Sep 25, 2018

Axios Future

By Bryan Walsh
Bryan Walsh

Welcome back to Future. Thanks for subscribing. Consider inviting your friends and colleagues to sign up. And if you have any tips or thoughts on what we can do better, just hit reply to this email or shoot me a message at steve@axios.com.

Okay, let's start with ...

1 big thing: Still missing — a U.S. AI policy

Illustration: Lazaro Gamio/Axios

U.S. policy in the race to dominate the world’s most advanced technologies remains chaotic, pushing ahead in some areas, while others languish.

Axios’ Kaveh Waddell reports: The global contest in artificial intelligence and quantum computing, for instance, is in its most crucial stages. But experts say the U.S. lacks China's focus on winning everything.

In a report published today, the leaders of the House Oversight and Reform IT subcommittee — Republican Will Hurd and Democrat Robin Kelly — urge the government to step it up:

"The United States cannot maintain its global leadership in AI absent political leadership from Congress and the Executive Branch."

But U.S. political leaders haven't moved with the urgency the situation requires, Hurd tells Axios. China’s rapid rise in AI should shock the U.S. into action, Hurd and Kelly write.

The pair lays out high stakes for failure. "Whoever masters AI will have an outsize role in this world," Hurd says. Their report offers four recommendations:

  • Increase funding for R&D through agencies like the National Science Foundation, National Institutes of Health, Defense Advanced Research Project Agency (DARPA), and NASA.
  • Publish government data sets, a potential boon for training data-hungry AI.
  • Develop standards for measuring the progress and dangers of AI.
  • Plan more DARPA Grand Challenges, like the one that motivated much of the early work on autonomous driving.

Earlier this month, DARPA announced a $2 billion investment in research into more flexible and powerful AI. That’s progress, Hurd and Kelly write, as was a May summit on AI at the White House Office of Science and Technology Policy.

The bottom line: "There's not the level of interest and urgency and immediacy that we need from government right now," says Paul Scharre, director of the technology and national security program at the Center for a New American Security. "There is no national strategy."

  • The executive branch needs to take the lead, Scharre says, especially in driving funding.

Go deeper: Read Kaveh's whole post.

2. China's dug-in supply lines

Huangpu port. Photo: VCG/Getty Images

The Trump administration is not only seeking what it calls fairer trade with Beijing, but much more — to upend the bedrock of the Chinese economy by forcing the chain of manufacturing supply out of the country and pushing it elsewhere.

But Axios' Erica Pandey writes that there are doubts about this end game of Trump's trade war. After decades of development and tens of millions of dollars of investment, few U.S. companies seem likely to move their manufacturing facilities out of China.

"It can’t be a replay of the Cold War division between the West and the Soviet Union — industries were nationally based back then and innovation wasn’t global.  All this changed after 1990."
Jim Lewis of the Center for Strategic and International Studies

China has been the world's biggest exporter for almost two decades and has poured millions into its logistics network so global companies can quickly move goods from factories to cargo ships. China also has one of the best-trained manufacturing workforces in the world.

  • U.S. companies can consider moving production from China to Cambodia, Malaysia, Vietnam or elsewhere — both to dodge tariffs and avoid the threat of intellectual property theft. But elsewhere, they face disadvantages like dirt roads between factories and ports and inexperienced workers, the NYT reports.
  • And China has levers to make sure it isn't frozen out of global supply chains, Lewis says. For example, it could threaten to take away a foreign company's access to its massive market.

Yes, but: Some companies are at least putting a halt to new Chinese production capacity.

  • Feng Tay Enterprises, a Taiwanese footwear manufacturer that services Nike and Adidas, stopped building up Chinese production over 10 years ago, per Nikkei Asian Review. It is now adding capacity in India and Southeast Asia.
  • Foxconn, the iPhone and laptop maker, earned kudos from Washington for promising to bring a chunk of production to Wisconsin.

What to watch: If a few big names try to cut reliance on China — even if they don't move out entirely — President Xi Jinping could make concessions to avoid disrupting the interconnected supply chain.

  • For now, that's unlikely to happen with Xi. "He's overestimated China’s ability to make advanced technology without Western help, but it will take a while for him to admit this," Lewis says.
3. When a few simple rules would do


Photo: Ariel Skelle/Getty

Eager to use AI, some companies may be unnecessarily complicating easy business problems.

Kaveh writes: Companies are over-using complex AI techniques when they would be better served with simpler approaches. Rule-based systems, for instance, show their work, thus allowing non-experts to pop the hood and see why an algorithm is misbehaving, unfair or biased.

Deep learning, the most advanced AI technique of the day, approximates how our brains work, with layers of "neurons" that together identify patterns. The catch is that their reasoning is opaque.

  • That means a rudimentary deep-learning system for doling out loans could illegally deny applicants based on their race, just because it noticed that African-American borrowers default on loans more often than white borrowers. But it would be hard to know that was the reason behind the system's denial.
  • By contrast, if they relied on a rule-based system, they would be using an explicit set of instructions for making decisions ("if x is true, then do y"), making any flaws easier to ferret out.
  • With rule-based systems, a user can work to deliver the best outcome — and then check precisely how it's doing.

Today, Adobe Research told Axios they now have a system that speeds up one technique for finding a good group of rules.

  • The result is a 10,000-fold decrease in the time it takes to find a good decision set, allowing the use of far bigger databases.

"There may be a problem for which accuracy is the only guiding principle," Adobe data scientist Ritwik Sinha says. But when decisions directly impact human lives, an increase in transparency may be worth a decrease in accuracy.

4. Worthy of your time
Expand chart
Data: World Bank; Note: South Asia's latest value is from 2013; Chart: Andrew Witherspoon/Axios

Extreme global poverty keeps shrinking (Stef Kight — Axios)

Navigating the age-old crisis of immigration (Reihan Salam — WSJ)

How the octopus got its smarts (Elizabeth Finkel — Cosmos) (h/t Jacob Feldman)

The coders of Kentucky (Arlie Hochschild — NYT)

Drowning in the fountain of youth (Emily Mullin — Youth, Now)

5. 1 reading thing: books and sausages

And a sausage. Photo: Jay Colton/LIFE/Getty

In Germany, like the U.S., book readers have become fewer as increasing numbers of people are drawn to shorter offerings online. But the NYT's Sally McGrane finds a small bookstore in the Hessian town of Bad Sooden-Allendorf whose owner has an answer for keeping his customers:

Sausage and bread.

What's going on, per NYT: In 2013, facing dwindling revenue, Wolfgang Fruhauf eagerly pounced on a local baker who had lost his lease and needed a new place to produce bread rolls. Then he added sausages, eggs, tomatoes and cucumbers.

  • As a result, a new set of clientele began to frequent Frauhauf's Bookstore, which goes back a century in the family. Along with their food orders, some customers began picking up a book or two.

The bottom line: Fruhauf found a way to save his store. McGrane quotes a German official explaining why it reflects a model for how to resurrect dying rural business centers elsewhere in the country. "Suddenly there is a meeting place, again.”

Bryan Walsh