Thoughts in Between

by Matt Clifford

Matt's Thoughts In Between - Issue #40

Forwarded this email? Subscribe here.

National culture and the weaponising AI

South China Morning Post has an extraordinary piece on a new Chinese education programme to train elite Chinese high school students in weaponising AI. Do read the whole thing, but here’s a representative excerpt(!): 

“We are looking for other qualities such as creative thinking, willingness to fight, a persistence when facing challenges,” he said. “A passion for developing new weapons is a must … and they must also be patriots.”

There's lots to think about here, but it reinforces how important national leadership and culture will be in shaping the character of AI in the coming decades. Ian Hogarth's essay on AI nationalism, which I've linked to before, remains the must-read post on the topic.

It's certainly true that the US DOD is exploring military applications of AI, but the prevailing attitude in Silicon Valley couldn't be further from training teenagers in AI warfare. If anything, Silicon Valley luminaries would prefer to keep their kids away from tech altogether, if we're to believe this (rather data-light) NYT article:

“The debate in Silicon Valley now is about how much exposure to phones is O.K.”

The comparison is unfair, of course, but it's illuminating.

Trump, the midterms and realignment

There's endless commentary on last week's US midterms, but the best piece I saw was this National Journal article that notes that the Republicans lost the suburbs in striking fashion, while shoring up support in small towns and rural settings. This represents a realignment relative to fairly recent US history and will reshape the electoral map for 2020.

In short, Trump is hugely popular among his base, but even more unpopular among everyone else. This means you have the highest midterm turnout for a century, without a Democratic wave (though that's debatable). David Leonhardt argues that populism is, well, popular - but racism is not, which may mean Trump's surprising 2016 midwestern wins are at risk in 2020. As the National Journal points out, given the booming economy, a "normal" Republican might be a shoo-in in 2020... But Trump is anything but normal and probably temperamentally can't "run normal", even if he believed it was the best strategy.

My two favourite titbits from the evening, though, were incidental to the actual results. First, the surprising finding that Texan natives actually leaned left and voted for the Democratic Senate candidate, Beto O'Rourke - but Republican Ted Cruz was saved by immigrants (albeit American ones). Second, that the markets watch Nate Silver so closely than when his model briefly (and incorrectly) predicted a Republican House victory, US bond yields shot up... The Singularity is near, indeed.

Has the West "lost it"?

There have been a number of heavyweight books in recent years on the rise of Asia and the new geopolitical balance of power. One that’s had less attention is Kishore Mahbubani’s Has the West Lost It? Mahbubani served as Singapore’s Ambassador to the UN, which, given Singapore's (see previous TiB coverage) unusual position bridging east and west, makes Mahbubani’s perspective particularly interesting. It’s both much more ideologically “pro-West” than many contributions to the genre, while also extremely critical of the practice of Western statecraft over the last 30 years.

The quickest way to get up to speed with the book’s ideas is this podcast interview. Mahbubani’s thesis is threefold: (1) that the West “matters” because its fundamental ideas have transformed the world (basically the Pinker worldview); (2) that the West “fell asleep at the wheel” at precisely the wrong moment in world history (around the fall of Communism and just as the Asian great powers were enjoying a renaissance); and (3) that the future success of the its values requires the US in particular to embrace ideas that it’s moved away from - multilateralism and “minimalism” in foreign affairs.

The podcast Q&A is fascinating, particularly the point that the most important geopolitical relationship is always between the top established power and the top emerging power - and in 30 years the US will be neither. If you believe that, then how the US behaves now and the structures it leaves behind will shape the 21st century. It’s an important contribution from a critical but sympathetic outsider. 

Quick links

  1. Heavyweight standard. We're changing how a kilogram is defined. It's surprisingly interesting.
  2. AI is literal. When AI does what you tell it, not what you want (Related - what AI means for "normal" people today)
  3. Facebook and chill. Social media's surprisingly positive impact.
  4. Everything is predictable. Your wrist tracker knows how you'll feel tomorrow.
  5. Never again. Extraordinary (and devastating) new photos of Kristallnacht

Your feedback

If you like this, forward to a friend or two and encourage them to subscribe. Thanks for reading!

Until next week,

Matt