3 min read

Global demographic stagnation

The world's declining population will have consequences; putting dodgy language models into humanoid robots is probably a bad idea; buy now, pay never set to join the financial innovation scrapheap; and how should the US respond to China's growing influence.
Cartoon from Japan showing "Oldzilla".
The world will soon be grappling with issues that were once confined to countries such as Japan. Source.

1—Global demographic stagnation

Global population growth in 2021 was the smallest in a half-century, "and by 2050, some 61 countries are expected to see population declines while the world's population is due to peak sometime later this century".

According to Joel Kotkin and Wendell Cox, that has consequences:

"We are well past the time when we need to concern ourselves with Paul Ehrlich's long-standing prophecy that humanity will "breed ourselves to extinction." On the contrary, we need to worry about the potential ill-effects of depopulation, including a declining workforce, torpid economic growth, and brewing generational conflict between a generally prosperous older generation and their more hard-pressed successors. The preponderance of low fertility in wealthier countries also presages a growing conflict between the child-poor wealthy countries and the child-rich poor countries."

Read the entire article here (~16 minutes).


2—Like bulls in a china shop

What happens when you combine massive pretrained language models and humanoid robots? Potential disaster:

"Google's new robot project PaLM-SayCan is incredibly cool. Humans talk, and a humanoid robot listens, and acts.
...
[However,] large language models are superficial statistical mimics, rather than systems that traffic in rich cognitive models of the world around them. Building a robot on top of language system that has such little comprehension of the world can’t be a recipe for success.

But that precisely what Google's new system is, a stitching-together of superficial and incorrigible language understanders with powerful and potentially dangerous humanoid robots. And as the old saying goes, garbage in, garbage out."

That's from Gary Marcus at The Road to AI We Can Trust. You can read the full article here (~8 minutes).


3—Buy now, pay never


4—The China trap

How should the US respond to "the challenge of a near-peer rival whose interests and values diverge sharply from those of the United States"? That's the question posed by Jessica Weiss in the Foreign Affairs magazine, and while we might query the premise – China nowhere close to being a "near-peer rival" to the US – there's no doubting its growing influence on the world stage.

Weiss warns that if the US takes the wrong approach, it risks falling "into the trap of trying to counter Chinese efforts around the world without appreciating what local governments and populations want":

"Lacking a forward-looking vision aligned with a realistic assessment of the resources at its disposal, it [the US] struggles to prioritize across domains and regions. It too often compromises its own broader interests as fractious geopolitics make necessary progress on global challenges all but impossible. The long-term risk is that the United States will be unable to manage a decades-long competition without falling into habits of intolerance at home and overextension abroad. In attempting to out-China China, the United States could undermine the strengths and obscure the vision that should be the basis for sustained American leadership."

Weiss argues that the US should promote "an inclusive, affirmative vision of the world", which would naturally "result in the United States opposing many of Beijing's activities, but that opposition would be accompanied by a clear willingness to negotiate the terms of China’s growing influence".

There's plenty more in Weiss' full article, which you can read here (~25 minutes).


5—Further reading...

📈 "Even 4 centuries of data isn't enough to tell us whether [total factor productivity] growth is linear or exponential."

👩‍🎓 Passing the New York State final algebra exam: "Instead of a 65% scaled score, you need a 50%. That is, instead of a raw score of 26/86 or 30.3%, you need 17/86, or 19.8%, to pass this test... A cat driving a car could get those 17 points. C for cat. C is choice 3 in the scoring key column here. Answer all Cs, you get 9 right. Each one is worth two points. That's 18 points, enough to pass."

☀️ Get some sun: A new study linked "insufficient" vitamin D levels to considerably higher SARS-CoV-2 mortality among unvaccinated Caucasian adults.