Category: Economics

DeepMind Mastered “Go” Only After It Was Told the Score

(p. C3) To function well outside controlled settings, robots must be able to approximate such human capacities as social intelligence and hand-eye coordination. But how to distill them into code?
“It turns out those things are really hard,” said Cynthia Breazeal, a roboticist at the Massachusetts Institute of Technology’s Media Lab.
. . .
Even today’s state-of-the-art AI has serious practical limits. In a recent paper, for example, researchers at MIT described how their AI software misidentified a 3-D printed turtle as a rifle after the team subtly altered the coloring and lighting for the reptile. The experiment showed the ease of fooling AI and raised safety concerns over its use in real-world applications such as self-driving cars and facial-recognition software.
Current systems also aren’t great at applying what they have learned to new situations. A recent paper by the AI startup Vicarious showed that a proficient Atari-playing AI lost its prowess when researchers moved around familiar features of the game.
. . .
Google’s DeepMind subsidiary used a technique known as reinforcement learning to build software that has repeatedly beat the best human players in Go. While learning the classic Chinese game, the machine got positive feedback for making moves that increased the area it walled off from its competitor. Its quest for a higher score spurred the AI to develop territory-taking tactics until it mastered the game.
The problem is that “the real world doesn’t have a score,” said Brown University roboticist Stefanie Tellex. Engineers need to code into AI programs so-called “reward functions”–mathematical ways of telling a machine it has acted correctly. Beyond the finite scenario of a game, amid the complexity of real-life interactions, it’s difficult to determine what results to reinforce. How, and how often, should engineers reward machines to guide them to perform a certain task? “The reward signal is so important to making these algorithms work,” Dr. Tellex added.
. . .
If a robot needs thousands of examples to learn, “it’s not clear that’s particularly useful,” said Ingmar Posner, the deputy director of the Oxford Robotics Institute in the U.K. “You want that machine to pick up pretty quickly what it’s meant to do.”

For the full commentary, see:
Daniela Hernandez. “‘Can Robots Learn to Improvise?” The Wall Street Journal (Sat., Dec. 16, 2017): C3.
(Note: ellipses added.)
(Note: the online version of the commentary has the date Dec. 15, 2017.)

The paper by the researchers at Vicarious, is:
Kansky, Ken, Tom Silver, David A. Mely, Mohamed Eldawy, Miguel Lázaro-Gredilla, Xinghua Lou, Nimrod Dorfman, Szymon Sidor, Scott Phoenix, and Dileep George. “Schema Networks: Zero-Shot Transfer with a Generative Causal Model of Intuitive Physics.” Manuscript, 2017.

The paper, mentioned above, from the MIT Media Lab, is:
Athalye, Anish, Logan Engstrom, Andrew Ilyas, and Kevin Kwok. “Synthesizing Robust Adversarial Examples.” Working paper, Oct. 30, 2017.

Will Ending Firm Hierarchy Create “a Blissful Business Utopia”?

(p. 18) “The Kingdom of Happiness” doesn’t take place in Silicon Valley per se, but it is definitively about tech culture. Groth follows Tony Hsieh, the creator of Zappos, as he pours $350 million of his personal wealth into downtown Las Vegas with the goal of reinventing the area as . I won’t be giving away the story by pointing out that it doesn’t end well for Hsieh, . . .”
. . .
When she’s sober, Groth documents Hsieh’s attempt to integrate “holacracy” into his organizations, a term that rids a company of hierarchy and titles, and instead creates an all-for-one do-what-you-want mentality. (No, I’m not kidding.) It gave me a panic attack just thinking of working in a place like that.

For the full review, see:
NICK BILTON. “Denting the Universe.” The New York Times Book Review (Sunday, FEB. 19, 2017): 18.
(Note: ellipses added.)
(Note: the online version of the review has the date FEB. 14, 2017, and has the title “Pet Projects of the New Billionaires.”)

The book under review, is:
Groth, Aimee. The Kingdom of Happiness: Inside Tony Hsieh’s Zapponian Utopia. New York: Touchstone, 2017.

Rise in Cobalt Price Will Increase Quantity Supplied, and Increase Search for Substitutes

(p. B14) . . . the dreaded shortage of cobalt, which is used in the cathode of the batteries, is a bit more complicated than industry projections would suggest.
. . .
Like cobalt, rare earths aren’t so rare. China’s move to restrict exports in 2010 exacerbated the perceived shortage, sending the prices of some varieties up 10-fold. Companies such as Molycorp, Rare Element Resources Ltd. and Quest Rare Mineral Ltd., which all had some connection to reserves, saw their shares surge based on supposedly rosy prospects. Since then, all have lost nearly all of their value.
Already, Mr. Heppel explains, other users of the metal, for example in the pigments industry, are searching for alternatives. Meanwhile, some batteries, such as a design by Tesla, use less of the metal. Lower-performing batteries use none at all, and those batteries’ capabilities may improve with technological tweaks.
Supply will react too. Companies that operate copper and nickel mines, where cobalt is co-produced, are targeting expansion, and there are some pure-play cobalt mines being planned that could start producing shortly after the projected shortage hits.
For electric vehicles, this looks more like a speed bump than a cliff.

For the full commentary, see:
Spencer Jakab. “Will a Shortage of Cobalt Kill Electric-Vehicle Makers?” The Wall Street Journal (Thurs., Nov. 30, 2017): B14.
(Note: ellipses added.)
(Note: the online version of the commentary has the date Nov. 28, 2017, and has the title “Will Tesla Die for Lack of Cobalt?.”)

Only 5% of Jobs at Risk of Total Automation

(p. B6) About 15% of all hours worked globally could be automated by 2030 using technology that is currently available, McKinsey estimates. The new report builds on McKinsey’s earlier research, published in January [2017], which found that 60% of all occupations could be at least partially automated with current tools, though fewer than 5% are at risk of total automation.
Like prior waves of technological change, the adoption of new tools like machine learning and artificial intelligence will likely create more jobs than it destroys, says the Institute, the think-tank arm of consulting firm McKinsey & Co.

For the full story, see:
Lauren Weber. “Forget Robots: Bad Public Policies Can Kill More Jobs.” The Wall Street Journal (Thurs., Nov. 30, 2017): B6.
(Note: bracketed year added.)
(Note: the online version of the story has the date Nov. 28, 2017, and has the title “Forget Robots: Bad Public Policies Could Be Bigger Job Killers.”)

Supersonic Technology Constrained by Regulators

(p. B5) Japan Airlines Co. 9201 -0.09% has become the first carrier to invest in Boom Technology Inc., a U.S. startup seeking to build a faster-than-sound airliner capable of flying more than four dozen premium passengers to Tokyo from the West Coast in roughly five hours.
. . .
With a one-third scale version now scheduled to start flight tests in late 2018–nearly a year later than initially planned–JAL’s involvement is expected to influence cabin design and various operational issues. Blake Scholl, Boom’s founder and chief executive, said such cooperation is intended “to determine whether airlines will really be happy to have this airliner in their fleets,” including from a maintenance perspective.
. . .
Boom’s project has initial support from several venture funds and is taking an unusual approach by adopting various technologies already certified by regulators.

For the full story, see:
Andy Pasztor. “Supersonic Jet Gets Boost.” The Wall Street Journal (Weds., Dec. 6, 2017): B5.
(Note: ellipses added.)
(Note: the online version of the story has the date Dec. 5, 2017, and has the title “Japan Airlines Invests in Fledgling Supersonic Aircraft Company.” The online version differs significantly in wording from the print version. Where different, the passages quoted above, follow the online wording.)

Kid Paid $100,000 to Skip College and Mine Asteroids

(p. 18) As I sat down for lunch at a restaurant in Los Angeles, I placed a copy of “Valley of the Gods,” by Alexandra Wolfe, on the table, and a waitress walking by stopped to peer at the cover. . . .
“It’s about Silicon Valley,” I began. “It follows this young kid, John Burnham, who gets paid $100,000 by this weird billionaire guy, Peter Thiel, whom you’ve probably heard of; he’s a big Trump supporter and spoke at the Republican National Convention?” — a blank stare from the waitress. “Anyway, Thiel pays him (and a bunch of other kids) to forgo college so Burnham can mine asteroids, but he doesn’t actually end up mining the asteroids and. . . .”
. . .
The book begins with the protagonist, Burnham (or antagonist, depending whose side you’re on), who isn’t old enough to drink yet but is debating dropping out of college to follow the Pied Piper of libertarian and contrarian thinking, Peter Thiel, to Silicon Valley. As Wolfe chronicles, Thiel, who has a degree from Stanford University and largely credits where he is today (a billionaire) to his time at that school, started the Thiel Fellowship, in 2011, which awards $100,000 to 20 people under 20 years old to say no to M.I.T., Stanford or, in Burnham’s case, the University of Massachusetts, to pursue an Ayn Randian dream of disrupting archetypal norms.
It won’t be giving away the ending by pointing out that it doesn’t end well for Burnham.

For the full review, see:
NICK BILTON. “Denting the Universe.” The New York Times Book Review (Sunday, FEB. 19, 2017): 18.
(Note: ellipsis at end of second paragraph, in original; other two, added.)
(Note: the online version of the review has the date FEB. 14, 2017, and has the title “Pet Projects of the New Billionaires.”)

The book under review, is:
Wolfe, Alexandria. Valley of the Gods: A Silicon Valley Story. New York: Simon & Schuster, 2017.

Congestion Pricing Rises Again, as Crises Loom

(p. A18) For decades, urban planners, economists, city officials and business leaders have revived again and again some version of a toll system both to manage the city’s worsening traffic and provide more revenue for public transit. Over and over it was batted down, only to be resurrected, most recently in August when Governor Andrew M. Cuomo declared that “congestion pricing is an idea whose time has come.”
Now a state task force, called Fix NYC, has been assembled with the goal of developing another congestion pricing plan. It has been nine years since the last major effort by Mayor Michael R. Bloomberg died in Albany after state legislative leaders refused to bring it to a vote. Mr. Cuomo, after once expressing doubt about congestion pricing’s chances, is expected to unveil a plan early next year and make it a centerpiece of his legislative agenda.
This time congestion pricing is back at a moment of crisis — above ground, streets are becoming increasingly snarled in large part because of the boom in ride-hailing apps, while below ground the problem is even worse as the city’s aging subway system is riddled with delays and in dire need of money. The state-run Metropolitan Transportation Authority, which operates the subway, faces a litany of problems, including antiquated signals and overcrowded cars, that have led to frequent breakdowns — much of it documented by smartphone-toting commuters for the world to see.

For the full story, see:
WINNIE HU. “A Solution to Gridlock, Years in the Making.” The New York Times (Weds., NOV. 29, 2017): A18.
(Note: the online version of the story has the date NOV. 28, 2017, and has the title “New York’s Tilt Toward Congestion Pricing Was Years in the Making.”)

Hundreds of Thousands of Californians Moving to Texas, Arizona and Nevada

(p. A18) For more than three decades, California has seen a net outflow of residents to other states, as less expensive southern cities like Phoenix, Houston and Raleigh supplant those of the Golden State as beacons of opportunity.
. . .
. . . , for many Californians, the question is always sitting there: Is this worth it? Natural disasters are a moment to take stock and rethink the dream. But in the end, the calculation almost always comes down to cost.
. . .
California was once a migration magnet, but since 2010 the state has lost more than two million residents 25 and older, including 220,000 who moved to Texas, according to census data. Arizona and Nevada have each welcomed about 180,000 California expatriates since the start of the decade.

For the full story, see:
CONOR DOUGHERTY. “Californians Brave Fires, but Flee Cost of Living.” The New York Times (Weds., DEC. 13, 2017): A1 & A18.
(Note: ellipses added.)
(Note: the online version of the story has the date DEC. 12, 2017, and has the title “Quakes and Fires? It’s the Cost of Living That Californians Can’t Stomach.”)

Robots May Be a Threat After They Learn How to Open a Door

(p. A1) Robots may enslave us all someday. In the meantime, if one of them goes berserk, here’s a useful tactic: Shut the door behind you.
One after another, robots in a government-sponsored contest were stumped by an unlocked door that blocked their path at an outdoor obstacle course. One bipedal machine managed to wrap a claw around the door handle and open it but was flummoxed by a breeze that kept blowing the door shut before it could pass through.
Robots excel at many tasks, as long as they don’t involve too much hand-eye coordination or common sense. Like some gifted children, they can perform impressive feats of mental arithmetic but are profoundly klutzy on the playground.
The machines stumble over tasks requiring even toddler-level balance, like kicking a ball, getting out of a car or (p. A9) climbing stairs. Grasping objects of varying size and weight is also perplexing.

For the full story, see:
Daniela Hernandez. “If the Robot Apocalypse Comes, Try Closing the Door.” The Wall Street Journal (Sat., Nov. 11, 2017): A1 & A9.
(Note: ellipses, and bracketed date, added.)
(Note: the online version of the story has the date Nov. 10, 2017, and has the title “How to Survive a Robot Apocalypse: Just Close the Door.”)

Union Blocks Firing of Teachers Who Do Not Teach

(p. A1) Francis Blake has not held a permanent position in a New York City public school in at least five years. At his last job, in a Bronx elementary school, records show he was disciplined for incompetence, insubordination and neglect of duties — he had been caught sleeping in a classroom when he was supposed to be helping with dismissal.
Felicia Alterescu, a special-education teacher, has been without a permanent post since 2010, despite high demand for special education teachers. According to records, in addition to getting a string of unsatisfactory ratings, she was disciplined for calling in sick when she actually went to a family reunion. She also did not tell the Education Department that she had been arrested on harassment charges.
This month, Mr. Blake, Ms. Alterescu and hundreds of other teachers who are part of a pool known as the Absent Teacher Reserve could be permanently back in classrooms, as the city’s Education Department places them in jobs at city schools.
The reserve is essentially a parking lot for staff members who have lost their positions, some because of school closings and budget cuts, others because of disciplinary problems, but cannot be fired. It grew significantly as a result of a 2005 deal between the Bloomberg administration, which wanted to give principals control over hiring, and the teachers’ un-(p. A17)ion. Since then, the union has fiercely protected the jobs of teachers in the reserve, resisting attempts to put a time limit on how long a teacher can remain there.

For the full story, see:
KATE TAYLOR. “Caught Sleeping or Worse, Idled Teachers Head Back to Class.” The New York Times (Sat., OCT. 23, 2017): A1 & A17.
(Note: the online version of the story has the date OCT. 22, 2017, and has the title “Caught Sleeping or Worse, Troubled Teachers Will Return to New York Classrooms.”)

High Demand for STEM Workers Is Mainly High for Workers in Info Tech

(p. 10) A working grasp of the principles of science and math should be essential knowledge for all Americans, said Michael S. Teitelbaum, an expert on science education and policy. But he believes that STEM advocates, often executives and lobbyists for technology companies, do a disservice when they raise the alarm that America is facing a worrying shortfall of STEM workers, based on shortages in a relative handful of fast-growing fields like data analytics, artificial intelligence, cloud computing and computer security.
“When it gets generalized to all of STEM, it’s misleading,” said Mr. Teitelbaum, a senior research associate in the Labor and Worklife Program at Harvard Law School. “We’re misleading a lot of young people.”
Unemployment rates for STEM majors may be low, but not all of those with undergraduate degrees end up in their field of study — only 13 percent in life sciences and 17 percent in physical sciences, according to a 2013 National Science Foundation survey. Computer science is the only STEM field where more than half of graduates are employed in their field.

For the full story, see:
STEVE LOHR. “Where the STEM Jobs Are/Aren’t.” The New York Times, Education Life Section (Sun., NOV. 5, 2017): 10.
(Note: the online version of the story has the date NOV. 1, 2017, and has the title “Where the STEM Jobs Are (and Where They Aren’t).”)