Readings: Alzheimer’s, Real estate, Tall people, Gods, and Complex societies


Alzheimer’s drug research is, charitably, even more of a mess than it was earlier this week. And, given the mess it was already in, that’s not easy. Nevertheless, that was the main takeaway from yesterday’s news that Biogen and its partner Eisai were halting two clinical trials of their anti-amyloid therapy — aducanumab — because, in short, neither worked.

In case you are keeping score at home, there are now more than 300 failed Alzheimer’s drugs, which is remarkable in itself, and doubly so when you consider that most of these drugs are based on a single theory, the so-called “amyloid hypothesis”.

And what is the amyloid hypothesis? Here is a summary from Science on this now decades-old theory that isn’t working very well:

[Alzheimer’s] is caused by deposition of amyloid β-peptide (Aβ) in plaques in brain tissue. According to the amyloid hypothesis, accumulation of Aβ in the brain is the primary influence driving AD pathogenesis. The rest of the disease process, including formation of neurofibrillary tangles containing tau protein, is proposed to result from an imbalance between Aβ production and Aβ clearance.

In short, to this way of thinking, AD is a sort of dental plaque on brain tissue. If enough forms, and you don’t scrape it off and prevent more from forming, bad things happen to brain function. And there is plenty of evidence of amyloid proliferation in the brains of AD patients, so the theory has that going for it. Further, there is an obvious mechanical connection that can be made between brain function and the presence of amyloid plaques. So, you can see the appeal of amyloid theory.

The trouble is, as is the case so often in medicine, a bioplausible mechanistic cause, when treated,  doesn’t change the outcome. Knees with meniscus tears can be painful, but the tear is often not the cause, etc. Experimental drugs have, to varying degrees, inhibited the formation of amyloid plaques to no avail, with patient outcomes largely indistinguishable from those of untreated patients. 

Rather than diving deeper into the science, however, I’m more intrigued, at least philosophically, by why the amyloid hypothesis hasn’t been abandoned. Are 301 trials required? 302? What constitutes plausible negative evidence that would cause researchers at pharma companies to abandon this bioplausible theory of AD causation and progression? 

In broad terms, this is a deep question, one that has puzzled scientists and philosophers for as long as both have existed. Perhaps the best-known example of an attempt to answer the question is that of Karl Popper, who, to vastly simplify his model, proposed that science proceeds through definite tests and falsification. All positive results are provisional, capable of being demolished if the right test comes along that falsifies a key element of the research.

The trouble is, of course, that isn’t really the way science works, as aducanumab and its predecessors’ failures show. If the amyloid hypothesis could be overturned by a well-crafted negative experiment (or 300), one would think that would have happened by now, which it demonstrably hasn’t. And it isn’t for want of bioplausible alternative approaches to Alzheimer’s therapy, of which there are now many. 

Thomas Kuhn would argue that this is exactly what we should expect in a field as inherently pack-oriented and emotional as science. He believed that science is a social endeavor and that one of the primary reasons why hypotheses aren’t abandoned, despite overwhelming evidence of their wrongness, is that scientists hold tightly to cherished beliefs, despite them having been falsified over and over again. 

His is not the only non-Popperian view, of course. Imre Lakatos vomited, a little, at Kuhn’s idea that the best way to understand science was as a kind of trend-driven teen fashion. Lakatos argued, instead, that research programmes should be thought of as “progressive” or “degenerating”, and no single experiment (or even collection of experiments) could turn a progressive programme into a degenerating. Instead what happened, to his way of thinking, was the scientists former “protective belts” around the “hard core” of the theory, leaving its main predictions intact in a way that allowed continued research to be conducted. 

This is where we are at, I think, with the amyloid hypothesis. There is still a hard core to the programme, with no-one denying the connection between plaques and AD, but there is increasing uncertainty about what to do with that information. As one researcher said this week, it is like “…blowing out a match (amyloid plaques) after the forest (the brain) is already burning (suffering the death of neurons and destruction of synapses)”.

To this way of thinking, amyloid-targeting therapies would work, if only they were delivered sooner. Fair enough, but short of having asymptomatic patients at high risk on such therapies from age 50 onward, it is hard to imagine where one goes from here. 





Readings: Significance, Kale, bankruptcy, and streaming media


[I]t will become clear that there is only one principle that can be defended under all circumstances and in all stages of human development. It is the principle: anything goes.
― Paul Feyerabend

Anything goes … means … everything stays.
― Alan Chalmers

It’s the best possible time to be alive, when almost everything you thought you knew is wrong.
― Tom Stoppard, Arcadia

The philosophy of science is a strange place in academia. To the extent that it is philosophy it isn’t science, and to the extent that it is science, it isn’t philosophy. Perhaps the only thing that both philosophers and scientists can agree on is that they can’t stand philosophers of science, and they wish they didn’t get so much goddamn attention for something that doesn’t actually work. We just don’t have a comprehensive theory for how science works; we just know that it works better than the alternatives, kinda, most of the time. 

I was reminded of it again this week in reading the noisy discussion about a recent Nature commentary on the merits of retiring the 0.05 significance hurdle in publishing research findings.  There are compelling arguments to be made about getting rid of an arbitrary hurdle, like 0.05, especially one so easily gamed in large data sets, especially of the sort of that pharmaceutical companies marshall to get expensive drugs with minimal effect past the U.S. Food & Drug Administration. Hey, the tests’ results were significant, c’mon. 

This, of course, is not unique to medical statistics. It is a restatement, of sorts, of Goodharts’ Law, that, “Any observed statistical regularity will tend to collapse once pressure is placed upon it for control purposes”. Once we have a measure nailed, in other words, and it is used to judge and control processes, especially a human-centric one, it ceases being useful. This is the dilemma we face in medical statistics, where, to a first approximation, we have little idea whether products whose effects are deemed statistically significant actually have significant effects. As Goodhart would say, the measure of significance has collapsed. 

But with what do we replace it? SImply cautioning people against blind fealty to 0.05 significance isn’t enough, as decades of academic research have shown, where faculty slavishly still report the measure, despite much private eye-rolling, knowing full well that without it they won’t get published. And having no hurdle comes with its own problems, as Chalmers’ rejoinder to Paul Feyerabend highlights. While anything may “go” in science — if said research works — if it truly goes then we have no basis for deciding what matters and what works — other than what works, of course, which may turn out to be all we really had in the first place.

Are we genuinely prepared for a 0.05-less world of research? Or do we prefer to continue with statistical thumbsucking and gameable measures? My bet is that very little changes. 





From the German youth version of The Voice (who knew such things existed?), two sisters perform Radiohead’s Creep, to remarkable effect. It is yowling, dissonant, and dangerous, like something that should be played over the end credits in Fight Club II as the world tumbles into itself.

Readings: The Worried Well, Awe, Game Engines, and Farming


We are getting healthier, wealthier, and we live longer, but we are more worried than ever. This, in large part, is the paradoxical message of a new OECD survey, The Risks That Matter, on what people around the world are worried about. 

Here is a key graph:

The situation is more or less the same around the world: We worry. A lot. Especially about health, especially about health as we live in healthier societies. It is remarkable, especially when you consider that most health trends are getting better in most places, with, for example, twenty-year trends in heart outcomes showing sharply reduced likelihood of death. We are also better than ever at understanding the huge impact on health of even small doses of activity, like biking to work

Granted, not everything is in our favor. For example, the food industry isn’t exactly going out of its way to make food healthier, with a recent study showing a sharp multi-decadal increase in fast food portion size, caloric load, and macronutrients (read: bad things). And we are no better than ever at detecting causality when something seems (wrongly) to be good for us (the case of multivitamins), or seems (wrongly) to be bad for us (the case of eggs), or even at figuring out whether data on the topic is fraught, useful, or just a mess.

Part of the problem, yes, is innumeracy, an inability to think critically about arguments or to understand risks in a practical sense. And, yes, there is some truth to the Rosling/Pinker argument that the world is getting better, and we just aren’t paying proper attention to the right graphs. But that’s sort of beside the point. It’s not merely that we don’t notice, it’s that we choose not to care given our worries, however unfounded. 

You can see this in the latest DSM manual, the fifth edition, where the fastest-growing section is on somatic disorders. These are things that we worry about, but for which there is no clinical data showing actual illness. We are, increasingly, to use a favorite medical phrase, the “worried well”: healthy people who worry about being ill because being ill is the thing we are most worried about, given the steady disappearance of other things to worry about, whether descending deities, predators outside the cave, or imminent economic collapse. Instead, we worry about that pain behind the left eye, the troubling feeling in your stomach, your inability to sleep the way you once did, and so on.

It doesn’t help, of course, that we are overwhelmed with data on all the ways we aren’t living in the healthiest possible ways. And even that data is highly fraught, as the huge Apple Heart Study just showed, with uncertain benefits and unknowable costs. People’s bodies, if observed over a long enough timeframe, will almost always do something weird, but nothing that can’t be made worse by a quick trip to the ER. A similar situation applies in almost everything we are busily instrumenting, from weight to sleep patterns: we creating new cohorts of somatic symptom sufferers.

After many decades of illness understandably triumphing over health in books, it is strange to see health unseated again recently by illness, as this Google Ngrams graph shows.  

You can see echoes of this in Derek Thompson’s excellent recent Atlantic essay on how “work-ism” is making Americans miserable. Having increasingly replaced religiosity with economics-osity, Americans are finding that this new system is even less satisfying. Rather than delivering transcendence, as promised, people are growing disillusioned and sad, even in Silicon Valley. 

“The best minds of my generation are thinking about how to make people click ads.” Jeffrey Hammerbacher

We worry because we are struggling with what to worship. As David Foster Wallace said in his “This is Water” speech years ago, worship power, and you become fearful; worship your body and health, and you always feel ugly and unhealthy. Or, at the very least, as Wallace writes, “… you will die a million deaths before they finally plant you”. People who sit around with nothing to worry about, simply worry about different things, and do it more intensely, and more dangerously. 

“Everybody worships. The only choice we get is what to worship.” David Foster Wallace




Readings: Trade, 8s, Kangaroos, Liverpool, and The Simpsons


Free-er Trade

While no-one apparently cares anymore as we lurch toward isolationism and xenophobia, free-er trade remains among the free-est of free economic lunches. That’s the conclusion in an important new study with this remarkable claim: fully liberalizing migration would increase welfare about threefold and would significantly affect the evolution of particular regions of the world.

Granted, these are economists saying this, so they would say that. But the study is compelling, even if highly analytical and model-based, with resulting limitations (about which it is upfront to the point of almost arguing with itself now and then). More broadly, the ideas are rich and provocative, like that under trade liberalization and free movement, Venezuela, Brazil, and Mexico would become some of the most productive regions of the world, and sub-Saharan Africa could rise to first world status by 2090. It’s a remarkable and unsettling piece of work — which will almost certainly affect nothing, other than maybe inspire some excellent SciFi stories.

Accounting Performance Art

I am intrigued by the PwC settlement today with the FDIC over Colonial Bank audits back during the financial crisis. I often argue that accountancy is largely a pomo performance art, sort of self-referential financial theater. This seems that. After all, as former FDIC chair Gruenberg neatly summarizes, “PwC did not detect that hundreds of millions of dollars of assets claimed by Colonial did not in fact exist, had been sold to others, or were worthless”. This is sort of a trifecta of material, undetected financial thingies that PwC was, in principle, being paid to detect.

One wonders, or at least I do, or did, what PwC was doing in lieu of noticing material amounts of non-existent mortgages at a company whose main financial product was mortgages. Perhaps they were distracted, like in the classic “invisible gorilla” experiment, and PwC auditors were busy counting basketballs rather than noticing missing mortgages. Or noticing a gorilla.

Either way, it’s always nice to see audit performance art being performed at its highest level. And, equally, it is appealing to see PwC admit no liability in having apparently missed most of what might matter in doing an audit of a troubled mortgage company’s non-mortgages. Kudos.





  • All five major US sports leagues now have sports gambling sponsors /v @fmtofficesport


I’m a sucker for these sorts of long-run races, so seeing the changing ranking of cities by population for the last five hundred years is remarkable. In particular, I was struck by two things: for how long Istanbul was the largest city in the world; and how recent brief New York’s ascendancy to top spot was.

Readings: Private Equity, Ski Towns, and $20 Bill Auctions

Finance & Economics

Private Equity and Calpers

Like most right-thinking people, I’m alternately puzzled and repelled by private equity. To be fair, I’m puzzled, mostly, at the people who aren’t repelled by a field that somehow takes all the worst parts of accountancy, middle-school bullying, and train-spotting, and turns them into an occupation that brings in MBAs the way McKinsey once did. Nevertheless, it is striking that CalPERS (the California public employees retirement fund) wants more money invested in private equity, and wants it there so badly that a CalPERS executive was recently quoted as saying, “We need more of it, and we need it now”.

The preceding is rarely a good thing to say out loud. The investments industry is, after all, nothing if not good at satisfying demand with supply. Hey, you want more private equity? Have I got some private equity for you, etc. The trouble is that most such pension fund managers want only the best funds, and the best funds can only run so much money on a finite planet, no matter how fee-ishly appealing it might be to general partners to run more. As a result of this sort of demand, much of which has come from underfunded and struggling pension plans like CalPERS, the supply of private equity funds has boomed in recent years, and their recent performance has been passable, even if there are unsettling signs that when you decompose private equity returns of late, a material fraction of the returns have come from tricks of the trade, like funds selling overpriced assets to other funds, which will strike most observers as a not particularly sustainable model. (I will save the whole discussion of why pension fund managers overvalue illiquidity for another day.)

Nevertheless, the entire CalPERS meeting transcript on the topic is worth a read, especially if you mentally /s/r “private equity” for “Public Equities Around 1999” as you go along nodding your head at the Deep Investing Thoughts.

It becomes fairly clear that the key to convincing yourself that private equity is nothing even remotely like a bubble is having an underfunded pension plan.


The College Admissions Scandal Thing

I remain fascinated by the studiously obtuse reactions to the college admission scandal thing. I’m sure my fascination will soon pass, but today’s installment comes from the studiously obtuse New York Times, which opines that it’s really a sports scandal, not an admissions scandal. This is impressive Shiny Object Over There work, and, most importantly, manages to avoid asking why sports is so important at colleges anyway.

Say, for example, colleges wanted athletes there so that they could exploit them in some way, perhaps for other income, or even from fund-raising with alumni. I’m sure neither of these things happen, but, as a thought experiment, you could imagine there would be more reasons to run flawed athlete-related admission programs that could be exploited by the unscrupulous — even if colleges claimed they were going to crack down on such programs to prevent further sports-related admissions scandal things.

Your Moment of Zen

GoPro: Kilian Jornet – Running Ridges 

Readings: Fading stars, Neanderthals, Aging, and the Premier League



  • A devastating new paper takes apart the claim that “new economy” companies are more dominant than ever, or more productive, or more … pretty much anything, other than normal, that is. This is careful work, and while nothing is ever definitive in this area, should be highly cautionary when making claims about technology-driven “star” companies’ dominance.

2019 03 14 07 19 57

Finance & Economics

  • When pretty much everything consumers do in economic life is an example of a cognitive bias best explained by behavioral finance, then what you have there, more than anything, is a bias is how you think about consumer behavior. The latest example comes from temporal discounting in loans, with a new study that people’s willingness to take on expensive loans isn’t just about them being seduced by money now over subsequent costs/
  • Despite being the richest league in world football, the Premier League has only had four clubs in  League last eight twice in the last decade. Why doesn’t wealth translate into dominance at a league leve?

The University Admissions Scandal Thing

  • The explainers continue to come thick and fast in the college admissions scandal in the US, but this is a good question that shows how badly people misunderstand the nature of elite colleges in the US: Why Were These Rich Parents So Fixated on Elite Colleges? The answer, of course, is that getting into elite colleges has nothing to with increasing their wealth, but everything to do with inter-family status baubles in a great and pointless competition.


Readings: Dr. Google, Columbo, cheating, and distracted shareholders


  • One would think that when doctors create a plan to deal with a known patient risk, said patients would be less likely to find themselves in an emergency room. Well, hold that thought. In an interesting new study on “backfire effects” in medicine, researchers found that it actually led to a small but significant increase in hospitalizations, as doctors respond to increased risk awareness by putting more patients in hospitals, with commensurate increases in risk. As ever, Illich was right about medical iatrogenesis.
  • Despite no evidence with respect to efficacy, interest in stem cell injections for osteoarthritis is exploding, as shown by Google Trends data. We ignore Dr Google at our peril when we think about health risk, effectiveness, and costs.


  • Accounting is mostly excellent at showing you what sorts of indefensible financial things you can do and still defend it as proper accounting.  Perhaps unsurprisingly, Hollywood is very good at this game, as shown most recently in the entertaining tussle between  Columbo’s creators and Universal.  After deducting myriad expenses with flimsy connections, at best, to the popular show starring Peter Falk, they claimed for decades that Columbo was unprofitable, and the creators weren’t owed profit-sharing.

  • I’ve long been puzzled by the inherent conflict between social and for-profit ventures, and so it’s fun to see an explanation. It turns out that social entrepreneurs adoption of a “fantasmatic form appears to temporarily neutralize tensions and anxieties while preserving attachments to pro-social ideals”. In short, they are telling themselves stories to make themselves feel better.
  • When large investors aren’t paying attention and companies want to goose earnings, said companies race straight to cutting R&D costs, especially if their CEO isn’t very good. It suggests that a quick way of detecting companies playing accounting games might be track R&D changes.


The HIgher Education Scandal Thing

  • The college bribe scandal is turning into a lovely example of motivated reasoning, the idea that people mostly project their emotional biases when making an argument. Academics see it as a loophole in college admissions, college counselors say it isn’t the sort of thing they would do, social engineers see it as failed social engineering, and still others see it as proof that college is so easy that people don’t mind paying up in anticipate of being shuttled through. About the only thing people can agree on is that they are shocked how much they aren’t shocked by the entire affair.

Readings: Death robots, bullies, streaming, and smoothing


  • We’ve long known of the uncanny valley, the idea that people are made more squeamish by things that seem almost human than by things that are less human. It’s nice to see Kaiser Permanente testing this in a hospital context by using a video call to tell an admitted patient he was dying. This is almost as off-putting as a doctor doing this in hurried fashion during rounds, so, obviously, it’s worse.


2019 03 11 07 22 33



Readings: Google Trends, Deadoption, and Hockey Tape






Readings: Land Rovers, fake patients, complexity, and movie memes


  • Fake patients are as many as one in five people in psychiatric emergency rooms. Of course, this begs the Rosenhan-ian question of whether we know who is a patient in the first place.
  • Hidden FDA database contains reports of medicine device malfunctions and patient injuries, outside of public view. There is a future at Goldman Sachs for the FDA regulator who came up with the idea of saying that known problems can be filed out of sight, because, hey, they’re already known.


  • Complexity in large US banks has only declined marginally since the 2007 financial crisis. Like Mark Gattis said about why his bank character should be the sole survivor in Game of Thrones, the banks always survive.

2019 03 07 06 58 38Tech

  • Plagiarism-detection company Turnitin is now a billion-dollar business, with 34-million students using it, half of whom are in China. I have this idea for a plagiarism-detection company that also own the sites that create and store fake essays, thus making the circle complete. You can’t always rely on Someone On The Internet.
  • A company turns old Land Rovers into electric 4x4s. While this either won’t work, or it will have the driving range of a Bird scooter, I still want to dream a little here, so don’t disabuse me of the idea.

DSC7290 1


2019 03 07 07 04 37

  • When highways are congested, we would all benefit from randomness in “fastest-path” routing algorithms. I’ve written on this before, the idea that over-optimization leads to under-optimization, so I’m a big fan of the notion that algorithms do better if they do worse, which is more or less the case here, sort of like HHGTTG Improbability Drive, without the tea.