Feynman

What Feynman says in this video is false.

Scientists Against Time

Back in WWII, air defense was difficult. You were shooting at high altitude, fast-moving targets. Missing was the most natural thing in the world. One out of thousands of anti-aircraft shells hit. So, there was a need for better methods.

The solution was the proximity shell – a shell with a tiny radar inside that would make the shell explode when it got _close_ to a target. This was technically challenging: it’s not easy to cram a radar into an artillery shell, not easy to build one that still works after experiencing accelerations of thousands of gravities. Moreover, people needed an answer _soon_.

The British had made some progress, but lacked sufficient resources. The US picked up the project, starting a lab that eventually became APL (Applied Physics Lab, associated with Johns Hopkins, which still exists today) . Its leader was Merle Tuve, who understood what the wartime priorities were:

” I don’t want any damn fool in this laboratory to save money. I only want him to save time.”

The New York Times, written & edited by orcs, for orcs, is criticizing Emergent Biosolutions for screwing up covid vaccine production – and those criticisms may well be valid. But they are also criticizing the original decision to throw money at Emergent Biosolutions for vaccine production – and I doubt if that is valid. The right thing to do, which by some odd chance we actually did, was ( besides getting rid of procedural obstacles) to try several vaccine approaches, several manufacturers, and use whatever was shown to work and could be produced rapidly. We didn’t know which companies would succeed ( Merck didn’t), so a shotgun approach was the logical way forwards.

The cost of the whole vaccine effort was nothing compared to the other costs of covid, while an effective vaccine was by far the most likely way of getting us out of this crappy situation.

The Silly Season

Inside a Battle Over Race, Class and Power at Smith College

In midsummer of 2018, Oumou Kanoute, a Black student at Smith College, recounted a distressing American tale:
She was eating lunch in a dorm lounge when a janitor and a campus police officer walked over and asked her what she was doing there.

The officer, who could have been carrying a “lethal weapon,” left her near “meltdown,” Ms. Kanoute wrote on Facebook, saying that this encounter continued a yearlong pattern of harassment at Smith.

Ms. Kanoute was determined to have eaten in a deserted dorm that had been closed for the summer; the janitor had been encouraged to notify security if he saw unauthorized people there. The officer, like all campus police, was unarmed. ”

Here, in what is likely the sign of some internal power struggle, the NYTimes is actually dissing some young moron who thinks that all the janitors of the world are out to get her, rather than honoring her lived experience

But there is a deeper significance: this must be the most boring story ever told. When I was bitten by a spider in my back yard and alternated between agony and a strange crisp energy, that was more interesting. When ants came over the wall from next door (where the lady lived who once worked on the Manhattan Project) and kidnapped our ants, that was more interesting.  When my youngest boy tried to crawl through the fence to the condo behind us and got his punkin head stuck, _that_ was more interesting.

When my Dad’s sister won the golf tournament at the country club but was denied the prize because she wasn’t a guy,  followed by my uncle Dean’s protest (crapping in every hole in the golf course just before he went off to the Army)  – that too was more interesting.

I would imagine that many of my readers have had experiences  ( possibly in the last half hour) even more worthy of coverage in the Paper of Record than Ms. Kanoute’s.

 

 

 

 

 

 

 

Black Doctors, Black Babies

There’s a paper out claiming that black infant infant mortality is much higher when they’re treated by white doctors, rather than black doctors.

Could it be that MCAT scores have negative predictive value?

No, there’s a simpler explanation: the report is nonsense.  A metaphorical cee-gar to the first person to explain why.

And the next question is: why do the pinheads that authored this paper have jobs?

A Random Walk on Scientists

A number of  epidemiologists and virologists  did not expect to see significant adaptive evolution for increased transmission in covid-19, and continue to argue against that hypothesis.  Vincent Racaniello, well-known virologist, takes this position.  So does  David Dowd, an infectious disease epidemiologist at Johns Hopkins.

They think that chance can drive a new variant with no transmission advantage to high frequency, even when there are many cases ( tens of thousands or more) .  The B.1.1.7 variant went from ~1% to a big majority of cases in England.

I’ve just run some sims ( with the sort of transmission dispersion observed in cov-19).  R (fitness)  = 1, in both cases.  I propagate 100 times in each run. Variant A starts out with 5,000 cases, B with 2000.  How often did B catch up with A? 6 out of 100 runs.

How often did B catch up with A when A started out with 50,000 cases and B with 20,000 cases? zero, out of 100 runs.

Start out with 2000 B and 50,000 A, same fitness: how often did B catch up? zero out of 100 runs.

Start out with 2000 B ( with a fitness of 1.05) and 50,000 A with a fitness of 1: How often did B catch up? 100 out of 100 runs.

 

The top curve shows the relative frequency of B.1.1.7 as a function of time in Denmark.  Dowd can look at that and believe it is a random fluctuation. Wow.

 

 

Lloyd Fredendall

Lloyd Fredendall was a general in the American Army in WWII, serving in North Africa.

He is known primarily for being a fuck-up. His early career may have been a sign: he dropped out of West Point twice.

He commanded II Corps in its advance into Tunisia, so his relative competence mattered.

He had a weird habit of talking in his own private slang. He called infantry units “walking boys” and  artillery “popguns.” Instead of using the standard military map grid-based location designators, he made up confusing codes such as “the place that begins with C.”  His subordinates had trouble understanding what the hell he was talking about.

He spent lots of effort building an underground fortress ( his headquarters) 70 miles behind the front lines  and spent most of  his time inside it, rather than visiting the front lines and talking with his commanders.

Tactically, also a mess: he split up units and scattered them widely. Which turned out poorly (Kasserine Pass).

After Kasserine Pass, Ike fired him. But how did Fredendall get anywhere in the first place, and why did removing him take so long?

Well, the most talented people didn’t much go into the American armed forces in those days, least of all the Army.  The Army wasn’t prestigious, wasn’t well-funded, wasn’t very meritocratic.  Promotion was slow, pay was lousy. The  US Army was about the size of the German Army while it was still obeying the Treaty of Versailles – but the Black  Reichswehr was an elite, taking only the best, secretly preparing for der Tag. Every sergeant was ready to be a captain.  The US Army was not like that.

The Army leadership all knew each other.  Most were West Pointers.  It was fairly easy-going.

Put to the test in WWII, we found out that our generals often weren’t very good. Ike himself had to learn an important lesson: how to fire people, including old friends. After a while American leadership became fairly good at that, for example when Nimitz fired Ghormley.

The Soviets already knew how to fire people ( sometimes with extreme prejudice) but Stalin learned to judge by performance and fire people intelligently: promote the winners, fire ( and sometimes execute) the losers. Act as if winning is the most important thing.

Generally, the governing classes in the US, for the last generation or two, has not acted as if they think that winning, actually achieving your goal,  is very important. Promotion follows failure: indeed, being right when almost everyone else is wrong just shows how undesirable you are.  Iraq is a good example.

Covid-19 is another example. The professionals weren’t very good, aren’t very good. They didn’t know a lot of important, knowable things. Probably the most talented people were going into something other than epidemiology or virology.

We don’t have to make them unpersons, don’t have to send them to Kolyma. We don’t have to pull out their teeth and fingernails.  There’s no reason to put on a black leather jacket and shoot them in the back of the head. That would be wrong.

But we can fire them.  And we should.

Let me count the ways

As many virologists have stated, their expectation was that the evolution of noticeably higher-transmission variants of Cov-19 was quite unlikely.

There is solid evidence that this has now happened at least three times (D614G, A222V, and B.1.1.7)  with at least two others likely ( in South Africa and Brazil).

They failed in an important aspect of their job.

 

 

 

 

 

 

Theory

Once upon a time, I was talking to a young engineer about some new wrinkle in solar concentrators.  He was enthusiastic: he thought that with a little effort, you could focus sunlight enough to generate a temperature higher than that of the Sun itself.

I said ” Nope. ”

Theory is your friend. Correct theory, that is.

New Guy in Town

These graphs show the results of evolutionary experiments by Richard Lenski, in which a bacterial species ( e coli) has been evolving under constant conditions for many years: tens of thousands of generations.

These bacteria were not perfectly adapted to the experimental environment, and so there is selection for changes that allow them to do better in these conditions. Adaptive change is rapid at first and slows down with time, as the culture approaches an optimum phenotype. Fitness increases rather like the logarithm of time.

The probability of a beneficial mutation fixing is proportional to the advantage it confers. Large-effect beneficial mutations are more likely to fix and dominate the early phase. As the bacteria get closer to an optimum, the possible gain from a beneficial mutation is smaller, and so those smaller-effect beneficial mutations ( the only ones possible) are less likely to fix. Thus they take longer to fix (on average they need to occur many times before succeeding) and they also fix more slowly, since their growth advantage is small.

relevance: a new virus in humans is like the situation near the origin of graph B.  The virus is not yet close to an optimum, so change is fairly rapid – particularly if the virus is infecting vast numbers of people ( like covid-19) which greatly increases the number of copies of the virus and thus the chance of favorable mutations ( Fisherian acceleration). Favorable to the virus, that is.

An old virus in humans, say measles ( > 1000 years old)  is closer to an optimum: change is much slower.

It seems that most professional virologists are used to viruses that have been around for quite a while – understandable, since new viruses do not sweep through the human race every year.

You could have predicted the emergence of new higher-transmission variants of covid-19 from this theoretical perspective. I did, arguablywrong did, probably others have as well. But virologists did not.

 

 

 

 

New variants

There’s a new variant of cov-19, the UK strain, that is something like 70% more infectious than previous strains. At this point its lethality seems about the same as earlier strains.

Virologists, most of them, did not expect this.  I did.

Let’s consider some simple examples. Imagine that you have someone – a single individual – with what we will call a normal strain, A. On average, in  the current situation,  it has an R0 well above 1.

What is the most likely outcome? Most likely, he doesn’t spread it to anyone, and it dies out.  Overdispersion: most people don’t do much spreading, and a few do a lot. Let’s say that 20% of those infected do all of the spreading.  Right off the bat, 80% of new strains die out, just because of this pattern.

Now, imagine that we simultaneously introduce two new strains, A  and B with a 50% greater R0%.  for each, a single individual.

There are four possible outcomes: A spreads widely, B spreads widely, both and A & B spread widely , both A and B disappear.

Most likely both will disappear (~64% chance) .

There’s a fair chance that A will spread while B is lost, and a moderately larger chance that B will spread while A is lost.

There is essentially zero chance that both will spread widely: even if both manage to avoid being lost by chance in the beginning, B will grow faster than A and replace it.

So, suppose you introduce one person with A, and one person with strain B: can you judge the relatively infectivity by which one succeeds?  No – there’s a significant  chance that the less-infective one will win out.

Now consider a situation in which A is already common, and a single case of B  is introduced.  what are the possible outcomes?

1.  B is lost by chance.  ( > 80% probability)

2.  B replaces A – happens if B is lucky enough to get past the risk of extinction when rare.  But once it gets up to a few hundred copies, it will surely replace A.

What can we conclude if B is rapidly replacing A ( as has been the case with the new UK strain)?

That it surely has significantly higher transmission, significantly higher R0.

Many virologists thought this very unlikely, and some said that you could never know that a new variety had higher transmission from mere incidence data: you must understand the biological mechanism.  Are they correct?  Obviously not.

Why did they think that a new, more transmissible variant of COVid-19 was unlikely?  I would say there are several reasons. One, they typically deal with viruses that have been around for a long time, like measles ( > 1000 years) .  An old virus is going to be pretty well-adapted to to humans.  Probably it’s at a local optimum, where small changes would reduce infectivity. But you don’t expect that high degree of optimization in  a virus that’s brand new in humans: while spreading to very many people, more than 100 million,  greatly increases the chance of  transmission-increasing mutations.  Fisherian acceleration.

Like most biologists and MDs, most virologists don’t know any theory, and in fact don’t _believe_ in theory.   For this they occasionally pay a price.