Boston, City of Other People’s Bad Memories

My home city has had a remarkably long run of peace.  The last time war touched Boston was when it was besieged for almost a year by the colonials during the American Revolution, some 240 years ago.   Few places in the world have been at peace for so long – not the US South or West, nor Asia or India, nor most of Europe.

Perhaps that’s why it’s filled with memorials to other people’s tragedies.  The events have nothing to do with Boston itself, but refugees from them have washed up on its calm shore.  Even generations later, they still want to keep hold of their stories.  The city obliges.   It tracks all public art at this site, which I’ve used for the links below.

Here, for instance, is the New England Holocaust Memorial (sculpture 1995, event 1939-1945):

It’s the largest memorial in the city, and is squarely in its oldest section.  It looks deliberately jarring amongst all the old brick.   There’s one tower for each million victims, and they’re covered in numbers, meant to evoke the camp tattoos.

Not far away is the Boston Irish Famine Memorial (sculpture 1998, event 1845):

The healthy figures are leaving the starving ones behind, a disturbing image.

Rather more obscure are the Hungarian Revolution Freedom Fighters (sculpture 1986, event 1956):

This was from when the Soviets crushed an uprising in Hungary in the early Cold War.

Then there’s the Armenian Heritage Monument, commemorating the “immigrant experience”, but mainly about the Armenian Genocide.  Scuplture 2012, event 1915-1923

Here’s the Polish Partisans Memorial (sculpture 1979, event 1940): This remembers the heroic and doomed resistance to the Nazi and Soviet invasions of WW II.  The sculpture had been on the Boston Common, but perhaps was too obscure and grim for the city’s main public space, and so is now near the World Trade Center T stop.

The city has plenty of memorials to its own past figures, of course, but they’re generally uplifting in some way.   These are all reminders of tragedy.  Note that they’re relatively new, all from the last couple of decades.   Remembering a past loss is a recent trend.   Is it a memento mori for groups that have done quite well in the new world?  A reminder to other Bostonians that this could happen to them?

More recent immigrant waves have their own tragedies: the Vietnamese in the early 70s, the Cambodians in the early 80s, the Russians fleeing the death of their dream, the Chechens not long after (and the city has anti-memorials to them from the Marathon bombing), Haitians, Somalians, and even students from Tianemen Square.   They don’t yet have the political clout to present their own stories, but they will at some point.

The city and its universities have more volumes in their libraries (about 40 million) than anywhere else in the world, except New York City (50 million).  It already acts as a global memory.  It’s acting as a sculptural memory too.

Posted in Uncategorized | Leave a comment

Plug For “Let’s See This Work – An Engineer At the Movies”

Alec Guinnes in one of the best STEM films, “The Man in the White Suit”

I’ve started a new blog that’s specific to movies, looking at them with a technical eye:  Let’s See This Work .  It’s not so much about the technical construction of movies, but rather notes from a technical person on what movies are about.    There’s a main post, STEM Movie List,  that gives notes and links to films with Science / Technology / Engineering / Mathematics characters, but most of the posts are reviews of films and their themes.   Others are analysis: The Funniest Screenwriter of All Time, Movie Inventors – Oddballs, or a summation of IMDB keywords in What Are Movies About?  The older posts are carried over from this blog.  Hope you enjoy them!

Posted in Uncategorized | Tagged | Leave a comment

How Did Wind Get So Cheap?

Rather surprisingly, wind power is now the cheapest form of electricity in the US:

Lazard LCOE Dec 2016. Click to enlarge.

This comes from the tree-hugging socialists at  Lazard Asset Management.  Unsubsidized wind comes in at $32 to $62 per MWh, depending on the site.   Natural gas is substantially more at $48 to $76, and coal and nuclear are right out at $60 to $143 and $97 to $136 respectively.  Wind is now generating 5% of all power in the US, and 45% in Denmark.  Its cost has fallen by 1/3 in the last five years.  Solar gets all the publicity, perhaps because those black panels look magical compared to the simple mechanics of a wind turbine, but wind is actually ahead, especially in the Midwest and Northeast.

So how did wind mills change from being a stereotype of communes to major industrial machinery?  It looks like a combination of standard engineering improvements and quite non-standard citizen involvement in Denmark.  The story gets told in the comprehensive new book “Wind Energy For the Rest of Us” by Paul Gipe:

Click for author site

Click for author site

It discusses the historical development of turbines since the 19th century, the reasons for their key features, how to evaluate a site, how to choose a machine, and how to finance it.  It ends by noting that it would take about 750,000 standard 2 MW turbines to power all of the US.  They’re about as hard to build as a heavy truck, and much easier than aircraft, and so could not only save money and cut pollution but revitalize manufacturing.

A key reason why they’re getting cheaper is geometrical scaling – a wind turbine that’s twice as big sweeps out four times the area and so gets four times the power.  If it costs less than four times as much to build, you’re ahead.  Every tweak that makes the blades longer or the towers taller is a win.  They’re now up to 195 m tall, with 80 m blades and a peak power of 8 MW, although those are only used off-shore.

Overall it looks like their development followed the typical pattern for a new machine: try lots of different approaches initially, then narrow down to an optimum one and drive for high volume and low cost.   The standard configuration today is an open rotor with three fiberglass blades on a horizontal axle facing upwind.  They have overspeed protection by twisting the blades and having brakes on the axle.   They feed a variable-speed under-sized generator, and it’s all on a tall cylindrical steel tower.   Every variation on these features has been tried:

  • Put a duct around the rotor to increase wind speed (too heavy) or have an open rotor.
  • The blade number: one (too heavy), two (too much vibration) or three.
  • The blade material: wood (rots), steel (too heavy), aluminum (too weak) or fiberglass (a hollow shell around a structural spar.  Does have to be made as one piece, making it hard to transport).
  • The axle orientation: vertical (can’t furl in high winds) or horizontal (turn sideways to high winds to lessen the force on the rotor).
  • The blade direction: facing downwind (makes a whomp noise every time a blade goes through the wind shadow of the tower) or upwind (have to worry about the blades hitting the tower, which is highly bad).
  • For overspeed protection: little parachutes on the tips (really?), slats on the blades (too prone to jam), twist the whole blade to make the airfoil stall and put a brake on the axle as a last resort.
  • Use a fixed-speed generator to match the 60 Hz of the grid (needs variable gearing or loses energy by turning rotor too slowly) or a variable-speed generator that’s converted to AC (possible with modern high-power switching transistors).
  • Size the generator to handle high winds (adds cost and weight) or size for only medium winds (improves the average power output since it runs full out more of the time, thus needing less backup.   The average power is now up to 40-50% of the rated limit versus 30% with the old over-sized generators).
  • The tower style: a lattice mast like a radio tower (rusts, gets covered in bird shit, makes wind noise, unsafe to climb), use guy wires (noise, needs land), or have a closed cylinder (which can be climbed in any weather).
  • The tower material: wood (can’t get tall enough), concrete (ugly and slow to assemble), steel (painted white for aesthetics instead of gray-blue to make less visible).

So it took a long time to figure it all out!  But now that the major parameters are set, one can really go to town on optimization.  If you’re building $10B of wind turbines a year (as happened in 2016 in the US), an 0.1% improvement is worth $10M, which sure pays for your salary.  For example, the Betz Limit says the maximum amount of power that can be taken out of an air stream is (1 – (2/3)^3) = ~60% of the kinetic energy in the stream.   Modern turbines get within 75% of that, so a tweak of the airfoil shape (perhaps depending on the wind speed distribution at a site) could get another percent or two.

And now that people have a lot of experience with the designs, they can estimate how long they’ll really last.   Early machines failed after only a couple of years, but modern ones last for twenty.  That makes a big difference in the financing, since you no longer need a risk premium for early failure.

What’s more interesting, though, is that it got started in a quite unconventional way – at a folk craft / technical school in Denmark, the Tvind School.   They built the first modern wind turbine, the Tvindkraft, with secondhand parts and student labor:

900 kW turbine and the Tvind School in Jutland. Click for link

900 kW turbine and the Tvind School in Jutland. Click for link

It’s still working after almost 40 years!  It pioneered the use of cantilevered fiberglass blades and a particular kind of flange for the blades that uses glass fibers wrapped around the bolt holes.  It’s big, 900 kW, and generated so much power that the grid couldn’t handle it, and so they dumped it into heating the school.

It was built during the oil crises of the 1970s.  The Danes had to buy oil from the Mideast and coal from Germany, and liked neither option.   Nor did they like buying nuclear-generated electricity from Sweden, who put a reactor immediately across the strait from them.  It had to be wind, but as a small country they didn’t have the resources to do a big research program.

The US had revived wind research at the same time, but gave it to NASA, GE, and Boeing.  They applied aerospace ideas to it, thinking that it was just another airfoil.  But it’s not – weight doesn’t matter.   What really matters is reliability, and the GE and Boeing machines failed after less than a year.

The Danes came at it from a cost and reliability standpoint instead of performance.  They concentrated on making the blades strong and durable – each of the Tvindkraft blades weighs about five tons.   Even more importantly, they published their designs. A lot of small firms sprang up making similar-looking mills.   Some even specialized in making just pieces like the blades.  Danish windmill owners also joined up in cooperatives and forced the builders to have standard features likes brakes on the generator.  Some people didn’t like the look of the towers, but once they got a share of the revenue they minded much less.

Eventually a farm-machinery company, Vestas, got involved and took it over.   They’re now about the largest wind company in the world, with 2016 sales of about 10 billion euros.  They’re closely followed by another Danish company, Bonus Energy, now a part of Siemens.  GE is still a large player, and is covering the Midwest in towers.  The Chinese bought Danish and German tech and now dominate the field, since they’re desperate to get away from coal.  At the end of 2015 there were more than 300,000 turbines operating around the world, with a capacity of more than 430 GW, and that grew by 63 GW in just that year.

So this huge industry started at an obscure school in Denmark, grew to a cluster of small companies, got taken up by some larger ones, was incentivized by feed-in tariffs in Europe and investment tax credits in the US, and over 40 years worked out all the kinks in the technology.  It benefited from natural geometric scaling, from airfoil simulations, from power electronics, and most of all from operational experience.  Another factor of two in cost reduction is in sight, and off-shore tech is coming on strong.   It’s the newest class of large machinery in the world (rockets and nuclear reactors come from the 50s, and jets and container ships from the 60s), and it’s taking over.

Posted in Uncategorized | Tagged , | Leave a comment

The Auto Industry Does Its Bit

I recently leased a 2017 Chevy Volt.   It’s a nice mid-range car with good interior space, a lot of zip, and is really quiet.   It’s even rather stylish:


And it gets 75 miles per gallon in terms of CO2 emissions, 3X the US new vehicle average.  That is, it gets 40 mpg when running on gas, and the equivalent of  90 mpg when running on electricity at 2.7 miles per kilowatt-hour.   A kilowatt-hour is about what a small air conditioner consumes every two hours.   I drive on electricity about 85% of the time, so that averages out to 75.

The number is so high because Massachusetts has pretty clean power.  The EPA tracks this here: EPA Power Profiler.    It says that MA burns about 50% natural gas with the rest as nuclear (30%), hydro (6%), wind (4%), coal (3%), landfill gas (2%), and some solar photovoltaic and biomass.  The US as a whole averages about 55 mpg because they burn a lot more coal.   This data is all from 2012, though, and coal is way down since then.  MA has dropped its emissions from electricity by almost a factor of 2 between a peak in 2007 and 2013, according to the the state tracking site here: MA GHG Emission Trends.

Electric car sales are growing fast.   About 160,000 battery-electric and plug-in hybrids were sold in the US in 2016, a 37% increase over 2015.  That’s still only 2% of overall US car sales of ~7M, and only 1% of car and truck sales of ~17M, but it’s a lot.  At $50K per car, about $8B of electric cars were sold last year, which is only a little smaller than the movie industry.   The breakdown by model is:


It’s nice that the top 4 models are American, and that the Volt and Fusion Energi are  union-made.   These were record years for the Teslas and Volts.

The common rule is that greenhouse gas emission have to drop by 80% by 2050 compared to 2005 levels in order to keep global warming under 2 degrees C.  The average new car and truck in 2005 got about 20 mpg, or 5 gallons per 100 miles, and this car does about 1.3 gallons/100 miles,  a 75% drop.   It’s almost there already!  The EU currently has a limit for new cars of 130 gm CO2 per km (42 mpg), and is going to 95 gm/km in 2021.  This car does about 62 gm/km, and will get cleaner still as the the power system de-carbonizes.

So don’t blame the auto industry for climate change going forward.  They’re offering mid-priced, comfortably appointed, well-driving cars that are way ahead of government regulations and about up to future requirements.  This car even feels much better when it’s running on the battery.   It’s smoother, quieter, and has more acceleration.   When the gas engine comes on, I’m reminded of what 20th century cars felt like.   It feels like phones that had to be wired to the wall, TVs that only played shows when it wanted to, and information that you had to go to the library to find.  This feels like a 21st century car.


Posted in Uncategorized | Tagged | Leave a comment

When Modeling Goes Bad – “Weapons of Math Destruction”

The political modeling that I talked about in the last post now affects most decisions that institutions make with respect to individuals.   This is nicely described in in the recent book Weapons of Math Destruction by Cathy O’Neil.   She has worked on these systems herself while at the hedge fund D. E. Shaw and at various e-commerce startups.


Clink for link to author blog

This modeling attempts to classify millions of people based on anything that can be gleaned about them online.  She has chapters on each of these categories of decisions:

  • College Admissions – are driven by metrics related to the US News and World Reports rankings, which conveniently don’t include tuition.
  • Sentencing and Parole – Who is likely to commit more crimes before and after jail time?
  • Hiring – Study people’s social media, credit scores, and judicial records to see if they’re a good match for a firm.
  • Firing – Teachers are especially closely judged these days because of right-wing opposition to the whole concept of public schools.   In particular, the No Child Left Behind Act almost forces teachers to be ranked and fired.   This has had the predictable consequences of teachers leaving low-performing school districts, skewing lessons towards the tests, and cheating.
  • Borrowing – How are credit scores actually arrived upon?  FICO is actually a clear and straightforward metric, but lots of banks use mysterious e-scores these days.
  • Insurance – Who gets covered and for how much?  The ACA forced consistent standards on the medical insurance industry, but that’s about to disappear.
  • Voting – How can the news and advertising that people see be tuned to persuade them to vote one way or the other?  She actually discusses the work of Cambridge Analytica, which had a role in Trump’s victory, even though the book came out long before the election.

These decisions are largely made by computer these days because it’s cheap.   Interviewing students or borrowers or applicants takes real people with real skills, and that’s more expensive than just screening them by algorithm.   That means it gets done for the upper classes, who otherwise get annoyed by impersonal rejection, but not for the middle class and below.

But cheap methods are usually crummy, and that’s true here too.  They don’t have nearly enough statistical power to do a good job.   They’re using way too few data points (E.g. teacher evaluations are based on only 20 or 30 scores from wildly different children), are using proxies that have no real connection to what’s being decided, and have poor feedback paths to adjust the models.

Worse still, the methods are completely opaque to the people they are affecting, and often to the people using them.    An answer spits out, and there’s no recourse.  No one knows why they got turned down.   If they’re using a neural net, even the coders don’t know why it gives the answers it does.

Even worse still, the goal of the algorithm is entirely for the benefit of the organization running it.   No larger social goal can be applied, nor can any larger sense of fairness.  Thus the algorithm can easily cause death spirals. E.g. by denying mortgages to certain neighborhoods, the area declines, making it less attractive for investment, causing further declines.  By denying people bail or parole, whole classes of people can be put in decline.  The algorithm may optimize the short-term profit of the people running it, but is too mysterious to service long-term goals even for them, much less society as a whole.

She contrasts this with player evaluations in major league sports.  The statistics about a player’s performance are all publicly known, and are plentiful if they’ve been in the game for any time.  They are directly related to the main question – how much will this player help the team win?   The model can be constantly run to verify its predictions, and adjusted when wrong.   But if you have a model for who makes a good hire, you really only get to see a little about who gets picked, and then only at infrequent reviews.  You don’t learn anything about the people rejected.

So what is to be done?  Her suggestions don’t seem that helpful to me.  She focuses on an ethics code for programmers of such algorithms.    That’s been a valuable approach in civil engineering, where people really are conscious that their work can kill when it fails.  Few other engineering disciplines insist on this, though.  The connection between one’s work and its consequences is much more remote in programming than it is in construction.

It’s better to have public and independent inspections.   That’s how bridges get certified.   It’s coming to be how components get certified in cars and airplanes.  An outside party reviews the design process and the safety behavior of a device and gives it a rating.

That’s hard for big software systems like these, especially since they’re considered to be a business advantage.  What people can do is test the system with simulated applications.  She describes how researchers can create fake online personas to see how their social media gets steered, or their search results, or their college and loan applications.

The Big Data companies like Facebook and Google hate this, though, and do everything they can to prevent it.  They don’t want people to know how they’re being judged, for fear that users will game the results.   There’ll be an arms race between the parties trying to understand what Big Data is doing to society and the increasingly malevolent firms themselves.

Anyway, the book as a whole is clearly written and thorough.  Her blog is excellent too!   It’s a good overview of a problem that will only get worse.

Posted in Uncategorized | Tagged , | Leave a comment

Weaponized Psychology Helped Elect Trump

The US has just elected a president who is an outright criminal – a man who cheats contractors, steals from investors, and assaults women. What on earth happened?  Everyone has a theory, but let me add one more – his campaign made use of weaponized psychology.   I noted in my review of the SF novel Affinities that large data sets and serious mathematical analysis were getting traction even in the most difficult of subjects like psychology.  We’re now seeing real-world consequences of these advances.

Analytica CEO Alexander Nix, which is a good supervillain name

Analytica CEO Alexander Nix, which is a good supervillain name

To be specific, Trump used the services of a company called Cambridge Analytica to do his voter analysis and message management.  They are a US subsidiary of a UK firm with the blandly sinister name Strategic Communications Laboratories.  They’ve been conducting propaganda and disinformation campaigns all over the world for the last twenty years, often for the US DoD and UK MoD.  They worked on the Leave side of Brexit.

Cambridge Analytica was backed in the US by one Robert Mercer, a right-wing hedge fund billionaire with a PhD in computer science from the University of Illinois.   He was a major backer of Ted Cruz, but even CA couldn’t save that campaign.   The standard joke was “Why do people take an instant dislike to Ted Cruz?  It saves time.”  When Cruz folded up, Mercer shifted his funding to Trump.   His daughter Rebekah is now a member of Trump’s transition team.  Trump’s former campaign advisor Paul Manafort was apparently against hiring CA, but was overruled by Jared Kushner, Trump’s son-in-law. Mercer himself spends money on teenage-boy projects like huge model railroad sets, a collection of machine guns including that of the robot in “Terminator”, and enormous yachts.  He believes the US should return to the gold standard.   He was sued by his mansion staff for stiffing them on pay, and so should fit right in at the new Administration.

CA’s job was to find the narrow path to electoral college victory for Trump.   Clinton had wrapped up the Northeast and Far West, and most of the South and the Mountain West was solidly Trump, but the upper Midwest could be exploited.  CA signed up with Facebook and got access to the profiles of hundreds of millions of voters.   They built models of each voter based on the OCEAN personality profile system.   This stands for Openess to experience, Conscientiousness, Extraversion, Agreeableness, and Neuroticism.  It differs from other systems such as Myers-Briggs in that it arose from factor analysis, of grouping people by common traits, instead of having an underlying theory.  A nice feature of it is that the traits can be inferred from writings and links instead of requiring questionnaires.

Once you know the personality types you’re dealing with, you can judge the effect of political messages on them.   Again, you use feedback from social media to estimate how well an approach is working.  In the weeks before the election, CA saw that early voter turnout was higher among older, rural white voters and lower among blacks than expected.  They reset their poll weightings and saw their opportunity.  They did big ad buys in the northern Midwest and advised Trump to focus there.  He changed his campaign schedule to include stops in Michigan.   Pundits thought that was crazy, since Michigan was solid blue, but he actually took the state.

In the end it took only a hundred thousand votes to swing Wisconsin, Michigan, and Pennsylvania.  That’s less than 0.1 % of the votes cast, but it will change the entire direction of the country for the next four years, and likely well beyond.   That’s the point of leverage that this kind of analysis can find.

Now, the Clinton campaign had their own analytics firm, Timshel, founded by a veteran of Obama’s 2012 campaign, Michael Slaby.   Obama’s campaign was also notable for its use of Big Data to try to capture the intent of every single voter.  I remember some discussion about who Trump would use early in the campaign, and the consensus was that no serious technical would ruin their reputation by associating with Trump.   The only exception was Peter Thiel and his mysterious Palantir Technologies, but they don’t seem to have gotten involved.   What Trump’s victory showed is that this kind of technology can be used by either side.   It’s not something that only sophisticated progressives can handle.  Thinking that was snobbish.

Was it an important factor in his win?   Maybe not compared to anti-Clinton misogyny, xenophobia, interference by Russia and the FBI, weariness with eight years of Dem rule or any of the other swirl of explanations.   Any or all of them could have contributed.   What’s likely, though, is that this level of manipulation will only increase.  You may think you’re voting based on a rational analysis of the issues, but what you see and hear will be adjusted for you personally by vast systems driven by models of your psyche.   It may sound like wild conspiracy theory, but people are making businesses out of it.

Update 2/27/17 – The Guardian reports that Robert Mercer loaned the services of Cambridge Analytica to, the pro-Brexit organization headed by Nigel Farage.  The expense was not reported, which is illegal. He is also an investor in Breitbart News and recommended Steve Bannon to Trump.   The Right now has another deep-pocketed backer besides Adelson and the Kochs, one with dangerous technology.


Posted in Uncategorized | Tagged , | 1 Comment

The Winningest SF Authors Are Women

Item Number bb004400

Le Guin in the 1970s, click for bio

The New Yorker recently published a charming interview by Julie Phillips of Ursula K. Le Guin.  It described her upbringing in a house full of myth and story headed by her father the great anthropologist Alfred Kroeber, her difficult relationship with Radcliffe where she got a degree in French, and her brilliant spurt of work starting in 1966 at age 37 with “A Wizard of Earthsea”, and extending over the next 8 years to the seminal works “The Left Hand of Darkness”, “The Lathe of Heaven”, “The Farthest Shore” and “The Dispossessed”.  She’s now 87 and as sharp as ever, tangling with Amazon over monopoly and Google over the digitization of literature.

“A Wizard of Earthsea” made a particular impression on me when I read it at age 13.  It starts with the standard SF trope of the Big Zoom.  That’s where a young person from a humdrum background comes to realize over the course of the story just how big and wonderful the world actually is.   That’s the plot of Heinlein juveniles like “Have Space Suit – Will Travel”, whose title alone tells you what’s going to happen.  Usually those stories have the protagonist going from misunderstood loner to reaching their proper place in the world, which makes them highly satisfying to young fans.  In Wizard, the young hero Ged does in fact become Archmage of Earthsea, but also realizes how circumscribed his vast power must be.  Even at age 13, I realized that Le Guin was working at a whole different level than most SF authors.

In reading about her elsewhere, I discovered a remarkable thing – she has won more of the top awards in SF, the Hugo and Nebula for best novel, than any other author except Lois McMaster Bujold.   They’re both tied at 6.   Le Guin won the Hugo and Nebula for “The Left Hand of Darkness” in 1970, both again for “The Disposessed” in 1974, and Nebulas for “Tehanu: the Last Book of Earthsea” (1990) and “Powers” (2008).  Bujold won the Hugo for “The Vor Game” (1991), “Barrayar” (1992), “Mirror Dance” (1995) and “Paladin of Souls” (2004), which also won the Nebula.  She also won a Nebula for “Falling Free” in 1988.

Connie Willis then comes in at 5 best novel awards with 3 Hugos and 2 Nebulas, but she has won more total awards, 18, than anyone else.   Le Guin is tied for second with Harlan Ellison at 11.  Joe Haldeman and Robert Heinlein also have 5 Best Novel awards, although Heinlein’s were mainly before the Nebulas existed.

By this most basic measure, then, Le Guin, Bujold, and Willis are the best living SF writers.  Over the last 20 years, 8 of the Hugo Best Novel winners have been women, and 11 of the Nebula winners.   For a field mainly known for rockets, rayguns, and boldly going where no man has gone before, it has become quite egalitarian.

Posted in Uncategorized | Tagged | 2 Comments