Monday, January 12, 2009

Unexpected social good - a post on roaring fans

My tech blog posts are partly for my own use, and partly I'm trying to give back.

Sometimes it seems to work; some posts get very kind comments.

There's one, however, that really stands out.

An otherwise unremarkable May 2008 post on how badly written OS X print drivers cause howling, roaring, fans has earned me more “thank yous” than almost anything else I’ve done in the past few years.

Funny how that works.

China 2009 is about to become Japan 1986

So you’ve probably been thinking that after the last ten years we’re due for a respite. A bit of calm amidst the floating refuse of our crated life raft. It was a big waterfall for sure, but now we get to float for a while.

Then you hear the roar of the next cataract.

Very few people now remember than by 1986 Japan had just about finished off the American car industry and was about to conquer the American computing industry.

What'? You don’t remember the congressman taking a sledgehammer to a Japanese made IBM-clone?

I remember because my very first PC was an 8086 clone made by Panasonic. It was the most over-engineered device I’ve ever owned. It was built to last thirty years, and it cost less than the other clones on the market.

IBM, Compaq, and the like were terrified. Congress was appalled – America had lost the automobile*, now we were going to lose the computer too.

De facto “voluntary” quotas were enforced. Panasonic and other Japanese PC vendors left the desktop market. Dell became dominant.

Flash forward 23 years …

Chinese Electric Automaker Challenges Big Three

…Chinese car maker BYD Co. BYD developed an electric automobile that only costs consumers $22,000…

Gee, this feels familiar.

Wait until Chinese/Google “Chromestellation” netbooks start selling at Walmart for $124.99 (batteries not included).

Eeeeeeeyyyyyaaaaaaa …..

* The same set of “voluntary” quotas forced Japan to manufacture in the US, and also allowed GM, Chrysler and Ford time to more or less compete with Japan.

Guru-level air travel tips

I am a lowly travel worm, unworthy to learn the techniques of the travel Ninja (emphases mine) …

Being John Glaser 1/6/09

… Preventing the person in front of you from reclining their seat (Spilling soda on their head is minus 5 points. Pointing the air vent so that it blows on their head is plus 5 points.)…

… Stopping the person next to you from reading over your shoulder. (Saying “Read my stuff again and I’ll kill you” is minus 5 points. Turning the book/newspaper upside down and continue reading earns the contestant 5 points.)…

Captured via Jacob Reider’s Google Reader shared item list.

Google responds to CO2 emission criticisms

I don’t know the back story here, but clearly Google is sensitive to accusations of environmentally incorrectness. Their response is fascinating …

Official Google Blog: Powering a Google search

… Recently, though, others have used much higher estimates, claiming that a typical search uses "half the energy as boiling a kettle of water" and produces 7 grams of CO2. We thought it would be helpful to explain why this number is *many* times too high. Google is fast — a typical search returns results in less than 0.2 seconds. Queries vary in degree of difficulty, but for the average query, the servers it touches each work on it for just a few thousandths of a second. Together with other work performed before your search even starts (such as building the search index) this amounts to 0.0003 kWh of energy per search, or 1 kJ. For comparison, the average adult needs about 8000 kJ a day of energy from food, so a Google search uses just about the same amount of energy that your body burns in ten seconds.

In terms of greenhouse gases, one Google search is equivalent to about 0.2 grams of CO2. The current EU standard for tailpipe emissions calls for 140 grams of CO2 per kilometer driven, but most cars don't reach that level yet. Thus, the average car driven for one kilometer (0.6 miles for those in the U.S.) produces as many greenhouse gases as a thousand Google searches.

I think it would be interesting to compare Google search CO2 prediction to CO2 production from a 1 km walk.

Sunday, January 11, 2009

2009 will be the year of machine translation

Speaking of the singularity, I've subscribed to my first foreign language blog, thanks to Google Reader's integrated machine translation services.

I've read a few predictions for 2009.

Here's mine.

This will be seen as the year that machine translation went mainstream.

The world's IQ is about to go up 10 points.

The press makes much of Apple and Palm's recent twitches, but those companies are dull as dust compared to Google. I don't think we grasp what's happening there.

Singularity watch: Google maps is now smarter than me

Mercifully, the Singularity seems to be receeding.

On the other hand, the days when I was smarter than Google Maps are now only a fond memory.

It's now clearly better than I am at finding routes through the Twin Cities. I've taken to checking Google Maps' recommendations on routes I routinely travel just to see where I shave a few minutes.

Barring a true global dacopalypse Google Maps is only going to get smarter, and barring a scary miracle I'm only going to get dumber. Our ships have crossed in the night.

I'm getting that old John Henry feelin'.

Saturday, January 10, 2009

Health care reform. Be humble. Be afraid of the accounting rules.

We're going to change American health care again.

It would be good to learn from past mistakes. In particular it would be good to remember that accounting is terribly dull and terribly powerful, which means it's terribly dangerous.

Consider the way we account for ambulatory care today.

In the 1980s, when I was doing my family medicine training, the Feds wanted to encourage primary care physicians. They looked for a "fair" way to divert money from procedural specialties to primary care and came up with the "RBRVS" (emphases mine) ...
Resource-Based Relative Value Scale - Wikipedia, the free encyclopedia

... RBRVS was created at Harvard University in their national RBRVS study from December 1985 and published on September 29, 1988. William Hsiao was the principal investigator who organized a multi-disciplinary team of researchers, which included statisticians, physicians, economists and measurement specialists, to develop the RBRVS.

In 1988 the results were submitted to the Health Care Financing Administration (today CMS) to be used in the American Medicare system. In December of the following year, President George H. W. Bush signed into law the Omnibus Budget Reconciliation Act of 1989, switching Medicare to an RBRVS payment schedule. This took effect in January 1, 1992. Starting in 1991, the AMA has updated RBRVS continually. As of May 2003, over 3500 corrections have been submitted to CMS.

Physicians bill their services using procedure codes developed by a seventeen member committee known as the CPT Editorial Panel...
To implement the RBRVS rules an accounting system had to be developed. That accounting system is expressed as the rules for "Evaluation and Management" codes.

Who developed that accounting system?

The AMA.

Who funds and controls the AMA?

Surgeons and procedural specialists.

So the RBRVS, aimed at improving reimbursement for non-procedural work, was transformed into accounting rules by an organization dominated by proceduralists.

It worked about as well as one might expect. From the mid 1990s through 1999 the "E&M" rules went through 3 revisions amidst a bitter struggle between physician specialties. The controversy was so great that the 1999 rules were never implemented, and today variations of the 1995 and 1997 rules are both accepted.

The rules are, and I say this carefully, insane. I've designed expert system solutions that worked with the rules, and no human should ever have to think about them. The saving grace is that despite all the complexity and branches and calculations and summations and variations, there are only 3 likely code choices for the care of a particular patient. Physicians pick what feels right, knowing that they must be breaking the rules some of the time and so, on average, tending to bill a bit on the low side.

The worst effect, however, is how the accounting rules perverted primary care. The penalties for error are theoretically severe -- for a while the FBI devoted a special team to looking for cheaters. Physicians are required to document all that's done, and thus care was radically changed to focus on documentation of material of virtually no clinical value. The modern medical note is almost unreadable.

Lastly, since the accounting rules involve points for bits of the body examined, physicians are incented to spend too much time on rote and pointless examination and documentation, at the cost of thinking about a patient's needs and problems.

Primary care is in bad shape today. I wonder if William Hsiao has any idea what came of his original work. It's a beautiful example of the unintended consequences arising from accounting rules.

Let's try to remember the lessons of the RBRVS and the E&M coding story.

Accounting is dull, Accounting is terrible.

Be afraid of accounting.

Friday, January 09, 2009

Election turnout numbers: Highest in MN, GOP voters stayed home

My home state of Minnesota is just so darned fabulous ...
FactCheck.org: Is it true that 36 percent to 37 percent of eligible voters failed to vote in the recent presidential election?

... Minnesota ... had the highest voter turnout, at 77.8 percent. Hawaii and West Virginia are tied for the lowest turnout, with 50.6 percent each...

... Before Americans went to the polls on November 4, much was made in media reports about record levels of voter registration and high enthusiasm levels among the electorate. And while the 61.6 percent turnout number doesn't seem that impressive – in 2004, after all, 60.1 percent of eligible voters cast a ballot for the highest office – it is the highest turnout in the U.S. in decades. As the CSAE report says, "If the rate of voting exceeds 61.0 percent of eligibles, turnout will have been the highest since 1964."

But why was it not even higher? Republican turnout, according to CSAE, dropped, while Democrats voted in higher numbers. The percentage of those voting for the Republican presidential ticket dropped by 1.3 percentage points and those voting for the Democratic ticket went up by 2.6 percentage points from 2004. Curtis Gans, the center's director, said he, too, thought even more Americans would vote in 2008. "... we failed to realize that the registration increase was driven by Democratic and independent registration and that the long lines at the polls were mostly populated by Democrats...
The GOP voters stayed home, the Democrats came out, and we had a historic win with a small increase in overall turnout which was, still, high by historic standards.

Incidentally, I'm fine with a 60% turnout. As best I can tell about 40% of Americans have no idea what's going on, and far too many of them vote for people like Cheney/Bush. I'd rather they stayed home ...

Update 1/11/09: As noted in comments, MN has a special advantage -- we're not on the west coast. After PA was called west coast voting really fell off. So my comparison is a bit unfair ...

Ultrasound and the developing brain – lessons from manipulation of mouse neurons

There’s a bit of wing-nuttery on the net about a possible relationship between the widespread use of obstetric ultrasound and an increase in the percentage of children diagnoses with autism (though there’s also been a simultaneous decrease in the percent of children diagnosed with mental retardation).

Sometimes the discussions have even had humorous consequences.

Still, there’s some reason for interest.

Which brings us the use of intermediate intensity ultrasound for altering the brain …

Sound and no fury | The Economist

… William Tyler and his colleagues at Arizona State University..

… knew from experiments done by other groups of researchers that ultrasound can have a physical effect on tissue. Unfortunately, that effect is generally a harmful one. When nerve cells were exposed to it at close range, for example, they heated up and died. Dr Tyler, however, realised that all of the studies he had examined used high-intensity ultrasound. He guessed that lowering the intensity might allow nerve cells to be manipulated without damage.

To test this idea, he and his colleagues placed slices of living mouse brain into an artificial version of cerebrospinal fluid, the liquid that cushions the brain. They then beamed different frequencies of low-intensity ultrasound at the slices and monitored the results using dye molecules that give off light in response to the activity of proteins called ion channels. (An ion channel is a molecule that allows the passage of electrically charged atoms of sodium, potassium, calcium and so on through the outer membrane of a cell.)

The purpose of all this was to coax the cells to release neurotransmitters. These are molecules that carry information from one nerve cell to another. When they arrive, they cause ion channels to open and thus trigger the electrical impulses that pass messages along nerve fibres. When those pulses arrive at the other end of a fibre they, in turn, trigger the release of more neurotransmitters.

Disruption of this system of communication is characteristic of several medical conditions, including Alzheimer’s disease, Parkinson’s disease, depression and epilepsy. Ways of boosting the release of neurotransmitters may thus have therapeutic value. And the ultrasound did indeed boost their release.

How that came about is not absolutely certain, but Dr Tyler thinks the shaking that his ultrasound gave to the cells in question opened up some of their ion channels. The cells were thus fooled into acting as though an impulse had arrived, and released neurotransmitters as a consequence…

So the obvious question is how does the intensity and duration of the ultrasound used in these experiments compare with the intensity of ultrasound used in obstetric scans? After all, “disruption of this system of communication” is also a characteristic of autism.

It feels like it would be wise to do further animal model studies, and to discourage obstetric ultrasound done largely for entertainment purposes rather than to truly guide and manage pregnancy.

Thursday, January 08, 2009

Palm Pre - unimpressed

I don't care if the Palm Pre is the world's greatest mobile web wonder.

I have that problem solved.

The problem I have is work/home calendar/task/contact integration in a world where corporations hold tight to their Exchange data.

If Palm were addressing that problem, I'd be interested. The original US Robotics PalmPilot made a stab in that direction, and in the 90s we could connect a serial cable and suck data from Outlook at work while synchronizing at home.

It wasn't pretty and it didn't work that well, but the Palm Pre appears to be in a totally different zone.

A zone owned by the iPhone.

Apple's product cycles - handy for purchase planning

Phil Schiller described Apple's product cycles for David Pogue ...
Gmail - Circuits: A Strange Macworld Expo

.. the holiday season (Novemberish), the educational buying season (late summer), the iPod product cycle (October), the iLife development cycle (usually March), the iPhone cycle (June)..
So we have by month
  • March: iLife
  • June: iPhone (? and MobileMe?)
  • August: Educational (iWork? Hardware tweaks?)
  • October: iPod
  • November: holiday things - hard to know what this means
This doesn't include dates for the professional software (Aperture, etc) or for major OS releases. Since we know the iLife news (March) we have quite a period of predictability ahead.

Wednesday, January 07, 2009

NYT in the ICU - why?

Readers of Henry Blodget's (yes, that Blodget) articles for Silicon Valley Insider know that the NYT is on its deathbed.

What we don't get is an explanation of why the New York Times, and most other large papers, are so ill.

We know that many media owners carry a lot of debt unrelated to newspapers, but in that case healthy newspapers would face sale -- not extinction.

We know that the internet killed the classified business, but no-one's suggesting that was huge for the Times.

We know subscribers have left paid print for free online reading, but online circulation has vastly lowered the average cost of delivering product while also increasing the pool of readers. In theory subscription losses should be outweighed by advertising revenues.

Ahh, but there's the rub. If I'm reading this VentureBeat article correctly, the problem is that the advertising model isn't working, either because online ads in newspapers don't seem to work and thus aren't worth much, or because there are a huge number of equally useful (or useless) routes to reader eyeballs. No one route can reach readers ...
If the New York Times dies, does the news die? VentureBeat
The death of an institution isn’t far off, writes the Atlantic in an article titled End Times, and with it an entire industry may be preparing to slip underwater. Low on cash, high in debt, the legendary New York Times is reeling from the recession. There’s no guarantee that it, or many others of our best newspapers, will survive the next year.
The immediate effect of the Times ending its storied run (or degrading to a lesser entity) will no doubt be the journalistic equivalent of a nuclear explosion...
... The New York Times has done an excellent job of growing its web property. ComScore says the company’s pageviews are approaching 200 million a month; that’s a lot for any website... 
... Getting a New York Times-caliber feature article requires paying a Times-caliber writer for a week or more of research and writing. That will set you back between $1,000 and $10,000, depending on who is doing the work. That figure doesn’t include the editing and expenses, by the way.
For most sites, that means they need 100,000 to a million pageviews to break even, for a single article...
If the NYT does die, I think it will take decades to replace it. I keep returning to the death of BYTE in the 1990s -- we still don't have anything like it on the web.

I'm optimistic an escape route will appear. It is a rough spot though ...

Tuesday, January 06, 2009

Kurzweil - not so good at prediction after all

Kurzweil, to his credit, has kept his 1999 predictions for 2009 online.

He doesn't do very well, though I give him credit for trying. Most of the things he mentions seem credible, but the hardware items are more 2018 and the software items 2028 or later.

Last September Scott Aaronson critiqued Kurzweil's prediction of a 2045 Singularity. Aaronson voted for 2300. I worried then about 2100.

After seeing how far we are from Kurzweil's 1999 predictions 2045 looks extremely unlikely and I'm feeling better about 2100 (meaning it will be later than 2100).

So not in my lifetime, or, and this is rather a problem for him, not in Kurzweil's lifetime either.

GD II: How large is the unused economic capacity?

I must be up to at least 30 Crackpot points in my recent economics threads, but I just can't help myself.

I swear, I'm going to stop soon, just not quite yet.

In past recessions, and even recently, I've read of estimates of the unused capacity of the US economy, and how much must be done to move us closer to using all of our capacity.

So today's question is whether the US economy is the right denominator.

What if we're now so entangled economically with China that the right denominator is not the US workforce, but rather the US and Chinese (and perhaps Indian) workforce? That's a rather bigger number; it could absorb a lot of stimulus ...

Crank roll:
  1. Lewis and Einhorn: repairing the financial world
  2. The role of the deadbeats
  3. Complexity collapse
  4. Disintermediating Wall Street
  5. The future of the publicly traded company
  6. Marked!
  7. Mass disability and income skew
  8. The occult inflation of shrinking quality

Dyer - the last articles of 2008

It's been a while, but Dyer has 5 new 2008 articles:
If he follows past practices he'll have a 2009 page up soon -- there's nothing there yet. I'll have to edit my Page2RSS monitor to track that new page.