Saturday 22 March, 2020

If you might find it more useful to get this blog as a daily email, why not subscribe here? (It’s free, and there’s a 1-click unsubscribe). One email, in your inbox at 07:00 every morning.

The American Cemetery, Madingley this morning. Click on the image for a larger size.

Solitude vs loneliness

As the world struggles to adjust to lockdown, quarantine and social-distancing there’s an interesting and timely book on the horizon. It’s A History of Solitude by my friend and colleague David Vincent, who is one of Britain’s most distinguished social historians. It comes out on April 24. The timing is fortuitous but accidental: David has been working on the book for several years, starting on it after he had finished his previous book, Privacy: A Short History. I haven’t seen it yet, but Terry Eagleton, the literary critic, has and he’s written an interesting review for the Guardian. Snippet:

Solitude is not the same as loneliness. Lonely people feel the need for company, while solitary types seek to escape it. The neatest definition of loneliness, David Vincent writes in his superb new study, is “failed solitude”. Another difference between the two groups is that hermits, anglers, Trappist monks and Romantic poets choose to be alone, whereas nobody chooses to feel abandoned and bereft. Calling yourself “self-partnering”, meaning that you sit in the cinema (should they be open) holding your own hand, may be either a genuine desire for solitude or a way of rationalising the stigma of isolation. The greatest difference of all, however, is that solitude has rarely killed anyone, whereas loneliness can drive you to the grave. As the coronavirus rampages, some of us might now face a choice between physical infection and mental breakdown…

Thank God for experts


Producing vaccines under intense political pressure poses serious risks

How anti-vaxxers win — If any eventual vaccine harms even a tiny percentage of those who get it, “the anti-vaxxers can set back not only this vaccine but all vaccines,” said Barry Bloom, a professor at the Harvard School of Public Health. The anti-vaccine movement has been growing in the United States, and contributed to the country’s worst measles epidemic in 27 years in 2019. (From Politico’s nightly summary.)

This is yet one more reason why Trump is a menace. He keeps talking nonsense and stoking unrealistic expectations. This makes him the second biggest public health risk to the American public. And while we’re on that topic, here’s Larry Lessig:

If in January, Trump had “declared war” on this virus with the resolve of FDR, or Churchill, or even President Thomas J. Whitmore (Independence Day), he would have united the world against this common foe, and for once, the world could wage a war as one, without hesitation, and without regret.

Yet so tiny is the mind of our Idiot King that he could not even glimpse this extraordinary gift. His single focus was on the single indicator that seemed to say that he was, indeed, a genius — the stock market. And so he dissembled and obstructed to the end of faking the market out. Who knows if the man is really stupid enough to have believed that a virus that had brought China to its knees, once discovered to have infected 15 Americans, would “within a couple days go down to close to zero.” It doesn’t matter. The political system had taught Trump that he had the power to distort reality. The economic system has now taught Trump that he can’t distort economic reality. America’s economy — and the worlds’ economy— will now collapse. The election in November will be in midst of a great recession, compounded by unimaginable loss of human life. No President gets re-elected in times like that. Not the good ones. Not even the buffoons.

Remote conferencing with Zoom

Some of my research colleagues and I had a key meeting scheduled for this week and planned some weeks ago. As the University (Cambridge) went into lockdown we obviously couldn’t meet fate-to-face but were reluctant to cancel the discussion. Previously, we would have used conventional phone-conferencing, but I have become so pissed-off with the inadequacies of that medium that I suggested we used Zoom instead.

It was MUCH better. Two things in particular made all the difference: firstly one could see all the participants (as live images in small frames at the top of the screen); and secondly, whenever anyone started to speak, the software foregrounded them. This latter feature wasn’t perfect, but it was generally very effective. And the audio quality was sometimes a bit harsh, but still perfectly comprehensible.

My conclusion: the tech isn’t perfect, but I never want to go back to phone conferences again.

Why modelling is the rational way to make policy in a complex system

The Economist has an excellent explanation of the Imperial College epidemiological model That persuaded the UK government to change tack (though not quickly enough). The modellers

assigned covid-19 a “basic reproduction number” of 2.4. This means that in a population not taking any precautions, and where no one is immune, each case leads, on average, to 2.4 secondary cases.

Under those conditions the model showed the disease infecting 80% of the British population in three to four months. If 4.4% of the people infected became ill enough to be hospitalised and 30% of those deteriorated to the point of needing intensive care, then by mid-April demand for beds in intensive-care units (icus) would outstrip the health service’s “surge” capacity. In May the number of critical patients would be more than 30 times the number of icu beds available. Estimates of the fatality rate in China range from 0.5% to 1.5% of infections. Using a conservative 0.9% for Britain, the model put the death toll by the end of the summer at over half a million.

The Italian tragedy

One of the tragedies of this pandemic is the way it shows how social structures that we generally think of as embodying sociality and stability — extended families with several generations living closely together, for example — can be especially vulnerable. It turns out that Italy has a higher percentage of elderly people than most European countries, and about two-thirds of adults aged 18-35 live with their parents, with many houses containing three generations — which meant they were sitting ducks for Covid-19.

Wednesday 11 March, 2020

Wash your hands.

Rupert Beale, who works at the Francis Crick Institute, has a really good blog post about COVID-19 on the London Review of Books site. Towards the end, he reports a message he received from a colleague about the epidemic. It made three points.

  1. This is NOT business as usual. This will be different from what anyone living has ever experienced. The closest comparator is 1918 influenza.

  2. EARLY social distancing is the best weapon we have to combat Covid-19.

  3. Humanity will get through this fine, but be prepared for major changes in how we function and behave as a society until either we’re through the pandemic or we have mass immunisation available.

Lunch with Freeman Dyson

A New York Times piece by Siobhan Roberts to celebrate the late Freeman Dyson. It’s a lovely piece, which captures something of the man himself. For example:

Lunch with Dr. Dyson was never short of fascinating, fun or lengthy. He was a slow eater, and he did nearly all the talking. Listening, while trying to capture the last few peas of my salad, I’d realize that my lunch mate had made little progress with his meal; it was work, cutting and chewing the meat.

I’d try to fill airtime — and trigger his silent but shoulder-bobbing laugh — with trivial bits, like recounting a tale relayed by his son, George Dyson, an author and historian of technology, regarding an email the elder Dyson once received from a woman with a cleaning business. Subject line: “vacuum — unsatisfied.” Cindy had spent $500 on the DC14 model and had come to hate it with a passion, she explained in great detail. The suction on the rug was so strong that it threw “my shoulder out (NO LIE) having to push so hard.” She signed off, defeated: “I know that I will not hear from Dyson.”

Dr. Dyson, ever the reliable correspondent, hit Reply: “Thank you for the hate mail which I enjoy reading. I get quite a lot of it because my name is Dyson. But I am sorry to tell you that I am the wrong Dyson. My name is Freeman and not James. I suggest that you take the trouble to find James’s address and send the message to him. I wish you good luck and good health.”

Worth reading in full.

Paul Krugman on Thomas Piketty’s vast new book

Readable review of *Capital and Ideology — friendly by not uncritical. I’m an admirer of Piketty and got a lot from his earlier Capital in the 21st Century. But this 1000-page doorstop will have to stay on the nice-but-not-for-now list. So Krugman’s review is helpful.

Krugman starts by referring to that earlier tome. For the book-buying public, he observes, its big revelation was simply the fact of soaring inequality.

This perceived revelation made it a book that people who wanted to be well informed felt they had to have.

To have, but maybe not to read. Like Stephen Hawking’s “A Brief History of Time,” “Capital in the Twenty-First Century” seems to have been an “event” book that many buyers didn’t stick with; an analysis of Kindle highlights suggested that the typical reader got through only around 26 of its 700 pages. Still, Piketty was undaunted.

His new book, “Capital and Ideology,” weighs in at more than 1,000 pages. There is, of course, nothing necessarily wrong with writing a large book to propound important ideas: Charles Darwin’s “On the Origin of Species” was a pretty big book too (although only half as long as Piketty’s latest). The problem is that the length of “Capital and Ideology” seems, at least to me, to reflect in part a lack of focus.

To be fair, the book does advance at least the outline of a grand theory of inequality, which might be described as Marx on his head. In Marxian dogma, a society’s class structure is determined by underlying, impersonal forces, technology and the modes of production that technology dictates. Piketty, however, sees inequality as a social phenomenon, driven by human institutions. Institutional change, in turn, reflects the ideology that dominates society: “Inequality is neither economic nor technological; it is ideological and political.”

But where does ideology come from? At any given moment a society’s ideology may seem immutable, but Piketty argues that history is full of “ruptures” that create “switch points,” when the actions of a few people can cause a lasting change in a society’s trajectory.

Krugman admires Piketty’s intellectual ambitions in trying to tell such a vast story, but he also implies that sometimes Piketty’s reach exceeds his grasp. For example:

For me, at least, the vast amount of ground it covers raises a couple of awkward questions.

The first is whether Piketty is a reliable guide to such a large territory. His book combines history, sociology, political analysis and economic data for dozens of societies. Is he really enough of a polymath to pull that off?

I was struck, for example, by his extensive discussion of the evolution of slavery and serfdom, which made no mention of the classic work of Evsey Domar of M.I.T., who argued that the more or less simultaneous rise of serfdom in Russia and slavery in the New World were driven by the opening of new land, which made labor scarce and would have led to rising wages in the absence of coercion. This happens to be a topic about which I thought I knew something; how many other topics are missing crucial pieces of the literature?

The second question is whether the accumulation of cases actually strengthens Piketty’s core analysis. It wasn’t clear to me that it does. To be honest, at a certain point I felt a sense of dread each time another society entered the picture; the proliferation of stories began to seem like an endless series of digressions rather than the cumulative construction of an argument.

Piketty sees rising inequality as being at root a political phenomenon. The social-democratic framework that made Western societies relatively equal for a couple of generations after World War II, he argues, was dismantled, not out of necessity, but because of the rise of a “neo-proprietarian” ideology. But, asked Krugman,

why did policy take a hard-right turn? Piketty places much of the blame on center-left parties, which, as he notes, increasingly represent highly educated voters. These more and more elitist parties, he argues, lost interest in policies that helped the disadvantaged, and hence forfeited their support. And his clear implication is that social democracy can be revived by refocusing on populist economic policies, and winning back the working class.

And his summing up?

The bottom line: I really wanted to like “Capital and Ideology,” but have to acknowledge that it’s something of a letdown. There are interesting ideas and analyses scattered through the book, but they get lost in the sheer volume of dubiously related material. In the end, I’m not even sure what the book’s message is. That can’t be a good thing.

Phew! I don’t think I have to read it.

Slouching towards Dystopia

My longish (3000-word) essay in this week’s New Statesman. It’s basically a account of how we got into our current impasse with digital technology. Sample:

The important point is that surveillance and our passive acceptance of it lies at the heart of the dystopia we are busily constructing. It doesn’t matter which technology is used to identify people: what matters is that we can be identified, and then correlated and tracked across everything we do. Mass surveillance is increasingly the norm. In countries such as China, a surveillance infrastructure is being built by the government for social control. In Western countries, led by the US, it’s being built by corporations in order to influence our buying behaviour, and is then used incidentally by governments.

What’s happened in the West, largely unnoticed by the citizenry, is a sea-change in the social contract between individuals and the state. Whereas once the deal was that we accepted some limitations on our freedom in exchange for security, now the state requires us to surrender most of our privacy in order to protect us. The (implicit and explicit) argument is that if we have nothing to hide there is nothing to fear. And people seem to accept that ludicrous trope. We have been slouching towards dystopia.

Some historical perspective on the dominance of current tech giants

From this week’s Economist:

As big tech’s scope expands, more non-tech firms will find their profits dented and more workers will see their livelihoods disrupted, creating angry constituencies. One crude measure of scale is to look at global profits relative to American GDP. By this yardstick, Apple, which is expanding into services, is already roughly as big as Standard Oil and US Steel were in 1910, at the height of their powers. Alphabet, Amazon and Microsoft are set to reach the threshold within the next ten years.

Remember what happened to Standard Oil and US Steel?

Tuesday 4 February, 2020

Newton’s notebook

The young Isaac Newton was a painstaking recorder of his expenditure, probably because he was relatively poor. This is one of his early notebooks, where he records his expenditure on frivolous ‘sweetmeats’ — as sugary treats were then called.

This particular notebook is included in the Fitzwilliam Museum’s current (and fascinating) Feast and Fast: the Art of Food in Europe 1500-1800 exhibition.

Surprise, surprise! There are lots of scammers on Airbnb

Good Vice investigation reveals what a lot of people already know:

The stories quickly started to fall into easily discernible categories. Scammers all over the world, it seems, have figured how best to game the Airbnb platform: by engaging in bait and switches; charging guests for fake damages; persuading people to pay outside the Airbnb app; and, when all else fails, engaging in clumsy or threatening demands for five-star reviews to hide the evidence of what they’ve done. (Or, in some cases, a combination of several of these scams.)

The article has an interesting list of the ways people (hosts as well as guests) can be scammed. And, to be fair, Airbnb seems to be willing to accept some responsibility for the bad stuff that goes on on their platform. In that sense, it provides a welcome contrast to Facebook.

Talk, don’t fly

One predictable consequence of Corona. “Zoom Video Stock Soars as Coronavirus Travel Bans Boost Focus on Videoconferencing.”


Towards another Facebook Presidential election

Four reasons that make Frederic Filloux believe that we are heading towards another American Presidential Election swayed by Facebook.

  1. Mark Zuckerberg’s stubbornness in exonerating political advertising from any fact-checking process. In short: any false statement showing up in a newsfeed can be debunked by either Facebook team or any TPFC (third-party fact-checking) it relies on and taken down. But if the same statement appears in a paid-political ad, it is allowed to stay (unless it points to previously debunked fake news).
  2. There is no change of Facebook’s core principle, which is to reward emotion, and incendiary statements — a principle that clearly favors right-wing rhetoric (starting with Donald Trump). Facebook never considers altering its algorithm to spotlight high qualityand more even-keeled content. The reason is that it brings less engagement, which is at the core of Facebook’s economics.
  3. Unlike 2016, this time, Facebook has a vested interest in seeing Democrats lose. Whoever the nominee be, he or she will go after Facebook, at least with severe regulation and at worse with an attempt to break the companies’ holdings up.
  4. Trump digital campaign is running on full throttle compared to Democrats’. In the exact same scenario as 2016, the Trump campaign is spending heavily on social media and runs 3x more ads than Pete Buttigieg and 7x more than Elizabeth Warren, who is following Hillary Clinton’s path: in 2016, while Trump was flooding voters with 6 million different ads, HRC ran only 66,000 different versions of its message. Toyda, the numbers are staggering: according to a detailed investigation published last week by the Guardian, the Trump campaign spent $19.4m on 218,100 different Facebook ads in 2019, which were seen between 633m and 1.3bn times.


Climate models are running hot — and nobody knows why.

Fascinating Bloomberg report. It’s not just one model, but lots of the major ones. They’re starting to predict much higher temperatures. And at the moment, there’s no consensus in the climate-research community about why this is happening. Is it just a quirk of these very complex models? Or the result of interactions that nobody’s understood? It’s a bit like the problem of inexplicable machine-learning systems. Only more worrying.

Ten tech trends that shaped a decade – Pew Research Center

  1. Social media sites have emerged as a go-to platform for connecting with others, finding news and engaging politically.
  2. Around the world and in the U.S., social media has become a key tool for activists, as well as those aligned against them.
  3. Smartphones have altered the way many Americans go online.
  4. Growth in mobile and social media use has sparked debates about the impact of screen time on America’s youth – and others.
  5. Data privacy and surveillance have become major concerns in the post-Snowden era.
  6. Tech platforms have given rise to a gig economy.
  7. Online harassment has become a fairly common feature of online life, both for teens and adults.
  8. Made-up news and misinformation has sparked growing concern.
  9. A majority of Americans see gender discrimination as a problem in the tech industry.
  10. Americans’ views about tech companies have turned far less positive in recent years.

No surprises, really. But useful to have empirical evidence.

See Source for details.

Serial Killers: Moore’s Law and the parallelisation bubble

Cory Doctorow had a thoughtful reaction to Sunday’s Observer column, where I cited Nathan Myhrvold’s Four Laws of Software. “Reading it”, he writes,

made me realize that we were living through a parallel computation bubble. The period in which Moore’s Law had declined also overlapped with the period in which computing came to be dominated by a handful of applications that are famously parallel — applications that have seemed overhyped even by the standards of the tech industry: VR, cryptocurrency mining, and machine learning.

Now, all of these have other reasons to be frothy: machine learning is the ideal tool for empiricism-washing, through which unfair policies are presented as “evidence-based”; cryptocurrencies are just the thing if you’re a grifty oligarch looking to launder your money; and VR is a new frontier for the moribund, hyper-concentrated entertainment industry to conquer.

“Parallelizable problems become hammers in search of nails,” Cory continued in an email:

“If your problem can be decomposed into steps that can be computed independent of one another, we’ve got JUST the thing for you — so, please, tell me about all the problems you have that fit the bill?”

This is arguably part of why we’re living through a cryptocurrency and ML bubble: even though these aren’t solving our most pressing problems, they are solving our most TRACTABLE ones. We’re looking for our keys under the readily computable lamppost, IOW.

Which leads Cory (@doctorow) to this “half-formed thought”: the bubbles in VR, machine learning and cryptocurrency are partly explained by the decline in returns to Moore’s Law, which means that parallelizable problems are cheaper/easier to solve than linear ones.

And wondering what the counterfactual would have been like: if we had found a way of extending Moore’s Law indefinitely.

Tony Brooker R.I.P.

Tony Brooker, the guy who developed Autocode, arguably the world’s first machine-independent programming language, has passed away at the age of 94. He did it to make one of the early computers, the Ferranti Mark 1 in Alan Turing’s lab in Manchester University, useable by human beings. There’s a lovely recording of him talking about it on the British Library site. Also, a nice obit in the New York Times.

Dining with Stalin

One of the glories of the blogosphere is its infinite variety. The economist Branko Milanovic has a fascinating post on his blog after he discovered an obscure book (probably based on an academic dissertation) in a secondhand bookstore in St Petersburg. The book is a detailed (400-page) account of 47 banquets that Stalin hosted between 1935 and 1949. The banquets, hosted in various reception rooms of the Kremlin, included between 500 and 2000 people and were, Milanovic writes,

sumptuous affairs, especially if contrasted with generalized penury of meat, fresh fruit and vegetables that often was the case in Moscow and even more so in the provinces. All produce and drinks however were Soviet-made. Compared to their equivalents organized by Hitler and his lieutenants and studied by Fabrice d’Almeda in The High Society in the Third Reich, Soviet banquets were more monotonous, less extravagant, and more modest. They were also more business-like in not (generally) including family members.

There were two groups of diners. The first (obviously) were members of the Politburo and top government officials. The guests were various groups of people. Many of the banquets were done after the May 1 or the Day of the October Revolution (November 7) military parades and thus included mostly the Army and the Navy. One especially favoured group, apparently, were Air Force pilots.

These provided some comic interludes. For example:

There were several special banquets for the pilots that in the 1930s achieved some notable successes for the Soviet aerospace, including flying to the North Pole, saving sailors stuck in the icy northern desert, and flying long-range non-stop flights to North America. These banquets seemed to put Stalin in an exceptionally good mood because he treated pilots with special consideration, allowing them liberties that very few were granted, including having his toast twice interrupted by the same pilot, at two different banquets. At times, there were unusual scenes that in a more bourgeois Western settings would have been unimaginable—as when Stalin invited the pilots to the leadership table and then began to hug and kiss each of them, which in turn led the entire Politburo to do likewise. With a dozen of pilots and more than a dozen of members of the leadership that implied perhaps as many as 150 or even 200 hugs and kisses. An almost California-like therapy of free hugs.

For those at the top table, however, things were anything but comical:

Even if the core was stable (Stalin, Molotov, Kaganovich, Kalinin, Voroshilov, and to some extent Mikoyan, Andreev and Zhdanov) included also the people who were, at various times, later purged and executed. For example (p. 158), “From June 1937 to April 1938, almost to his arrest, Kosior sat five times at that [leadership] table….In August 1938 Kosior’s wife was shot. And then he was arrested himself. He was taken to the higher level of punishment [probably torture]”. Overall, out of 21 people (excluding Stalin) who sat at the leadership table in 1937 and 1938, eight were shot and two killed themselves (p. 162). Thus almost half of the convives to that supreme table were killed by the main host. Not a usual occurrence.

Simultaneously creepy and fascinating.