Category: Science

The performance of the NIH during the pandemic in 2020

“A new research study by one of us and his Johns Hopkins colleagues found that of the $42 billion the National Institutes of Health spent on research last year, less than 2% went to Covid clinical research…

Here is the WSJ source.  Here is the full report from Johns Hopkins, and here is the executive summary:

● Of the $42 Billion 2020 NIH annual budget, 5.7% was spent on
COVID-19 research
● Public health research was underfunded at 0.4% of the 2020 NIH
● Only 1.8% of the 2020 NIH budget was spent on COVID-19 clinical
● Average COVID-19 NIH funding cycle was 5 months
● Aging was funded 2.2 times more than COVID-19 research
● By May 1, 2020, 3 months into the pandemic, the NIH spent 0.05%
annual budget on COVID-19 research
● Of the 1419 grants funded by the NIH:
• NO grants on kids and masks specifically
• 58 studies on social determinants of health
• 57 grants on substance abuse
• 107 grants on developing COVID-19 medications
• 43 of the 107 medication grants repurposed existing drugs

Ouch.  Here is a not entirely random sentence from the report:

The COVID-19 pandemic has only exacerbated the NIH institutional challenges and inability to reallocate funds quickly to
critical research.

Here is another damning sentence, though it damns someone other than the NIH:

…to date, no research has investigated NIH COVID-19 funding patterns to the best of our knowledge.

Double ouch.  Might the NIH have too much influence over the allocation of funds to be investigated properly?  Rooftops, people…

Sentences to ponder

…participants who trust science are more likely to believe and disseminate false claims that contain scientific references than false claims that do not. Second, reminding participants of the value of critical evaluation reduces belief in false claims, whereas reminders of the value of trusting science do not. We conclude that trust in science, although desirable in many ways, makes people vulnerable to pseudoscience.

Here is the full paper, via the excellent Kevin Lewis.  That is only one (non-replicated) paper, but the point is worth keeping in mind nonetheless.

Within one hundred years the future is going to get very weird

That is the theme of my latest Bloomberg column, here is one excerpt:

The notion that the future will be weirder than we think, and come sooner, is a possibility raised by Holden Karnofsky, the co-chief executive officer of Open Philanthropy. It’s an intriguing and provocative idea.

I consider genetic engineering, longevity research, finding signs of life on other planets, neural engineering, and AI as possible developments, plus a bit more.

…these changes are far more radical than those that occurred between 1921 and today. Compared to 1921, we are much wealthier and more secure — but a lot of basic structures of the world remain broadly the same. I don’t think that much of what we can do now would strike our 1921 predecessors as magical, though the speed and power of our computers might surprise them. Nor would visitors from 1921 think of us as somehow not human.

Of course none of these developments are inevitable. Another very weird future is entirely possible: that we humans use our creative energies for destruction, causing civilization to take some major and enduring steps backwards.

Either way, the future is not just more and nicer suburbs, better pay and new forms of social media. All those are likely to happen, but they won’t be the biggest changes. When it comes to the future of the human race, we — and our children, for those of us who have any — may turn out to be especially important generations. I very much hope we are up to this moment.


Our regulatory state is still, still, still failing us

The U.S. agency leading the fight against Covid-19 gave up a crucial surveillance tool tracking the effectiveness of vaccines just as a troublesome new variant of the virus was emerging.

While the Centers for Disease Control and Prevention stopped comprehensively tracking what are known as vaccine breakthrough cases in May, the consequences of that choice are only now beginning to show.

Here is more from Bloomberg, tragic and stupid throughout.

A Step Closer to General AI

From Google’s Deep Mind:

In recent years, artificial intelligence agents have succeeded in a range of complex game environments. For instance, AlphaZero beat world-champion programs in chess, shogi, and Go after starting out with knowing no more than the basic rules of how to play. Through reinforcement learning (RL), this single system learnt by playing round after round of games through a repetitive process of trial and error. But AlphaZero still trained separately on each game — unable to simply learn another game or task without repeating the RL process from scratch.

…Today, we published “Open-Ended Learning Leads to Generally Capable Agents,” a preprint detailing our first steps to train an agent capable of playing many different games without needing human interaction data. We created a vast game environment we call XLand, which includes many multiplayer games within consistent, human-relatable 3D worlds. This environment makes it possible to formulate new learning algorithms, which dynamically control how an agent trains and the games on which it trains. The agent’s capabilities improve iteratively as a response to the challenges that arise in training, with the learning process continually refining the training tasks so the agent never stops learning. The result is an agent with the ability to succeed at a wide spectrum of tasks — from simple object-finding problems to complex games like hide and seek and capture the flag, which were not encountered during training. We find the agent exhibits general, heuristic behaviours such as experimentation, behaviours that are widely applicable to many tasks rather than specialised to an individual task. This new approach marks an important step toward creating more general agents with the flexibility to adapt rapidly within constantly changing environments. (Bold added, AT).

Hat tip: Daniel Kokotajlo at Less Wrong who notes “This seems like a somewhat big deal to me. It’s what I would have predicted, but that’s scary, given my timelines.” See also the LW comments.

In other news, South Africa awarded the first ever patent to an AI.

Do looks matter for an academic career in economics?

It seems they do:

We document appearance effects in the economics profession. Using unique data on PhD graduates from ten of the top economics departments in the United States we test whether more attractive individuals are more likely to succeed. We find robust evidence that appearance has predictive power for job outcomes and research productivity. Attractive individuals are more likely to study at higher ranked PhD institutions and are more likely to be placed at higher-ranking academic institutions not only for their first job, but also for jobs as many as 15 years after their graduation, even when we control for the ranking of PhD institution and first job. Appearance also predicts the success of research output: while it does not predict the number of papers an individual writes, it predicts the number of citations for a given number of papers, again even when we control for the ranking of the PhD institution and first job. All these effects are robust, statistically significant, and substantial in magnitude.

That is from a recent paper by Galina Hale, Tali Regev, and Yona Rubinstein.  Via John Chilton.

They modeled this — why women might see fewer STEM ads

Women see fewer advertisements about entering into science and technology professions than men do. But it’s not because companies are preferentially targeting men—rather it appears to result from the economics of ad sales.

Surprisingly, when an advertiser pays for digital ads, including postings for jobs in science, technology, engineering and mathematics (STEM), it is more expensive to get female views than male ones. As a result, ad algorithms designed to get the most bang for one’s buck consequently go for the cheaper eyeballs—men’s. New work illustrating this gap is prompting questions about how that disparity may contribute to the gender gap in science jobs.

…As a result of that optimization, however, men saw the ad 20 percent more often than women did…

Tucker ran $181 worth of advertising via Google, for example, saying she was willing to pay as much as 50 cents per click. It ended up costing 19 cents to show the ad to a man versus 20 cents to show that same ad to a woman. These investments resulted in 38,000 “impressions”—industry-speak for ad views—among men, but only about 29,000 impressions among women.

Similarly, on Twitter it cost $31 to get about 52,000 impressions for men but roughly $46 to get 66,000 impressions for women. And on Instagram it cost $1.74 to get a woman’s eyeballs on the ad but only 95 cents to get a man’s.

Here is the full Scientific American article, via Luke Froeb, and do note those differentials may vary considerably over time.  Gender issues aside, I would say this reflects a broader problem with having a very high value of time — it becomes harder to maintain a relatively high proportion of people showing you valuable things you wish to see (as opposed to people bugging you, grifting you, etc.).

Second Doses Are Better at 8 Weeks or Longer

In Britain people are now being warned *not* to get their second dose at 3 or 4 weeks because this offers less protection than waiting 8 weeks or longer.

Warnings over the lack of long-term protection offered by jab intervals shorter than eight weeks come as scores of under 40s continue to receive second doses early at walk-in clinics, contrary to Government guidance.

…“There is very good immunological and vaccine effectiveness evidence that the longer you leave that second dose the better for Pfizer and eight weeks seems to be a reasonable compromise.”

Professor Harnden emphasised that “you’re definitely less protected against asymptomatic disease if you have a shorter dose interval”.

I’m so old I can remember when first doses first wasn’t “following the science.”

Scientists discover spiders are eating snakes all over the world

“They can outfight snakes 10 to 30 times their size,” says the University of Basel in Switzerland.

According to a new meta-analysis study, perhaps snakes should be fearful of spiders. It seems arachnids like to chow down on the reptiles, all over the world.

The study, published in The Journal of Arachnology, has the straightforward title Spiders (Arachnida: Araneae) Feeding on Snakes (Reptilia: Squamata). The researchers looked at 319 reports of spiders feeding on snakes from every continent except Antarctica. Most of the events occurred in the US and Australia.

The data showed that spiders representing 11 different families have been observed eating snakes. “That so many different groups of spiders sometimes eat snakes is a completely novel finding,” lead author and arachnologist Martin Nyffeler said in a Monday news release from Switzerland’s University of Basel. Guess we can add the discovery to our Surprising Insect Dining Habits file, alongside praying mantises eating hummingbird brains.

Here is the full story, via Shaffin.

Use Fractional Dosing to Speed Vaccination and Save Lives

I’ve been shouting about fractional dosing since January, most recently with my post A Half Dose of Moderna is More Effective Than a Full Dose of AstraZeneca and the associated paper with Michael Kremer and co-authors. Yesterday we saw some big movement. Writing in Nature Medicine, WHO epidemiologists Benjamin Cowling and Wey Wen Lim and evolutionary biologist Sarah Cobey title a correspondence:

Fractionation of COVID-19 vaccine doses could extend limited supplies and reduce mortality.

Exactly so. They write:

Dose-finding studies indicate that fractional doses of mRNA vaccines could still elicit a robust immune response to COVID-192,3. In a non-randomized open-label phase 1/2 trial of the BNT162b2 vaccine, doses as low as one third (10 μg) of the full dose produced antibody and cellular immune responses comparable to those achieved with the full dose of 30 μg (ref. 4). Specifically, the geometric mean titer of neutralizing antibodies 21 days after the second vaccine dose was 166 for the group that received 10 μg, almost the same as the geometric mean titer of 161 for the group that received 30 μg, and 63 days after the second dose, these titers were 181 and 133, respectively4. For the mRNA-1273 vaccine, a dose of 25 μg conferred geometric mean PRNT80 titers (the inverse of the concentration of serum needed to reduce the number of plaques by 80% in a plaque reduction neutralization test) of 340 at 14 days after the second dose, compared with a value of 654 for the group that received the standard dose of 100 μg (ref. 5). According to the model proposed by Khoury et al.6, if vaccine efficacy at the full dose is 95%, a reduction in dose that led to as much as a halving in the post-vaccination geometric mean titer could still be in the range of 85–90%. Although other components of the immune response may also contribute to efficacy, these dose-finding data are at least indicative of the potential for further exploration of fractionation as a dose-sparing strategy. Durability of responses after fractional doses should also be explored.

…Concerns about the evolution of vaccine resistance have been posited as a potential drawback of dose-sparing strategies. However, vaccines that provide protection against clinical disease seem to also reduce transmission, which indicates that expanding partial vaccination coverage could reduce the incidence of infection. As described in a recent paper, lower prevalence should slow, not accelerate, the emergence and spread of new SARS-CoV-2 variants8.

…In conclusion, fractionated doses could provide a feasible solution that extends limited supplies of vaccines against COVID-19, which is a major challenge for low- and middle-income countries.

Also a new paper in preprint just showed that 1/4 doses of Moderna create a substantial and lasting immune response on par with that from natural infection.

Here we examined vaccine-specific CD4+ T cell, CD8+ T cell, binding antibody, and neutralizing antibody responses to the 25 ug Moderna mRNA-1273 vaccine over 7 months post-immunization, including multiple age groups, with a particular interest in assessing whether pre-existing crossreactive T cell memory impacts vaccine-generated immunity. Low dose (25 ug) mRNA-1273 elicited durable Spike binding antibodies comparable to that of convalescent COVID-19 cases. Vaccine-generated Spike memory CD4+ T cells 6 months post-boost were comparable in quantity and quality to COVID-19 cases, including the presence of TFH cells and IFNg-expressing cells.

Finally, an article in Reuters notes that Moderna are preparing to launch a 50 ug dose regimen as a booster and for children. Thus, contrary to some critics of our paper, the technology is ready.

Frankly, governments are way behind on this–they should have been pushing the vaccine manufacturers and funding trials on alternative dosing since at least January. Indeed, imagine how many lives we might have saved had we listened to Operation Warp Speed advisor Moncef Slaoui who advocated for half doses in January. On a world scale, we could have vaccinated tens even hundreds of millions more people by now had we ramped up fractional dosing.

At this point, it’s my view that there is enough knowledge to justify rolling out alternative dosing in any hot spot or in any country worried about outbreaks. Roll it out in a randomized fashion (as Kominers and I discussed in the context of the US vaccination rollout) to study it in real time but start the roll out now. Lives can be saved if we speed up vaccination, especially of the best vaccines we have, the mRNAs. Moderna and Pfizer have together pledged to deliver (mostly Pfizer and mostly through the US) some 250m vaccine doses to COVAX in 2021 for delivery to less developed countries. If we go to half-doses that becomes 500m doses–a life saver. And recall these points made earlier:

Judging by neutralizing antibodies, a 50 ug dose of, for example, Moderna looks to be more effective than standard dosing of many other vaccines including AZ and J&J and much better than others such as Sinovac. Thus alternative dosing is a way to *increase* the quality of vaccine for many people.

A 50 ug dose vaccine available today is much higher quality than a 100 ug dose vaccine available one year from now.

If we have the will, we can increase vaccine supply very rapidly.

Facts and uncertainties about ear wax

Our attitude to ear wax is in some ways surprising. A review of impacted ear wax estimates that 2.3 million people a year in the United Kingdom suffer problems with wax needing treatment, with some 4 million ears being syringed annually.2 This makes it possibly the the most common therapeutic procedure carried out on any part of the body. Symptoms of excessive wax or impaction, especially in the elderly, include not only hearing loss but tinnitus, dizziness, infections, social withdrawal, poor work function and mild paranoia. Other problems include general disorientation and loss of an aural sense of direction. With unilateral wax, sounds can appear to be coming from the wrong side, leading to accidents as a driver or especially as a pedestrian. Inappropriate self-treatment (or even treatment by health professionals) can cause perforated eardrums and in very rare cases cochlear damage, leading to nystagmus and sensorineural deafness. In spite of this catalogue of harms, the clinical profile and management of excessive wax are poorly understood. The evidence base is poor and inconsistent, leading to few strong recommendations, even relating to the most commonly used treatments.

Low esteem for ear wax is surprising in other ways too. As a substance, it is unique in the human and mammalian body. This is due to its position in our sole anatomical cul-de-sac. Everywhere else on our body surface, dead and redundant skin cells fall off or are scrubbed away when we wash. In the ear canal – which points forwards and downwards and might otherwise turn into a dermatological garbage dump – ear wax binds these together, along with other assorted detritus that may have entered from the world outside. It is then moved up to the exit by jaw movements and as a result of the skin of the canal slowly moving outwards like an escalator. Wax also prevents multiplication of micro-organisms and infection. It is as essential as sweat and tears, although perhaps not quite as vital as blood. Wax is also fascinating in its own right.

Imagine an ear wax post that is not solely about Q-tips! (Have you ever wondered why they have to be so dangerous?  Can’t you just put them in a little way?  Or is there some indivisibility here?  I have never understood the anguished warnings here.  If you are not using Q-tips at all, you only have to put them in a little way to pull out a lot of earwax, right?  Solve for the equilibrium!)

Here is more by John Launer, about ear wax throughout, via Michelle Dawson.

Nematodes Maximize Expected Utility on the heels of the new paper showing that the trading behavior of mycorrhizal fungi is consistent with the predictions of general equilibrium theory we have that nematodes obey the generalized axiom of revealed preference. It would be amusing if economics turns out to work well everywhere except for humans.

Abstract: In value-based decision making, options are selected according to subjective values assigned by the individual to available goods and actions. Despite the importance of this faculty of the mind, the neural mechanisms of value assignments, and how choices are directed by them, remain obscure. To investigate this problem, we used a classic measure of utility maximization, the Generalized Axiom of Revealed Preference, to quantify internal consistency of food preferences in Caenorhabditis elegans, a nematode worm with a nervous system of only 302 neurons. Using a novel combination of microfluidics and electro-physiology, we found that C. elegans food choices fulfill the necessary and sufficient conditions for utility maximization, indicating that nematodes behave exactly as if they maintain, and attempt to maximize, an underlying representation of subjective value. Food choices are well-fit by a utility function widely used to model human consumers. Moreover, as in many other animals, subjective values in C. elegans are learned, a process we now find requires intact dopamine signaling. Differential responses of identified chemosensory neurons to foods with distinct growth potential are amplified by prior consumption of these foods, suggesting that these neurons may be part of a value-assignment system. The demonstration of utility maximization in an organism with no more than several hundred neurons sets a new lower bound on the computational requirements for maximization, and offers the prospect of an essentially complete explanation of value-based decision making at single neuron resolution.

Photo Credit: By Agricultural Research Service – source (15016 KB); Description page, Public Domain,

Hat tip: Derek Lowe.

My Conversation with Richard Prum

Prum is an ornithologist at Yale, here is the audio, video, and transcript. Here is part of the summary:

Richard joined Tyler to discuss the infidelity of Australian birds, the debate on the origins of avian flight, how the lack of a penis explains why birds are so beautiful, why albatrosses can afford to take so many years to develop before mating, the game theory of ornithology, how flowers advertise themselves like a can of Coke, how modern technology is revolutionizing bird watching, why he’s pro-bird feeders yet anti- outdoor cats, how scarcity predicts territoriality in birds, his favorite bird artist, how Oilbirds got their name, how falcons and cormorants hunt and fish with humans, whether birds exhibit a G factor, why birds have regional accents, whether puffins will perish, why he’s not excited about the idea of trying to bring back passenger pigeons, the “dumb question” that marks a talented perspective ornithologist, and more.

Here is one excerpt:

COWEN: Putting path dependence aside, if you were trying to give us the most fundamental explanation of why sexual dimorphism is different in birds compared to mammals, what would that be?

PRUM: Well, that’s actually a really big question. [laughs]

COWEN: Of course, but the most fundamental factor — what is it?

PRUM: The most fundamental factor is that most birds don’t have a penis.

COWEN: Talk me through the equilibrium there.

PRUM: [laughs] There’s a lot. That’s where we start: Most birds don’t have a penis, which means that one of the things that happens in avian evolution that’s distinct from mammals is that the kids require a lot of care. They’re growing up in the nest, they’re hatching out of an egg, but they’re very, very vulnerable until they can fly.

Birds have a very rapid period of rapid development. That means that they grow up and leave the nest, and you need two parents to do that efficiently in most diets or most kinds of ecologies. That means the dad’s got to be at the nest.

We usually thought that you have social monogamy, at least two birds helping raise the young, because the young are so needy and they have to grow up quickly. But there’s another possibility, which is that they could evolve to be so needy and grow up quickly because they managed to get males at the nest.

One of the things that happened in the phylogeny of birds — you’ve got ostriches and their relatives, and you’ve got chickens and ducks, and then you’ve got the rest of birds, and that’s a bunch. That’s the vast majority of them, and in that lineage leading to the rest of birds, the penis evolved away, and the question is why. My own theory is that female birds preferred mates that did not have a penis.

One of the ancillary benefits of that, one of the correlated benefits of that is that they were no longer subject to sexual coercion or sexual violence. They could be coerced behaviorally, but they couldn’t be forcibly fertilized. That means that they have freedom of choice, and what do they do with their freedom of choice? They choose beauty. One of the reasons why birds are so beautiful is that males don’t have a penis. They have to be subject to choice in order to effect reproduction, and also they have to invest if females require it.

COWEN: Now, sometimes albatrosses don’t breed until they’re 20 years old or even, on average, maybe it’s what — 10 years old. What are they doing in the meantime that’s so important?

PRUM: Well, that is a deep question.

Recommended, this was one of my favorite CWT episodes.

A problem in nuclear waste semiotics

Via Richard Harper, this thread asks how you might warn very future people away from a nuclear waste site. Alexandra Erin wrote:

An easy way to understand the problem of nuclear waste storage semiotics is to imagine what kind of warning could have been on an Egyptian tomb that would have kept Howard Carter from robbing it.

Here is some background material on how people are thinking about the problem at Yucca Mountain.  Here is a Wikipedia page on different signs and options.  Alex suggests color-changing cats.

I would think the question of how to inform a super-advanced civilization is a manageable one, at least if they have any patience at all.  Simply explain the whole truth in plain English, and give them enough English text, in durable micro form if needed, so they can unlock the secrets of English.  Also put up some images of radioactive decay.  Skull and crossbones may not mean so much to them.

What about our possible “Mad Max” descendants?  Of course that scenario means our own civilization has in some manner perished, so it is not a totally optimistic prognosis for human prudence.  So why think some silly red signs will make much of a difference?  After all, just try today to talk people out of alcohol.  Good luck.

So instead my mischievous thoughts turn to finance theory and portfolio diversification.  If the nuclear waste site is truly remote and previously unobserved and undiscovered, why not put something really good in there as recompense?

A seed bank.  Copies of The Great Books.  The text of the United States Constitution.  Proofs of Newton’s Laws.  Einstein’s theory of relativity (maybe wait on that one?…)  Design for a better medieval water wheel.  Compositions of Beethoven and Mozart.  Translation advice, some of it pictorial.  And so on.  Surely some of it will be useful, sooner or later.

Which is further reason why all of your ideas are less likely to work.  You can’t credibly commit to not giving people insurance against their bad decisions — just ask the Fed!

The economics of peer review

Springer-Nature, the Anglo-German publisher of the world’s leading scientific journal Nature, announced in 2017 that in some of its publications it was censoring articles that used words like “Taiwan”, “Tibet” and “cultural revolution”, when printing in China.

In April 2020 Nature ran an editorial apologising for its “error” in “associating the virus with Wuhan” in its news coverage.

…The magazines that publish scientific papers have become increasingly dependent on the fees that Chinese scientists pay to publish in them, plus advertisements from Chinese firms and subscriptions from Chinese institutions. In recent years observers have noticed that the news coverage of China in these magazines has begun to look a little less objective than it once did.

Here is the full Matt Ridley piece.