Skip to content

Flies at 6000 meters, carrying a backpack…

January 18, 2022

Great-Reed-Warbler-3There is much we don’t understand about birds, from their physiology to what triggers their migrations. Ecologists from the Max Planck Institute of Animal Behavior attach digital trackers to warblers to observe their migration flights from Europe to Africa. They describe those as “tiny backpacks.” But what is tiny to an ecologist might seem a fair heft to a bird that only weighs about an ounce. Yet the warbler flew at heights no one thought it would. That’s a great reed warbler. Fully ladened.

We could have done better…

January 17, 2022

OmicronHospitalizationUnsurprisingly, hospitalizations are starting to follow the current Covid surge upward. Moreso in the US, than in the UK and some north European nations. Eric Topol looks at some historic data on that. See the first graph, right. The second graph shows the vaccination history of those same nations, as partial explanation.

The US is the outlier, starting vaccination earlier, ending with the lowest fraction of population vaccinated, and now having the highest hospitalization. Norway is an outlier in the other direction. Which reminds that geography and culture are part of the causal mix in the course of a disease. The kind of data snapshot those graphs provide doesn’t substitute for a study that applies some rational methodology and that does some statistical analysis. We know from rigorous studies that the vaccines work. This merely illustrates the consequence of their greater and lesser use.

VaccinationRatesAs second illustration, Zeynec Tufecki points to some data from Alberta, where the vaccinated and boosted people over 80 years of age have less Covid hospitalization than the unvaccinated in the 12 to 29 age range. Technology doesn’t yet provide a potion of youth. But on this one disease, one simple act gives your immune system the reliability of someone a half century younger. To stave off many of the other infirmities, try this.

Why I won’t watch that video

January 12, 2022

Video is great for some purposes. It can entertain. It can show quite vividly why climbing a tree is a lousy way to escape a bear. Teaching typically is done visually because it can lay an easier introduction for those new a subject. 3Blue1Brown is a youtube channel that provides visual insight to a variety of math. Their explanation of Bayes theorem is worth watching by anyone trying to understand basic probability theory. And there is the rare person who is both entertaining and informative. I can watch most any video Sabine Hossenfelder makes, such as her explanation of hyperloops and the technical challenges they face.

Those accolades made, there are reasons research papers and most serious discussions at the edges of knowledge are written, rather than spoken into a camera.

  • Reading is faster. College students who are native English speakers typically read about 300 words per minute. Spoken English is half that or less. No one wants to commit to an hour video without good expectation of its worth. (The ones linked above are mercifully short.)
  • Reading proceeds at each reader’s own pace. Which means a passage that is difficult or questionable or presents an idea new to the reader can be digested slowly, while those that are more straightforward can be more rapidly done.
  • Text more readily carries the technical details that matter. It can be tricky to speak an equation or formula, and even more difficult to understand it spoken, rather than written. Even conveying correct units is more difficult in speech than in text.
  • Text is easier to navigate. Someone reading a research paper may read through the abstract and methodology, then look at some of the graphs and visualizations, thinking on alternate explanations, including data collection issues, then read the discussion, then go back and dig into some of the thornier theoretical content, then look at some of the references to see if they actually provide what is claimed, then return to the methodology or discussion. Different readers will take different paths. In a video, the maker chooses the path, and it’s not so easy to jump back and forth.
  • Text lends itself to excerpt and cite. I had to play with the pause and rewind in that hyperloop video to get this not-quite-sarcastic quote on the vibration problem: “Passengers may be willing to accept the risk of dying from leaks in a capsule surrounded by near vaccuum. But only as long as they are comfortable before they die.” Now that it is in text, it can be copied and pasted. With correct attribution, please.
  • All of which makes text a better object of revision and idea evolution. Someone who reads a paper can spot ambiguous phrasing, quote arguments they think need sharpening, list what is missing, explain that some graph doesn’t really show what is claimed, question the methdology and analysis. Their respondent can address those issues with reference to what is in the paper, by quote and heading and page.  Video is not so easily treated.
  • Which is why it is papers that get submitted to journals and refereed. That is a meaningful process. Especially for readers outside the field. I can read and understand some medical papers. But without the field background, I am largely dependent on the authors and reviewers that there isn’t something wrong in an important way, and in a way that isn’t obvious to readers like me who have some background in biology, but who lack medical expertise.

While Dr. Hossenfelder makes fun videos, when she wants to present a novel piece of physics, such as showing that the Born rule follows from some assumptions about transition probabilities, she writes a paper on that. Having written papers, I will guarantee the one published went through a variety of revisions. And that she has started quite a few papers that never made it through that process.

So no, I’m not going to watch a video by some doctor offering an alternate viewpoint on some medical issue, or by someone alleging an unheralded advance in science, or by many others who want to persuade us of some factual matter that, were their research good, should be shown with studies and papers. Were those studies done? Were those papers written? Were they reviewed by those in the field? Have you — who want us to watch that video — read them? If those exist behind the presentation, that is where you should point us. If they don’t, why are you so convinced by a mere video?

Cosmic rays and quantum computers

January 11, 2022

Cosmic rays are causing a problem for quantum computers. Phycisists and computer scientists I know are taking that problem seriously. There are a variety of paths to workarounds. It will take some work. 

Another Republican stands up to the big lie

January 10, 2022

South Dakota Senator Mike Rounds publicly stood up to Trump’s big lie:

As a part of our due diligence, we looked at over 60 different accusations made in multiple states. While there were some irregularities, there were none of the irregularities which would have risen to the point where they would have changed the vote outcome in a single state. .. The election was fair — as fair as we’ve seen. We simply did not win the election as Republicans for the presidency.

Like Kinzinger and Liz Cheney, I suspect he will find that simple honesty about an election puts him on the wrong side of his party today.

The unfolding

January 5, 2022

The first attempted coup by a sitting president doesn’t have a true anniversary. Trump made ongoing efforts at overturning his election loss from the time close states certified their results. An article at The Bulwark points how close he came to succeeding. January 6th was just the denouement, when all else had failed. Approaching that anniversary, the NYT editorial board joins arms and pisses into the wind. We can hope, in the new year, that the criminal prosecution of those involved works its way from the dupes and pawns up to some of the leaders. Perhaps Garland’s speech today will hint at ongoing  investigations. But would seeing their leaders prosecuted and their lies exposed much change the cult that Trump rode to power? That seems unlikely. Stephen Marche, a Canadian, explores some of the uglier possibilities in The Next Civil War, speculative fiction on how this unfolds for the US. Or rather, on how the US unfolds. Though he hits quite accurately when interviewed, I’m not eager to read that book.

The uncounted

January 3, 2022

While the tally of US Covid-19 deaths is near 825,000, there is good reason to suspect that the actual toll is significantly higher. Many deaths go poorly attributed:

The nation’s struggle with recording COVID-19 fatalities underscores a truism about death in the United States: Where people live and die has a lot to do with the accuracy of their death certificate. Some deaths are investigated with state-of-the-art technology and expertise. Others don’t go beyond a phone call from the family.

I suspect that happens some even in rich urban areas with relatively good processes for accounting the dead. When all three people in a household are found dead, that raises enough notice that the local coroner does autopsies and micrology tests. But when a lone and homeless alcoholic, with cirrhosis and few social ties, is found dead one morning on a sidewalk, how much effort is spent to determine whether they were suffering Covid-19 pneumonia? Or might some overworked examiner just attribute it to cardiac arrest, subsequent to exposure and long-term alcoholism?

Jean-Henri-Marlet-Morgue_de_ParisWe want our government institutions to accord people the same respect in death. We want good data. Like many things we want, there is only so much we’re willing to pay for it.

Well, we’ve made good progress since the Paris morgue in the 19th c. put unclaimed corpses on public display, so that those missing friends or family might walk by and find them. The morgue then became a popular tourist destination. See the contemporary engraving, right.

Those who are vaccinated seem to have lower all cause mortality excluding Covid-19 than those who are unvaccinated. (Cite.) That is not surprising. There are several plausible explanations.

  1. The Covid-19 deaths that are not identified: Those will show up more among the unvaccinated. The first article linke provides some explanation for the reduced other cause mortality. Some of that “other cause” is the same cause, misidentified.
  2. Unaccounted harm from the disease: Many diseases leave those who survive more likely to suffer other diseases following. A recent study finds some evidence of this for Covid-19. (Cite.) Similar research years past found that those who had the flu vaccine were less likely  to suffer cardiovascular events. There is a sense in which a vaccine for a disease provides a lens for measuring the long-term harm a disease does.
  3. Unaccounted benefit from the vaccine: There have been vaccines past that had health benefits other than prevention of the disease targeted. I don’t know any evidence that is the case for these vaccines.
  4. Demographic and behavioral differences between the vaccinated and unvaccinated: There are all sorts of ways these groups might differ, that would lead to different health outcomes independent of Covid-19 infection. Those kind of confounders are a reason clinical trials need to be randomized.

Of course, there is no way to determine which of these or other explanations are the case, except by doing studies designed to surface such effects. That reflects a common pattern in statistics. And more generally. Experiment or data that shows a result thereby excludes some alternatives, and favors others. At the same time, it reveals a new layer of uncertainty. Smaller in some sense than the previous. Still large enough to matter.

Update: Judy Melinek, a pathologist, argues that there could be quite a bit of unaccounted mortality from the disease, due to long-term harm.

A suggestion for the new year…

December 27, 2021

We recently bought a new refrigerator, for the usual reasons that people do. The new one has a digitally controlled compressor. That has two advantages. First, it is quiet. Second, its peak draw is 3 amps, rather than 7 amps.

That caught my interest because of how we power our house during power outages after hurricanes and other storms. I use a 2kW pure sine wave inverter, powered off our Prius. The refrigerator is the large and necessary load. The stovetop is gas, so we don’t need electricity to cook or boil drinking water. The old refrigerator sometimes tripped the inverter, when its compressor kicked on and presented a new 800 watt load, on top of whatever else we were running. It will be nice to cut that by more than half.

But what were those other draws, that a 800 watt spike would exceed 2,000 watts? There is all the digital stuff: laptops, chargers for phones and watches, modems, and routers. I suspect the larger draw was our lights. Our house still had quite a few incandescent bulbs. Those are 60 watts or 75 watts each, with some circuits having six bulbs. That quickly adds up.

So after getting the new refrigerator, I went through our house and methodically replaced the incandescents with LED bulbs. Those are now inexpensive. There are both dimmable and instant varieties, in any temperature. They have such long lives you might never replace them again. Their light is nicer, to my eyes. And they draw a seventh the power. That 360 watt circuit now is a 54 watt circuit.

While I did this to make our lives easier during power outages, it has the obvious advantage of lowering our power usage year round. Since that exercise, our weekly draw has varied from 35kWh to 64kWh, depending on how much heat we used. Our first full month electric bill was for 305kWh. Despite the fact our rates recently bumped up, that came to $42.

Well, one data point doesn’t mean much. The bill will double next August, when air conditioning dominates our use. Those caveats given, I believe the frequent claim that the power savings from an LED bulb pays for itself in a year or less. Having gone through the exercise, I encourage doing the whole house in a one sweep. Figure out what light temperature you want in different rooms. Decide where you need dimmables. Buy the bulbs in packs. Get out the ladder and get it done.

An ethics debate in Congress

December 20, 2021

Though unlikely to gain traction soon and not much covered or noticed, there has been a serious ethics debate in Congress about the behavior of our elected officials. Alexandria Ocasio-Cortez and Elizabeth Warren again have pointed to Congressmen trading individual stocks as a kind of corruption. The former explains the issue:

It is absolutely ludicrous that members of Congress can hold and trade individual stock while in office. The access and influence we have should be exercised for the public interest, not our profit. It shouldn’t be legal for us to trade individual stock with the info we have.

Nancy Pelosi, who is well-known for having made some successful stock investments, snapped back:

We are a free market economy. They should be able to participate in that.

Former director of the Office of Government Ethics (OGE) Walter Shaub responded sharply:

This is the opposite of government ethics. Nobody kidnapped these members of Congress when they were private citizens, dragged them to Washington and forced them to be in Congress. The American people are sick of members of Congress buying and selling stock and creating the appearance of trading on insider information.

My view is that the corruption runs deeper than insider information. Holding individual stocks, something I have done for decades, influences one’s attention. An investor naturally reads on the industries and companies in which they are invested. That creates a kind of insider attitude even when all the information gleaned is public. Legislators, judges, and other civil servants constantly make decisions that intersect business interests. Even if they consciously try to do so in a way that does not favor their own interests, they cannot be objective, because their attention has been focused partly by that investment.

Cortez and Warren have focused on the behavior of members of Congress. That seems where to start first. Unless Congress can regulate itself in this regard, it cannot regulate the rest of government. Eventually I would like to see broader law: no employee of the federal government, from the president down to the postman delivering mail, should be invested in an individual business.

Importantly, the laws Cortez and Warren propose allow investment in index funds. It would be considerable financial imposition on civil servants to prohibit entirely their ability to benefit from equities. In rebuttal to Pelosi, the modern financial world provides tools to invest in American enterprise generally, without being biased by attachment to specific enterprise. Shaub has it right that when one goes to serve the public, that should exclude interest in particular business.

All of those named above support capitalism. They recognize publicly owned corporations and equity markets as a key part of that. None of them are Marxists or other kind of radical trying to tear that down. The argument is not about whether people should own stock. It is a narrow argument about public service and conflict of interest. All that should go without saying. Alas, it doesn’t, because of the amount of nonsense now popular. Which also is the reason that this important issue likely is irrelevant for the time being.

Medical progress on Covid-19

December 16, 2021

Paxlovid may prove to be a large step forward in treatment.

IndianaPeruIt should go without saying that prevention has large advantage over treatment. Which is why getting vaccinated should be almost everyone’s first line of defense. Peru, where Covid has struck hard, has had a difficult effort there. While some of that stems from poverty and access, it seems the same conspiracy theories that do their damage here also do their damage there.