End of Work Reading List: Automation, Onshoring, Heart, and AI Texas Hold’em

Xconomy Seattle — 

One of Xconomy’s areas of focus in 2017 is the increasing impact of technologies such as artificial intelligence, robotics, and automation on human labor. Technology has always made some occupations obsolete, while changing others, and creating new ones. Today, and looking forward, these changes seem to be accelerating. This is the exponential economy from which Xconomy created its name.

There’s news every day and a lot of great stuff being written on this broad, essential issue. To help you make some sense of it, and to show you the sources that influence our thinking and coverage, we’ll be posting periodic digests: our End of Work Reading Lists.

In this inaugural EOW Reading List you’ll find: POTUS vs. PEOTUS on job creation and economic dislocation, with the former explicitly addressing the impact of automation and the latter so far ignoring it; The New Yorker’s distillation of research on automation’s role in the “reshoring” of manufacturing; Amazon’s big human hiring plans; and discussion of our innate human advantages—heart, instinct—and whether AI systems, including a poker-playing program going up against human experts this week, can effectively simulate those qualities.

—President Obama, in his farewell address to the nation Tuesday night, put his finger on what many believe will be the fundamental driver of job losses in the years to come:

“I agree, our trade should be fair and not just free. But the next wave of economic dislocations won’t come from overseas. It will come from the relentless pace of automation that makes a lot of good, middle-class jobs obsolete.

“And so we’re going to have to forge a new social compact to guarantee all our kids the education they need; to give workers the power to unionize for better wages; to update the social safety net to reflect the way we live now, and make more reforms to the tax code so corporations and individuals who reap the most from this new economy don’t avoid their obligations to the country that’s made their very success possible.”

The next morning, at his first news conference in 164 days, President-elect Trump took credit for announcements of U.S. factory expansions from automakers including Fiat Chrysler and Ford. “I think a lot of people will be following. I think a lot of industries are going to be coming back,” he said, before launching a broadside against the pharma industry. He added, “We’re going to create jobs. I said that I will be the greatest jobs producer that God ever created. And I mean that.”

—There’s a significant body of economic and technology research suggesting that manufacturers are returning to America because advances in automation obviate the need to go overseas in search of cheap labor.

A great summary of this research comes from The New Yorker staff writer Elizabeth Kolbert. In December, she distilled books by Martin Ford, a software developer, who wrote “Rise of the Robots: Technology and the Threat of a Jobless Future” (Basic Books); Jerry Kaplan, a computer scientist teaching at Stanford and author of “Humans Need Not Apply: A Guide to Wealth and Work in the Age of Artificial Intelligence” (Yale); and MIT researchers Erik Brynjolfsson and Andrew McAfee’s “The Second Machine Age: Work, Progress, and Prosperity in a Time of Brilliant Technologies” (Norton).

Kolbert quotes the MIT researchers who observe that much of the work sent overseas in the last two decades tends to be predictable, physical work—the easiest tasks to automate. She writes:

“Off-shoring jobs, they argue, is often just a ‘way station’ on the road to eliminating them entirely.

“In ‘Rise of the Robots,’ Ford takes this argument one step further. He notes that a ‘significant “reshoring” trend’ is now under way. Reshoring reduces transportation costs and cuts down on the time required to bring new designs to market. But it doesn’t do much for employment, because the operations that are moving back to the U.S. are largely automated. This is the major reason that there is a reshoring trend; salaries are no longer an issue once you get rid of the salaried.”

—In announcing plans to grow its full-time U.S. workforce by 100,000 jobs in the next 18 months to more than 280,000 people by mid-2018, Amazon made no mention of how many robots it would be adding during that period. It did say, “These new job opportunities are for people all across the country and with all types of experience, education, and skill levels—from engineers and software developers to those seeking entry-level positions and on-the-job training. Many of the roles will be in new fulfillment centers that have been announced over the past several months and are currently under construction in Texas, California, Florida, New Jersey and many other states across the country.”

At last count, late in 2016, Amazon reported having 45,000 robots, up 50 percent from the 2015 holiday season.

One more note from Trump’s press conference: He made sure everyone knew about his meeting with Jack Ma, chairman of Amazon’s China-based competitor Alibaba, saying at the press conference that the company’s interest in the U.S. is only the result of his election. “They would not be in anybody else’s office. They’d be building and doing things in other countries,” Trump said.

—A fundamental question, as machines perform tasks that are evermore humanlike, is what is it that makes us human?

New York Times columnist Thomas Friedman took up the question in a recent conversation with Dov Seidman, CEO of executive advisory firm LRN, and came up with this: “If machines can compete with people in thinking, what makes us humans unique? And what will enable us to continue to create social and economic value? The answer, said Seidman, is the one thing machines will never have: ‘a heart.’”

That hopeful message—that human value in the age of intelligent machines will be found in our ability to love and connect with each other, be compassionate, and dream—only goes so far. As Julia Bossmann, president of the Foresight Institute, wrote at the World Economic Forum last fall: We are at “the start of an age where we will frequently interact with machines as if they are humans; whether in customer service or sales. While humans are limited in the attention and kindness that they can expend on another person, artificial bots can channel virtually unlimited resources into building relationships.”

In other words, machines will likely be able to fake “heart” in many contexts. For evidence, look no further than Xiaoice, the Microsoft-built chatbot that is beloved by millions of its Chinese interlocutors.

—OK, if not heart, how about that good old human gut instinct—like when you just know your opponent is bluffing?

With Go and chess now mastered by purpose-built AI systems, the next game to fall to machines may well be poker. But unlike games in which all the information is visible to all players, poker is a different challenge, one of “imperfect information”—you don’t know what cards your opponent is holding. In the words of the great Kenny Loggins, it’s about “readin’ people’s faces / Knowin’ what the cards were / By the way they held their eyes.” It’s much more like real-world situations in that respect.

A tournament began this week in a Pittsburgh casino pitting a poker-playing program developed at Carnegie Mellon University against “world-class poker players,” reports MIT Technology Review. Another poker program, DeepStack, developed by computer scientists at the University of Alberta and two universities in the Czech Republic, has already shown success beating human players in head-to-head, no-limit Texas Hold‘em, in which each hand has some 10160 potential paths of play. DeepStack dealt with the complexity of the game “by applying a fast approximation technique that [the researchers] refined by feeding previous poker situations into a deep-learning algorithm,” writes Will Knight, MIT Technology Review senior editor for AI.

The researchers liken this approach to “a human player’s instinct for when an opponent is bluffing or holding a winning hand, although the machine has to base its assessment on the opponent’s betting patterns rather than his or her body language,” Knight continues. In the researchers’ own words: “This estimate can be thought of as DeepStack’s intuition. A gut feeling of the value of holding any possible private cards in any possible poker situation.” (Of course, a computer can calculate the odds of winning any hand much more quickly and accurately than a human.)

The CMU machine playing in the tournament is called Libratus and runs on hardware at the Pittsburgh Supercomputing Center, Knight reports. My question: Will the human players be able to look at the stacks of servers, watching for a tell when Libratus is bluffing?

Photo by Flickr user Morgan via a Creative Commons license.