internet communications capability (based on WiFi and eventually WiMax), the NSA could monitor this same information -- in real time.

Are you driving "suspiciously" in the neighborhood of some critical infrastructure (even if you're not aware of it)? Have you ever, perhaps, visited a part of town that could get you into blackmail problems? Better watch out, because Big Brother will know.

It would be reassuring to think that such concerns were merely paranoid fantasies. But how much should we, as individuals and as citizens, stake our future on a belief that the intentions of our government are entirely benign? Perhaps we should at least understand what they're capable of. Of course, they will never disclose their full capabilities, let alone what capabilities they're actually using. That stuff's highly Top Secret. It wouldn't do to allow the "terrorists" to find out, now, would it?

But of course, none of us are terrorists, right? And we never have cheated on taxes, and never will. So what do we have to be afraid of?

Well, simple mistakes for one thing. Many people, including Senators and Congressmen, have already not been allowed to board an airplane or even been detained at an airport, simply because some name that happened to be the same as theirs was on a list of suspected "terrorists".

But it's worse than that. Today, everybody who has a bank account, a credit card, or a commercial loan has a credit rating, whether they realize it or not. In the near future, if not already, everybody could also have a loyalty rating. Government computers at the NSA, and elsewhere, will be computing this score on a profusion of data collected from existing business and government databases and all the data flowing over the Internet itself. That's what "data mining" is all about. There are plenty of excuses they can come up with for doing this -- watching for "terrorists", tracking down fugitives and deadbeat dads and people with unpaid traffic tickets, screening for potential tax cheats, seeking out money launderers and drug dealers and pedofiles. For just about every law on the books, there's a good excuse for snooping to find anyone who's breaking it.

The point is, governments will be doing this data mining. And the "profiles" compiled about you by all these government computers will be passed around and shared with other computers as long as you're alive. In particular, this "loyalty" rating. Even if you're not a "terrorist" or "subversive'.

Who knows what might go into the ratings? If the government knows you like Christian music, maybe you'll get points added to your rating. But if you prefer music genres that are more in favor with other groups (Islamic music?), you'll lose points on your rating. No discrimination intended, mind you. It's just that, well, church-going Christians who give more than $1000 a year to the church are known statistically to need less thorough scrutiny than your average potential "terrorist". Same with your political preferences. If you're a known supporter of the party in power, your rating goes up. If you support the opposition party, it goes down. Again, nothing personal, you know. The statistics just say that contributors to the party in power are more trustworthy than those who hang out with people associated with the other party.

The govenment has the ability today to compute such "loyalty" ratings. Can you say you're confident that the government would never, ever, use the available technology to do it? (If so -- and the government will probably know this -- your rating will certainly be higher than if not.)

And here's one more thing. Even if we personally are as pure and loyal as the driven snow, what's to stop a government overly infatuated with holding onto power from snooping into the lives and opinions and finances of any politician or office holder belonging to an opposition party at any level of government? Could be helpful to know about any and all skeletons in their closets, no?

Just in case I haven't made you sufficiently concerned, here are a few more news items to think about:

Tax tech pays dividends

From data mining to predictive modeling techniques, information technology is an increasingly important tool for state governments that want to improve their tax handling performance by speeding tax processing and boosting tax collections. Such technologies are components of a financial enterprise in which one tax application links to another to produce an overview of individual portfolios.

Pay attention to the part about applications linking together. These same applications can feed into another set that generates a "loyalty" rating. Want to bet it will never happen?

IBM Announces Advanced Analytics to Help Revenue Agencies Optimize Tax Auditing Process

IBM today announced a Tax Audit and Compliance Solution that uses advanced analytics to help revenue agencies zero in on questionable tax returns. The solution is currently being used by U.S. state revenue agencies as well as international government agencies.

By providing a more scientific, data-driven approach to tax audit case selection, the Tax Audit and Compliance Solution goes well beyond existing methods of tracking tax compliance, which range from manually matching internal data with lists of taxpayers to using data warehouse and query tools. By identifying possible compliance problems at the time a tax return is filed, the solution, which is capable of data mining thousands of tax returns in seconds, can potentially help save years of tracking, investigation and collection costs.

Developed by IBM scientists, the solution uses algorithmic data mining and predictive modeling to compare individual taxpayer behaviors to those of similar taxpayers. The resulting analysis helps agencies to detect new tax evasion methods and provides auditors with lists of potentially fraudulent tax returns.

So, this kind of sophisticated data mining software is available for purchase off the shelf. Is there any doubt it will also be used by credit rating "services", marketing organizations, ... and political parties? In fact, such tools are already used to target voters in fund raising and GOTV operations. The party in power definitely knows who its friends are, and as for everyone else... points can be taken off their loyalty rating. The government need not fear disclosure of how it computes the loyalty rating -- that's all highly classified, of course.

Policing Trade to Nab Terrorists

In the first formal effort to combat these techniques, the U.S. in January teamed up with the governments of Argentina, Brazil and Paraguay to create "Trade Transparency Units" that allow the countries to share detailed information about each others' import and export transactions. Armed with a U.S.-designed data-mining computer program, investigators sift through the information looking for anomalies in commerce that could indicate terrorist financing or other criminal activity.

The same software, of course, can be used internally to the U. S. to scrutinize any financial transaction at all -- buying a house or car, trading stocks or options, contributing to political organizations... Given that such transactions do now, or will, take place over the Internet, cooperation from banks or any of the parties involved isn't even needed.

OK, enough of this for now. There are plenty of other things to fret about, some of them problems today, some of them for the not-too-distant future. Such as smart-card driver's licenses, passports, and ID cards that can get on the Internet themselves and rat on the people who carry them. Or future personal computers that come from the factory with government-mandated spyware embedded in the hardware. (Media companies and organizations like the RIAA are already clamoring for this sort of thing in order to "protect their intellectual property".) Together with Internet access, the NSA can then read any of your files, watch your every keystroke, and learn all your passwords and encryption keys.

All without any search warrant or court overview required, of course!

Worried yet?

Tags: surveillance, privacy, NSA, civil liberties, data mining

Next generation computer chip lithography

In the competitive world of semiconductor technology, the name of the game is to squeeze as many transistors as possible into every chip. For random-access memory chips this translates directly into more memory on every chip. For central processor chips this allows faster speeds for several reasons. More transistors per chip makes it possible to add more cache memory to the processor chip for faster memory access. In addition, communication between circuit elements is faster the closer together the elements are.

The main way of describing how densely transistors are packed is in terms of "feature size", which is roughly the size of the smallest geometric feature that can be produced. In the current state of the art, with the latest Intel Pentium processor chips, for example, the smallest feature size is 65 nanometers (nm).

In the process of semiconductor fabrication features are created through the process of photolithography. In essence, this amounts to printing the circuits containing all transistors and their interconnections on a semiconductor material in much the same way as photographic images have long been printed on paper.

The main problem with printing very small features using light is the phenomenon of diffraction, which makes it very difficult to print features that are substantially smaller than the wavelength of light. The shortest wavelength of visible light is about 400 nm (violet), which is totally inadequate for feature sizes such as 65 nm.

The current state of the art uses ultraviolet light with wavelengthts of 248 or 193 nm ("deep ultraviolet"). This is just barely able to handle feature sizes of 65 nm by using various tricks of "nanolithography", such as "liquid immersion". It is expected that by 2009 this process will be capable of creating feature sizes of 45 nm. Intel has already demonstrated an experimental system that can do this.

But that's not going to be good enough beyond 2009. New techniques of producing smaller feature sizes around 10 to 30 nm (just a few hundred atoms in width) are now the subjects of active research. Intel is investigating techniques of extreme ultraviolet lithography which, the company hopes, will achieve feature size objectives, using ultraviolet light with a wavelength of only 13.5 nm, by 2009. (See here.)

That approach, however, is not the only one possible. Other laboratories are claiming features sizes of 26 nm already, using other techniques:

Breakthrough Computer Chip Lithography Method Developed at RIT

Leading a team of engineering students, Bruce Smith, RIT professor of microelectronic engineering and director of the Center for Nanolithography Research in the Kate Gleason College of Engineering, developed a method—known as evanescent wave lithography, or EWL—capable of optically imaging the smallest-ever semiconductor device geometry. Yongfa Fan, a doctoral student in RIT’s microsystems engineering Ph.D. program, accomplished imaging rendered to 26 nanometers —a size previously possible only via extreme ultraviolet wavelength, Smith says. By capturing images that are beyond the limits of classical physics, the breakthrough has allowed resolution to smaller than one-twentieth the wavelength of visible light, he adds.

-----------------------------

Tags: semiconductor fabrication, photolithography, nanolithography, extreme ultraviolet lithography, evanescent wave lithography

Videoconferencing on steroids

It's hard to say how long the idea of a "telephone" that could carry visual images as well as voices has been around, but experimental implementations of videophones have existed since the early 1960s. Nevertheless, even today ordinary people are unable to buy such a thing that works over existing telephone lines.

But we're getting a lot closer. Elaborate (and expensive) implementations of a slight generalization of the idea, in the form of videoconferencing, have been done since the late 1960s. The idea here is to tie together TV cameras and video screens over some sort of connection with sufficiently high bandwidth that make it possible to hold virtual meetings among a number of people which approximate the experience of being in the same room. The approximations are getting better and better, and are available to any group -- mainly businesses -- that can afford them.

Thanks to the Internet and cheap personal computers -- using inexpensive video cams and broadband network connections -- the functional equivalents of videophones are finally available to the general public. Even if such video chats are not yet quite as simple to use as ordinary telephones. But with the rapid spread of voice over IP just about to happen, we're almost there.

The experience of using such a system is still not quite like being in the same room with others. For instance, one has to look into a camera to give the other person the illusion of eye contact, and in the process losing one's own subjective feeling of eye contact. And there remain problems of latency and lack of smooth motion in less expensive systems.

But things are getting pretty good at the high end:

Together across a continent: Bridging time and place with 'Shared Spaces'

"Modern videoconferencing hasn't worked well as it doesn't allow you to interrupt one another and has never managed to support the quality of interaction that people experience in real life. We wanted to change that," says John Roston, director of Instructional Multimedia Services at McGill University.

"Our technology provides a life-size, high-definition view on a large panoramic screen, which gives users the impression that they're talking to people in the same room with a window between them," adds Professor Jeremy Cooperstock of the Department of Electrical and Computer Engineering.

It's interesting to think about how such systems will impact our lives when they become generally available. For instance, it will become posible to have regular virtual visits with friends and relatives anywhere in the world (subject to time zone differences). We will be able to schedule an evening (or morning, or afternoon) with people we like who live across a continent or an ocean as easily as we do now with friends in the next town. Even easier, in one important respect, as there will be no time wasted in travel. (More bad news for the airlines.)

Will we finally shuck off our thralldom to the boob tube if real, live people we like are as accessible as the latest TV sitcom characters? (The advertising industry will be further devastated.) Or, on the downside, will we discover that our friends and relatives actually aren't as interesting as fictional characters on TV?

In another direction, what will happen with colleges and universities when it becomes possible to "sell" individual lectures by expert faculty members to students anywhere in the world? Will institutions of higher education prosper, or will they flounder as their best faculty go freelance and earn a lot more without the overhead expenses of physical campuses and administrative staffs? Will domestic educators and educational institutions suffer in the same way as other service businesses when their jobs are outsourced to India? What new mechanisms will evolve to certify that students in such a very "open university" have mastered the contents of a particular course?

It's going to be really interesting to watch what actually develops as basic -- and eventually more advanced -- forms of teleimmersion become widespread.

----------------------------------

Update, December 18, 2005 -- Story about a presently-available videoconferencing system: Videoconference system creates boardroom illusion

Tags: videoconference, video chat, teleimmersion

A new technological dark age?

The general asumption most people have is that technological innovation is not only progressing steadily, but that it's actually accelerating. However, one expert's careful study suggests that may not be so:

Entering a dark age of innovation

[Jonathan] Huebner is confident of his facts. He has long been struck by the fact that promised advances were not appearing as quickly as predicted. "I wondered if there was a reason for this," he says. "Perhaps there is a limit to what technology can achieve."

In an effort to find out, he plotted major innovations and scientific advances over time compared to world population, using the 7200 key innovations listed in a recently published book, The History of Science and Technology (Houghton Mifflin, 2004). The results surprised him.

This outlook it quite possibly too pessimistic. But an awful lot depends on future developments whose arrival is very difficult to predict, and which may not even occur any time soon. Among things that probably must happen in order to sustain technological advance are

• Finding cheap, abundant sources of renewable energy to replace ever more expensive fossil fuels.
• Inventing entirely new computing technologies such as "quantum computers" to take over when semiconductor-based iterative procedural computers reach their apparent limits in 10 or 15 years.
• Averting looming and potentially quite disruptive environmental problems such as global warming and dwindling fresh water supplies.
• Discovering a biotech means to prevent a global pandemic of an infectious disease such as avian flu or SARS, which could easily lead to a long-lasting economic depression.
  

That is all without even considering obvious global political problems, such as religious and cultural clashes, the proliferation of nuclear, chemical, and biological weapons, and conflicts arising from the vast disparity in wealth between the least and most affluent countries. It's hard to imagine any technological fix for problems like these.

And technology may not bail us out in areas where it can potentially be effective, if indeed innovation is about to slow significantly because most of the low-hanging fruit has already been picked.

It may well be that other goals such as sending humans back to the Moon and to Mars will have little useful value in dealing with any of these problems. Certainly we're not going to be able to move large numbers of Earth's population off the planet before technologies we can't even really imagine at present are developed. Space travel is probably not anywhere near our top priority right now. (Has the still unfinished space station been, or will it ever be, good for much of anything at all?)

What we're really going to need most to develop is more like a vast, renewable source of good sense and good luck.

Controlled nuclear fusion

When will we ever have thermonuclear ("fusion") electric power? For something like 50 years the answer has been "in about 30 years". A large part of the problem is political. Here's one assessment of the situation:

Nuclear fusion: power to the people?

But what's changed since JET was built to make it any more likely that fusion won't remain forever 30 years in the future? The materials, says Carpenter, and the computers. From helium-cooled superconducting magnets to tungsten chamber walls to supercomputers that can calculate how the plasma will behave far more accurately and quickly than ever before, the pieces are all there, waiting for the politicians to sign off the cheques and shake hands.

Biocomputation

This panel discussion at Edge, among J. Craig Venter, Ray Kurzweil, and Rodney Brooks, is a must-read.

It's too long to summarize concisely. But the gist is that computing and biotechnology are the leading technologies of the past decade, and almost certainly of the next several decades. While they are not exactly merging (yet -- that's another decade or two away), they are starting to overlap significantly at the cutting edge.

Venter leads off with a discussion of how rapid progress in genomics will affect developments in cancer, antibiotics and antivirals, and synthetic biology (synthesizing living organisms). Kurzweil speculates about how radical advances in treatments for infectious and systemic diseases may occur sooner than most expect. Brooks explains why he thinks we need to understand much more about biology before we can make further progress in such fields as robotics and artificial life.

After the introductory remarks, Kurzweil and Brooks debate how rapidly dramatic progress in the science and technology of biocomputation should be expected to occur. The disagreements revolve around questions about how long it will take to really understand extremely complex systems such as individual cells and the human brain.

The discussion as a whole is a good introduction to the science of the 21^st century.