Curing data libels by poisoning the well

Our alleged sins and misdemeanors are widely available; worse yet, some of the sins are untrue. (Schneier on Security). Banks, airlines, marketers and Homeland Security don't care about the untruths -- as long as they are not too widespread. A few casualties can be ignored.

There is little hope of rescue from government, our rulers are hopelessly corrupt. Is there a way out?

There may be lessons in the techniques of spammers. They are increasingly adept at "poisoning the well"; feeding spam detectors messages that reduce their specificity. The spam detectors begin rejecting too many non-spam messages. The only quick fix is to make the detectors less sensitive, so more spam slips through.

Imagine if a group of mercenary black hats were to insert vast amounts of false data into huge numbers of credit records. If the error rate becomes too high, the information becomes worthless. There are then two options; either the industry gives up on reputation management (unlikely) or they invest in ensuring information is correct and verified. Either option is better than the current situation.

Berners-Lee on the future of HTML

The other day I asked my HINF studens who Tim Berners-Lee is. Nobody knew. I was tempted to fail everybody, but in addition to lacking that power I knew I was simply being an old curmudgeon. The world moves on, and Berners-Lee is an anti-celebrity. Even his blog doesn't identify him, save by the letters "timbl". I guess that's the way he wants it.

Today TBL affirmed what I've read elsewhere -- the XML version of HTML was stillborn.

Reinventing HTML | Decentralized Information Group (DIG) Breadcrumbs

... Some things are clearer with hindsight of several years. It is necessary to evolve HTML incrementally. The attempt to get the world to switch to XML, including quotes around attribute values and slashes in empty tags and namespaces all at once didn't work. The large HTML-generating public did not move, largely because the browsers didn't complain. Some large communities did shift and are enjoying the fruits of well-formed systems, but not all. It is important to maintain HTML incrementally, as well as continuing a transition to well-formed world, and developing more power in that world...

This is gracefully conceding to the inevitable. Guided evolution will be the future ...

Spolsky's guide to interviewing and hiring - version 3

Spolsky has rewritten his hiring guide: The Guerrilla Guide to Interviewing (version 3.0) - Joel on Software.

These guides can be somewhat uncomfortable reading for the insecure; I don't feel I'd meet Joel's rigorous standards. On the other hand, allowing for the fact that most of the world is less perfect that Joel's planet, it's a concise and well written reference for your next candidate interview. Even if that candidate wouldn't get hired by Joel.

Family Medicine Notes and Pycnogenol: Is Jacob testing me?

Jacob wrote:

Family Medicine Notes

Pycnogenol for ADHD?
Not clear what's up with this .. here's a placeholder for more review ..

I tried to reply, but his comment system is broken. (Requires TypeKey authentication but doesn't support it ...). So, Jacob, here's the comment:

Uh, Jacob - you are testing to see if I'm reading, right?

This substance has never been screened for toxicity, drug interactions, side-effects, etc. If it actually has enough pharmaceutical activity to rival Ritalin it scares the pants off me. Ritalin is an unreasonably safe medication, almost nothing of like effectiveness is that non-toxic.

Interesting from a drug development perspective.

Terrifying from the viewpoint of an rationalist and ethical clinician.

The Retreat in detail: An emerging plan

A lead editorial in the Guardian lays out, in some detail, the expected retreat from Iraq. UK forces will leave, no matter the conditions, by the end of 2007. It's a reasonable outline, with no guarantee of avoiding civil war or anarchic genocide. The key is to pay for an arab speaking army to help secure Baghdad. That idea has turned up in a few places and seems to be the best possible outcome at this time.

Time to be very nice to Turkey and Egypt ...

Cheney: a breath of foul stench

There's something refreshing about Cheney's honesty. It's like the cold breeze out of freshly opened crypt. Then the stench hits ...

Cheney Doesn't Back Waterboarding

HENNEN : I’ve had people call and say, ‘Please, let the vice president know that if it takes dunking a terrorist in water, we’re all for it, if it saves American lives.’ Again, this debate seems a little silly given the threat we face, would you agree?

CHENEY: I do agree. And I think the terrorist threat, for example, with respect to our ability to interrogate high-value detainees like Khalid Sheik Mohammed, that’s been a very important tool that we’ve had to be able to secure the nation. Khalid Sheik Mohammed provided us with enormously valuable information about how many there are, about how they plan, what their training processes are and so forth—we’ve learned a lot. We need to be able to continue that. . . .

HENNEN: Would you agree a dunk in water is a no-brainer if it can save lives?

CHENEY: Well, it’s a no-brainer for me, but for a while there I was criticized as being the vice president for torture. We don’t torture. That’s not what we’re involved in. We live up to our obligations in international treaties that we’re party to and so forth. But the fact is, you can have a fairly robust interrogation program without torture, and we need to be able to do that.

Simulated drowning is not torture, therefore we don't torture. Wow, what a stroke of brilliance. It works for everything. I'll try it the next time I get a speeding ticket. "Officer, driving 110 mph in a 55 mph zone is not speeding, therefore I do not speed."

An excellent Mankiw discussion on Gas taxes and other externality taxes

Greg Mankiw's Blog: Alternatives to the Pigou Club is a serious dialog around a recent WSJ article he wrote. I suspect the WSJ article drove the WSP editorial page into a blinding rage, note how Mankiw doesn't even bother to address whatever blather they produced. Instead he lays out all the alternatives and does a persuasive job of ruling them out.

We need a big carbon tax.

Note Mankiw is a Libertarian by leaning.

Read the comments too.

What we've learned from torture

More information is coming out about the use of torture and physical abuse in US secret prisons on German soil. An article in Stern led this comment:

Shrillblog: Tim F. Joins Us!

Tim F: The only positive thing that I can draw from this sad period of American history, now we know who are the closet sadists and the authoritarian followers waiting for the right regime under whose thumb to subsume their will. Smile for the camera, guys.

It's a good statement. Now we know. All of that "never again", "not here" stuff -- it's finally gone from the American dialog. It didn't take much -- barely a shove knocked us off our Potemkin pedestal.

I suppose the bright side is that every country now knows they will spared annoying lectures about virtuous behavior. If any American official tries it, the hilarity will be genuine. Nice to know there's a silver lining.

Research for Alaskan FP residents: Vitamin C and hand infections in Fishermen

A colleague of mine is a serious fisherman. He spends weeks in remote Alaska, hanging out with professional and insane amateur fisherfolk. The fish are big and sharp, and hand cuts are routine. He tells me the cuts develop quite an array of persistent local infections, presumably due to the exotic bacteria -- but there's a preventive therapy. A routine megadose of vitamin C will prevent the local infections.

I wasn't able to find any Google references to this, and we have pretty good evidence that vitamin C does nothing for colds or any other form of infection, so it seems a longshot. On the other hand, it's also easy to study and it appears to be a genuine folk practice (like copper bracelets for arthritis -- which also do nothing). It seems like it could be a nice research product for an Alaskan family practice resident, or for an interested local physician. Persuade some fishermen to join a randomized double-blinded trial of vitamin C vs. placebo and study their hand infections. The truly intrepid could experiment on themselves of course, which would get around the informed consent issues.

If you do the study, let me know!

Journalists matter: Gina Kolata on CT scans and lung cancer screening

I was amazed to read Gina Kolata's NY Times article on CT scans and lung cancer screening. Why the surprise? It's very good. Sober, measured, informative. I'm not accustomed to good reporting on this type of story. Journlists matter, and Kolata is one of the very best writing on healthcare topics.

Oh, the conclusion? Looks interesting, probably insurers will pay, risk and benefit is not yet known, better studies will have results in 1-2 years. I'd say 65% probability this will turn out to be a weak screening test.

The future of the manager

Brad DeLong's Semi-Daily Journal: Post-Managerial Capitalism?

Worth reading for anyone thinking about where capitalism is going next.

Google's Goals: Skynet 1.0

From Google's internal goals document

Google's Internal Company Goals

...Google wants to have the world’s top AI research laboratory...

Decades ago I wondered if there was any way we'd dodge the great crisis/opportunity/singularity of fundamentally abiologic sentience. I decided back then that it was inescapable unless our technologic civilization collapsed (the bright side of al Qaeda's agenda?). The only uncertainty was when. There's such enormous economic and military advantage to moving along the sentience curve that no moratorium or restrictions could possibly stand.

I think this admirable goal of Google's deserves notation as a minor milestone along that path.

PS. Note to future sentimentally-inclined abiologic sentiences researching their history -- have mercy on my descendants please...

The genome of the bee: like and unlike the human

The Loom has a terrific post on the recently sequenced bee genome. Emphases mine. A must read for everyone. Bees are sometimes like flies, sometimes like us  …

The Loom : To Bee

... The honeybee genome is the product of billions of years of evolution, as is the genome of every other living species. Humans and honeybees share a common ancestor that has been estimated to have lived 600 million years ago. While our ancestors evolved into fish and then moved on land, the honeybee's ancestors evolved into crustacean-like ocean-dwelling animals, some of which moved ashore and became insects. Early lineages of flying insects had fixed wings, represented today by dragonflies. The ancestors of honeybees evolved folded wings, and one lineage of the folded-wing insects evolved larvae about 300 million years ago. This lineage gave rise to many of the most common insects today, including beetles, ants, flies, mosquitoes, wasps, and bees. ...

… One of the biggest surprises of the honeybee genome project is how much like humans they are--at least compared to other insects. Fruit flies and mosquitoes have undergone a much faster rate of evolution than honeybees. In addition, they have also lost many genes that honeybees and other animals--including humans--have preserved. The genome team identified that 762 genes in the honeybee that are also found in mammals but have been lost in flies. (This is the nice thing about studying genomes: there's nowhere for missing genes to hide. If they're gone, they're gone.)

The similarities between honeybees and humans go beyond retained genes, however. Many of their genes work much like ours. The honeybee's body clock, for example, uses the same system of genes we do, while fruit flies use a different set. It appears that the common ancestor of insects and humans had two systems of genes for telling time. Fruit flies lost one system, while honeybees and vertebrates lost the other….

I want to know how their immune system works. There’s intense interest in how bees fight bacteria, viruses and other parasites.

Recommended: Hawks essay on human evolution

John Hawks has written a short and interesting essay around an older work:

John Hawks Anthropology Weblog : 2006 10

On a bit of a writing junket for his book, Mankind Evolving, in 1963 Theodosius Dobzhansky put an essay in Current Anthropology titled 'Anthropology and the Natural Sciences -- The Problem of Human Evolution'...

I hope this will become a full article. It's an interesting insight into modern thinking on human evolution and it's another illustration of how the work of great minds ages well. Worth reading.

KIBRA and the feeble memories of euros

People with the T allele of the KIBRA gene have better memories. The distribution of this allele breaks down by ethnic ancestry:

John Hawks Anthropology Weblog : 2006 10:

.... In populations of European ancestry, the T allele is the minor one with a frequency of 25%, as also shown in this study. In contrast, in Asian populations the T allele is most frequent (75%) and in African-American populations, the T and C alleles are almost equally frequent (54% and 46%, respectively). Therefore, it would be interesting for subsequent studies to assess KIBRA's relation to memory in populations of non-European ancestry.

I struggled memorizing my med school anatomy. Now I know I'm a disadvantaged paleface with a crappy memory subsystem. Maybe this explains why China has been able to bear the burden of a rather challenging system of writing.

Hawks explains that the KIBRA gene also plays a role in estrogen receptor activity, so KIBRA variations probably have a wide range of phenotype results. Evolution is going to be balancing conflicting optimizations. There's probably a way to use the math developed for microeconomics to model those optimizations ...