Quantum computing: revenge of the analog

I'd assumed we were still 10 years away from a useful quantum computer. A Canadian company is now advertising one: Scientific American.com: First "Commercial" Quantum Computer Solves Sudoku Puzzles.

The experts believe this particular approach won't scale, but I'm stunned that they're already up to 16 bits. It also doesn't work do encryption tasks, so international finance can keep working. It apparently works rather like the analog computers my grad school adviser once played with. Don will be amused to learn that his analog skills are again on the cutting edge.

So, how long until we do applied counterfactual computing? Does anyone still think we can predict twenty years ahead?

The peculiar scarcity of books on quantum entanglement

I've been blogging periodically about the weirdness of a universe where quantum entanglement is real, but it was a somewhat supercilious essay in Wired magazine that convinced me I needed to read a real book ..

Wired 15.02: What We Don't Know

How do entangled particles communicate?

... In 1997, scientists separated a pair of entangled photons by shooting them through fiber-optic cables to two villages 6 miles apart. Tipping one into a particular quantum state forced the other into the opposite state less than five-trillionths of a second later, or nearly 7 million times faster than light could travel between the two [jf: probably instantaneously]. Of course, according to relativity, nothing travels faster than the speed of light - not even information between particles.

Even the best theories to explain how entanglement gets around this problem seem preposterous. One, for example, speculates that signals are shot back through time. Ultimately, the answer is bound to be unnerving: According to a famous doctrine called Bell’s Inequality, for entanglement to square with relativity, either we have no free will or reality is an illusion. Some choice.
- Lucas Graves, New York City-based writer

I was sure I'd find a lot of books on the topic, but the best I could do was 'Schrodinger's Kittens' -- written about 13 years ago! I did find some a more recent text, but it belonged to what a Wikipedia article calls Dirac's "shut up and calculate" school [1]. The other relatively recent texts I found were either fluffy or preferred to deal with familiar topics like modern cosmology.

So what gives? Is this such a scary topic that almost no-one dares to explore what it means?

I'm write some more as I work my way through the Kittens ...

[1] Easy for Dirac to say -- he died before we were doing quantum teleportation, quantum encryption, and irrefutable entanglements.

Update 2/15/07: After finishing the last chapter of Schrodinger's Kittens I understand why Graves setup the choice of 'reality is not what we think' versus 'no free will'. The 'transactional model' preserves our familiar "reality" of time and relativity, but the handshakes between past and future seem to constrain the future. In the enhanced (Wheeler) slit model, for example, the photon assumes its wave or particle behavior as it crosses the slit based on a handshake with the future absorbed photon. The nature of the future absorbed photon however, is based upon an observer choosing whether or not to "drop the screen". Since the photon adopts its configuration before the screen is dropped, however, the observer cannot really be choosing whether or not to drop the screen. The choice must be made in a way that's consistent with the form of the photon. Gribbin tries to dodge this trap with some handwaving about micro vs. macro causality, but it's obvious his heart isn't in it.

Hmmm. I'm beginning to see why Feynman warned us about 'going down the [quantum] drain', and why there are so many books on cosmology and so few on Bell's theorem. Next thing you know I'll be ready to start believing that our universe is both simultaneously vast and unbelievably small, that all things that will happen have happened, and that time's arrow really is an illusion ...

There is an escape clause. Gribbin's framed the transactional model as being dependent upon a closed universe. We appear to live in an open (perhaps excessively open) universe (but see below). In an open universe, might we get a true arrow of time and the possibility of choice? Maybe one day, if we ever "understand it all", we'll learn that you can determine whether a universe is open or closed by testing for action-at-a-distance.

Update 3/4/07: I'm reading through Gribbin from page one, and since I'd read the last chapter first I know to watch for references to a closed universe. I've found a few, it seems that more than a few of the foundations of QM, and even QED, do assume a closed universe. I'd thought that the universe is now thought to be "open", so I wrote Professor Gribbin asking if he'd written any updates. He graciously replied:

... It is entirely possible for the Universe to be closed but with accelerating expansion! All we know for sure is it is indistinguishably close to flat, and it is probably on the closed side of flat, pushed there by inflation.

Which reminds me of my old post about the respiratory rate of the universe ...

I'll just keep on reading, reading, reading ...

Update 5/13/07: My Amazon review of Gribbin's book. Five stars, despite being 12 years old.

The end of SARS and the use of synthetic pathogens to combat epidemics

You know - SARS - the plague that was going to destroy civilization. It went away. Puzzled the heck out of me. It was supposed to recur every year or two, but it hasn't. Nobody seemed to be curious about this. Was a giant conspiracy at work?

Well, no. A quick pubmed search showed that lots of epidemiologists have been quietly puzzling about what happened, especially in China. I liked this one:

A double epidemic model for the SARS propagation (9/2003).

... We find that, in order to reconcile the existing data and the spread of the disease, it is convenient to suggest that a first milder outbreak protected against the SARS...

I particularly like it because I wrote:

I still don't understand why all hell didn't break loose then. My best guess is that there were multiple strains of SARS circulating simultaneously, and an innocuous one spread faster -- immunizing the susceptibles in advance of the killer strain.

Ok, so I put my theory out more than a year after Ng et al published a full model supporting it! I gotta work on my timing. They must have started work on their model very shortly after the epidemic had started to fizzle. I didn't blog on this thought, but my theory back then (2004, not 2003!) was the Canadian nurses got so sick because they were so good at isolation -- they prevented exposure to the benign, immunizing, coronavirus and thus suffered the full impact of the malign virus.

All of which lead to some random observations and questions:

1. This is a fascinating story that ought to appear in a popular magazine, or at least in The Economist or Scientific American. I don't recall seeing anything. There's a curious "chaotic" aspect to what gets written when. I wonder if blogs will change any of that, or if they simply amplify the current fads.

2. As a non-practicing physician who works on clinical knowledge representation I often think about the limits of the mental models I once used to care for patients. Back in the day we were taught to think of 'one infection, one disease'.

Are medical students still taught to think that way, or are physicians now taught that illness (or its absence) may be the result of a number of interacting simultaneous infections (and of course susceptibilities, treatments, phases of the moon, etc, etc)? Of course I'm not sure what one would do with such knowledge! Still, it does help make one's predictions more modest.

3. This suggests a radical, but not entirely novel, approach to a future serious epidemic. Create a contagious synthetic pathogen that's relatively benign, but induces immunity to the major pathogen -- and spread it actively. I say not entirely novel, because this is how Polio was suppressed. The oral vaccine was an active contagious pathogen that was excreted in stool. It immunized a vast number of persons -- but some became sick, disabled, or dead. When Polio was less of a threat we switched to a non-pathogenic inoculation. The difference is the successful Polio strategy was probably unintentional (I suspect some people understood even in the 1950s), but in the future we'd be deliberately exposing an entire population to an immunogenic pathogen that would almost certainly harm many people.

Oh, about that conspiracy. Somewhere in central China in 2003, a brilliant scientists realizes that she can save the world by unleashing a synthetic coronavirus she's been developing in a top-secret bioweapon facility ... The novel almost writes itself ...

2/15/07: Emily points out that this is rather like fighting a fire by setting fires -- backburning I think it's called. My son has a book on it called 'Hotshots'. A useful analogy.

Walls around the world - Dyer

Gwynne Dyer surveys the worldwide industry of building walls between nations. There are many more of these than I'd imagined. Dyer makes a persuasive case that walls and fences are the future, an inevitable consequence of mobility and population growth.

It's a unique essay, well worth a quick read.

The dog: symbiote or parasite?

I've long thought of the Dog as a brilliantly successful cuckoo-like brood parasite. This only increases my admiration for the species. One can do much worse than develop a remora-like attachment to the most dangerous and dominant large predator the planet has ever known. (Ok, so our reign will be short. Not every species can be a shark.)

A biologist friend, however, scoffs at this. He claims the dog is obviously a symbiote. Well, to be honest, I know the boundary between symbiote and parasite is fungible. Dogs are great at cleaning up rotting meat, which probably has some health advantages. They may help geek-drones and women survive around alpha-males -- though they are prone to run away in a fight (they're not stupid). I'm sure they have other advantages ...

Whatever. We've been together long enough that we've probably altered each other's evolution. Jon Katz, another dog lover, weighs in on the debate:

The real reason we love dogs. - By Jon Katz - Slate Magazine

... Archer suggests, "consider the possibility that pets are, in evolutionary terms, manipulating human responses, that they are the equivalent of social parasites." Social parasites inject themselves into the social systems of other species and thrive there. Dogs are masters at that. They show a range of emotions—love, anxiety, curiosity—and thus trick us into thinking they possess the full range of human feelings.

They dance with joy when we come home, put their heads on our knees and stare longingly into our eyes. Ah, we think, at last, the love and loyalty we so richly deserve and so rarely receive. Over thousands of years of living with humans, dogs have become wily and transfixing sidekicks with the particularly appealing characteristic of being unable to speak. We are therefore free to fill in the blanks with what we need to hear. (What the dog may really be telling us, much of the time, is, "Feed me.")

As Archer dryly puts it, "Continuing features of the interaction with the pet prove satisfying for the owner."

It's a good deal for the pets, too, since we respond by spending lavishly on organic treats and high-quality health care.

Psychologist Brian Hare of Harvard has also studied the human-animal bond and reports that dogs are astonishingly skilled at reading humans' patterns of social behavior, especially behaviors related to food and care. They figure out our moods and what makes us happy, what moves us. Then they act accordingly, and we tell ourselves that they're crazy about us.

"It appears that dogs have evolved specialized skills for reading human social and communicative behavior," Hare concludes, which is why dogs live so much better than moles.

These are interesting theories. Raccoons and squirrels don't show recognizable human emotions, nor do they trigger our nurturing ("She's my baby") impulses. So, they don't (usually) move into our houses, get their photos taken with Santa, or even get names. Thousands of rescue workers aren't standing by to move them lovingly from one home to another.

If the dog's love is just an evolutionary trick, does that diminish it? I don't think so. Dogs have figured out how to insinuate themselves into human society in ways that benefit us both. We get affection and attention. They get the same, plus food, shelter, and protection. To grasp this exchange doesn't trivialize our love, it explains it.

I'm enveloped by dog love, myself. Izzy, a border collie who spent the first four years of his life running along a small square of fencing on a nearby farm, is lying under my desk at the moment, his head resting on my boot.

Rose, my working dog, is curled into a tight ball in the crate to my left. Emma, the newcomer who spent six years inside the same fence as Izzy, prefers the newly re-upholstered antique chair. Plagued with health problems, she likes to be near the wood stove in the winter.

When I stir to make tea, answer the door, or stretch my legs, all three dogs move with me. I see them peering out from behind the kitchen table or pantry door, awaiting instructions, as border collies do. If I return to the computer, they resume their previous positions, with stealth and agility. If I analyzed it coldly, I would admit that they're probably alert to see if an outdoor romp is in the offing, or some sheepherding, or some beef jerky. But I'd rather think they can't bear to let me out of their sight.

And the answer is ... symbiote. I happen to believe, with a bit of evidence, that humans have almost uniquely flaky brains (Temple Grandin mentions this in one of her books as well). We have lots of very significant relatively recent evolutionary hacks in our wetware, and we know that hacks produce bugginess and instability. We have a high rate of major and minor defects -- as near as we can tell we're much worse off than any other animal. It's the price we've paid for sentience. A social animal with a buggy brain has a major need for psychic support. The dog is a first-rate psychic crutch. That's a heck of a value proposition. All hail the Dog!

Teraflop 2007 and 1996

[I originally titled this 2007 and 1986. I think my mind wouldn't process that it's been 11 years since 1996.]

Now the "ancient time" comparisons are well within my memory range ...

BBC NEWS | Technology | Teraflop chip hints at the future

A chip with 80 processing cores and capable of more than a trillion calculations per second (teraflop) has been unveiled by Intel.... a piece of silicon no bigger than a fingernail ...

... The first time teraflop performance was achieved was 11 years on the ASCI Red Supercomputer built by Intel for the Sandia National Laboratory.

That machine took up more than 2,000 square feet, was powered by almost 10,000 Pentium Pro processors, and consumed more than 500 kilowatts of electricity...

... The Teraflop chip uses less electricity than many current high-end processors, making the design attractive for use in home computers.

It consumes 62 watts, and the cores can power on and off independently, making it more energy efficient.

I can't think of anything to add. I'm awed.

Building productivity solutions via mash-ups: Reider's office schedule

Jacob Reider ties a commercial (Misys) practice management system to Google calendar to build a distributed medical office schedule.

It's a fascinating example. Is it HIPAA compliant? Probably, though I can't swear to that. Is it reliable -- I'm less sure of that, Google is famous for breaking APIs. It is probably reliable enough for Jacob though, he can always fix it as needed.

Many eons ago, when dinosaur roamed the earth and Gopher was hot, I advocated for building our startup's solutions as an extensible ecosystem. I wasn't thinking web services, I was thinking more photoshop-style plug-ins. We would build things, others could build things, the sum of the ecosystem would be greater than a controlled solution. (I wasn't so foolish as to advocate open standards; even now the closest thing to that, the HL-7 RIM 3.0 spec, is a decade away from being robust.)

I got exactly nowhere with my proposal. The startup sold for a decent price, so I was probably wrong -- most startups simply fail. In a broader sense, however, I do believe in the original vision. There are win-win solutions in the universe.

I don't, however, buy Jacob's more radical proposal that one can craft an entire robust solution from off-the-shelf parts. It's the old probability curse. Building a 98% reliable solution from x integrating parts requires (1- x*y*z...) reliability from each component. That's hard to achieve even in a controlled web services architecture. It's worth remembering that even Gmail is not all that reliable -- there are times it's slow or even unavailable. We are ok with that because we can tolerate some offtime on our email. Gmail is only one controlled system.

Of course I've been wrong before ...

2/12/07: Jacob replies. Alas, no comments on his blog so I can't say hello that way. Really, the wrestler wasn't all bad! He was kind of a weird libertarian populist ...

The astronaut story: I'd want an MRI

A Navy pilot and astronaut engaged in some exceedingly bizarre behavior. The media are all over the 'feet of clay' story. Maybe, but this behavior isn't simply an extramarital affair or a straightforward substance problem. It's profoundly weird.

If I were her lawyers, and assuming she doesn't have a meth habit, I'd want an MRI and an organic brain syndrome workup. Sure, it could be bipolar disorder, but if she's no preexisting history she's a bit old for that. If there's no mass lesion I'd be thinking PET scan ...

Friends don't let friends buy a RAZR

My alter ego really despises his new Motorola RAZR V3M. He got lost on the Denver ring road while contemplating the depths of his dislike for this iPhone inspiration [1] ...

Gordon's Tech: Review: Motorola RAZR V3m Phone (Sprint):

... I have a bad marriage -- with my phone.

When I first bought my Motorola RAZR V3m I thought it was pretty feeble, but I needed something to tide me over until my Sprint contract expired. The main selling point was the mini-USB connector, I figured I could charge the phone from my MacBook (nyet) or my Dell laptop (yes). After a few days I thought it looked pretty interesting, especially when I cleaned out the default settings and installed Google's superb Gmail client.

Then I began to see the warts. They didn't go away, they just got wartier. Soon I decided the phone was a bit weak. That didn't last. I came to hate the phone with a deep and abiding passion. I was ready to rant against the RAZR in public spaces. I saw Paris Hilton holding the RAZR and I thought it was perfect for her...

The longish post has the full flogging of this foul marriage between decent hardware and really bad software. Don't let any of your loved ones buy this phone. If you own Motorola shares, sell them ...

[1] As in its so bad it may have convinced Jobs that Apple needed to bring the iPhone to market ...

Why Apple can't license Fair Play

Sometime in the past month I mentioned in a blog posting that DRM solutions require absolute control of the hardware chain. (update: it was 9/06, when iTunes stopped supporting the ROKR -- the only non-Apple FairPlay client - seems like I wrote that only yesterday ....) Today Apple (allegedly Steve Jobs) said the same thing:

Apple - Thoughts on Music

...The most serious problem is that licensing a DRM involves disclosing some of its secrets to many people in many companies, and history tells us that inevitably these secrets will leak. The Internet has made such leaks far more damaging, since a single leak can be spread worldwide in less than a minute. Such leaks can rapidly result in software programs available as free downloads on the Internet which will disable the DRM protection so that formerly protected songs can be played on unauthorized players.

An equally serious problem is how to quickly repair the damage caused by such a leak. A successful repair will likely involve enhancing the music store software, the music jukebox software, and the software in the players with new secrets, then transferring this updated software into the tens (or hundreds) of millions of Macs, Windows PCs and players already in use. This must all be done quickly and in a very coordinated way. Such an undertaking is very difficult when just one company controls all of the pieces. It is near impossible if multiple companies control separate pieces of the puzzle, and all of them must quickly act in concert to repair the damage from a leak.

Apple has concluded that if it licenses FairPlay to others, it can no longer guarantee to protect the music it licenses from the big four music companies. Perhaps this same conclusion contributed to Microsoft’s recent decision to switch their emphasis from an “open” model of licensing their DRM to others to a “closed” model of offering a proprietary music store, proprietary jukebox software and proprietary players.

In this case, I think Apple is telling the truth. Of course they've known this all along -- though you have to wonder about the ROKR fiasco. The situation for movies is even worse. You can be sure someone is storing all the currently encrypted movies they can find, knowing that sometime in the next five years they'll be able to hack them all at once.

Apple is saying that the music owners have to give up on DRM. I'm sure they'll agree ... :-)

More on DRM. Also, see my 2005 post on how DRM wrecked my media center experiments.

Buying dissenting voices: how it's done

Ever wonder where the 'smoking doesn't cause cancer' or 'CO2 isn't an important driver of our global warming' voices come from? They seem so much louder than their numbers suggest. Hint: It's the same way PACs buy politicians for wealthy interests:

Why Buy a Climate-Skeptic Cow When Milk is Cheap? | Cosmic Variance

... When I was an undergraduate (bear with me here) I spent a summer working at the Harvard-Smithsonian Center for Astrophysics. I worked with Sallie Baliunas, a CfA scientist who was a fellow Villanova astronomy grad, and was running an ambitious project to track chromospheric activity on a large sample of Sun-like stars. Sallie is an outstanding astrophysicist, and was a great advisor, as well as a friend. It’s no coincidence that I ended up going to grad school at Harvard’s astronomy department; the physics department didn’t like people from smaller schools and wouldn’t let me in, and Sallie helped convince the astronomy department to accept me.

Sallie also was, and continues to be, very right-wing, of the libertarian variety. Letting the free market do it’s job was the best strategy in nearly any circumstance, she firmly believed. Her interest in stellar variability led her to contemplating the role of Solar variability in the Earth’s climate, and she became convinced that changes in the Sun were essentially the only important factor in explaining changes in the Earth’s temperature. In particular, that human-produced emissions had nothing to do with it. Nothing about this belief was influenced in any way by large piles of cash offered by oil companies. But, once her views became known, they were more than happy to provide platforms from which to spread them; she’s now an editor at Tech Central Station, as well as a fellow of the George C. Marshall Institute.

Nobody could be more sincere in their views about climate change than Sallie is. I also happen to think that she’s dramatically wrong, as do the vast majority of (much more expert) scientists working on the question. But this is how the game is played — no need to bribe people when you can influence the public debate much more easily, and without fear that your targets won’t stay bribed. Unfortunately, oil companies have a lot more cash to spend on this purpose than the atmosphere does. Which is why public-minded scientists who agree with the carefully researched views of the IPCC need to keep hammering on the importance of doing something to fix this problem, before the damage is irrevocable.

Philip Morris developed many of these techniques to slow public health attacks on smoking. Dissenting voices are important for the health of science, but these economic interests give them vast power -- somewhat to the detriment of science. The effect can work the other way too, if science showed that chocolate prevented cancer then the candy companies would amplify that as well. Alas, I have a hard time thinking of examples where this effect works to the good. Drug companies, for example, vastly amplified voices promoting estrogen therapy for menopause -- beyond what the science justified. The result was a fiasco.

If there were an economic upside to promoting HIV dissenters, they'd have wreaked vastly more havoc than they have to date.

(Oh, those politicians? PACs don't pay a politician to change their vote. They find politicians who favor their positions and fund their election and reelection. It's not so much classic bribery as it is selection - often selection for very dim people. This selection effect probably explains the past the incompetence of GOP representatives over the past 20 years.

Salon: another spook speaks out

Yawn. Yet another spy points out the extraordinary incompetence and banal nastiness of the Bush administration. Nothing new here but I liked this comment on what happens when Cheney/Bush let loose the dogs of war...

Bush and Cheney's dirty secrets | Salon News

... From the perspective of the White House, it was smart to blur the lines about what was acceptable and what was not in the war on terrorism. It meant that whenever someone was overzealous in some dark interrogation cell, President Bush and his entourage could blame someone else. The rendition teams are drawn from paramilitary officers who are brave and colorful. They are the men who went into Baghdad before the bombs and into Afghanistan before the army. If they didn't do paramilitary actions for a living, they would probably be robbing banks. Perhaps the Bush administration deliberately created a gray area on renditions.

Robbing banks?

When does Amazon censor reviews?

Recently I wrote a negative Amazon review of the Motorola RAZR V3M. It's pretty but dumb.

Amazon didn't release it.

Odd. I've written hundreds of Amazon reviews, I don't recall one being rejected before. I edited my blog review a bit, and tried again.

Nothing.

Peculiar.

I went back and reviewed the other reviews of this phone. There were very few, and they were very weak. Some were negative, but they didn't say anything that might really hurt sales. My review would hurt sales.

It appears that Amazon is now censoring negative reviews of some products. Why this product? One hint might be that Amazon has a shady deal going with cell phones. If you buy on Amazon you get a significant cash discount -- but you sign up for a high margin plan. You have to keep the plan for at least six months, after which the cost of the plan equals the rebate you've received. If you don't cancel then, you pay through forever.

I've been trying out Amazon's "premium" plan (pay up front, get free 2 day shipping). I was toying with continuing it and paying for it. Amazon, after all, has been providing great value through their online reviews. Now Amazon has greatly decreased the value of those reviews for me. I trust them much less. I won't sign up for their premium plan after all.

An Amiga 2000 working for a TV network

The equivalent of using a horse and buggy to deliver cell phones:

Readers fess up to old-skool IT

... “While our local TV network has a high-speed network for editing newsrooms, video servers, resource management with lots of PCs running Windows XP and expensive software, live titling is still done using a 1989-built Commodore Amiga 2000. Bought almost 20 years ago, the beast still runs great using a Maxtor 240MB hard disk drive, with a 25MHz 68030 processor and 68882 math co-processor running Amiga DOS 3.1, and overlaying graphics to live video using a GVP G-Lock device. It has a total of 10MB of RAM. We have to go around in thrift shops hunting down Amiga 500s, Amiga 2000s for spare parts, keyboards and mice. Even a CDTV was victimised to get a Fat Agnus 8372A chip. The computer runs Scala IC500 as a titling program and it still runs great. We transfer data from the Windows boxes using 720KB formatted 3.5- inch diskettes.”..

Many of the other examples in this fun article are OS/2 machines. In addition to sheer cussidness, absolute security from bots and viruses is a winning advantage.

Hardware has improved exponentially over the past decade, though no consumer prooduct will ever again equal the tank-like construction of my original Panasonic [1] manufactured 8086 box. Software hasn't even managed a linear improvement; we had better wordprocessors 10 years ago than we have now.

[1] It's long forgotten now, but in the age when Japan was righteously crushing American car manufacturers Japanese computer manufacturers threatened to do the same to the nascent US PC market. Congress responded with aggressive protectionism, forcing Japan out of the desktop marketplace. Does Dell owe its present power to that action? [Yes, I know all PCs come from the "far east", but Dell is still considered a US company.]

Bandwidth: why the net is slow and getting slower

We don't have anywhere near the bandwidth we need to support video-on-demand (which is why Netflix and the post office still have years to go). Cringely has been delving into various solutions (Googleplex, P2P), but here he delivers a short tutorial on how ISP's work ...

I, Cringely . The Pulpit . WYSIWYB | PBS

... At the heart of this video distribution problem is the lie that ISPs tell about how much bandwidth we are really buying. While you may think your 1.5-megabit-per-second DSL service or your 3-megabit-per-second cable modem service is actually backed by 1.5 megabits or 3 megabits of Internet bandwidth, they really aren't. ISPs provision backbone access based on the expectation that people usually aren't on the Internet, and even when they are on the Internet most of their time is spent reading the screen, not actively sending or receiving packets. As such, ISPs have been able to get away with buying 20-30 KILOBITS per second of Internet backbone capacity for every MEGABIT per second of Internet service they are selling at retail. This 20-to-1 provisioning ratio of what's sold to what is promised (and believe me, 20-to-1 is me being generous to the ISPs since it is probably much higher than that) is what creates the burgeoning Internet video problem.

So if ISPs would simply provision the amount of bandwidth they are selling us there wouldn't be a problem, right?

Alas, that's not the complete story.

What would appear to be an obvious solution to this impending problem is increasing effective backbone bandwidth for every broadband user. There is no LOCAL bandwidth crisis, just an INTERNET bandwidth crisis. As I've explained over the last two weeks, Google is going to solve this for us in exchange for taking the majority of global advertising revenue. Maybe that's the answer, in which case we don't have to do anything. But if Google doesn't step in appropriately or you'd rather not rely on Google for some reason, there is the alternative of going with a different ISP.

If you want a beautiful Internet video experience, one answer might be to buy service from a broadband ISP that will allocate more backbone bandwidth per account. If you are a DSL user, you have a choice of ISPs in most markets. They all go through the local phone company, of course, but that phone company has no problem allocating the rated value of the DSL line because to the telco that last mile bandwidth is free. My DSL ISP is MegaPath and they buy a 1.5-megabit-per-second circuit from BellSouth (now AT&T I guess, but the trucks don't say that yet) and it can really carry 1.5 megabits per second 24/7, no problem. The problem is that MegaPath then makes me share a DS3 (45-megabit-per-second) connection to the Internet with every other customer of theirs in my LATA or service area, so in practice I get a LOT less than my 1.5 megabits.

One would think a national DSL ISP would have a huge business advantage if they could sell me a 1.5 megabit service and actually PROVISION a 1.5 megabit service. The wholesale cost of Internet bandwidth is right now around $15 per megabit per month. So delivering a REAL 1.5 megabits vs. the 30 kilobits I am probably getting right now will cost my ISP 75 times as much, but that's still only $22.50 vs. $0.45. So the question is whether I would pay $22.05 more than I am paying right now to be able to effortlessly watch video over the net?...

Along the way Cringely admits that ISP's do have a legitimate grievance with "net neutrality". Their business model depends on that 20:1 provisioning ratio, but some kinds of net use are truly incompatible with that sparse provisioning. For things like video they have to buy much more backbone capacity -- so they bear a lot of extra cost. Of course they ought to be able to pass that on to customers, but imagine the screams from customers who aren't getting value from the streaming video ...