Forwarded this email? Subscribe here. Enjoy this email? Forward it to a friend.

This week a new paper in Nature (see here for ungated coverage) shows that natural language processing (NLP) techniques can be used to read an entire corpus of journals and make discoveries years in advance of published research. Looking at materials science, the authors trained a model to “discover” compounds that have interesting properties.

People have applied machine learning to materials science before, but this required a huge amount of labour to create (small) labelled datasets, which include a tiny fraction of our knowledge of materials. The NLP method avoids this: it’s completely unsupervised and so can read all published articles in the field without human input. The technique used is based on Word2Vec - a kind of model that encodes word meanings as vectors, so that we can see things like the “closeness” of related concepts. This allows us to perform what you might crudely call word arithmetic - e.g. we can ask what is “King - Man + Woman” and get the answer “Queen”! (See here for some fun examples). 

It’s exciting to think that so much opportunity for discovery is already embedded in existing knowledge - not just in materials, but in medicine, engineering and more. The future is already here; we’ve just not (machine) learned it yet.

The ongoing US-China-Huawei struggle hinges on whether it’s dangerous to allow key technology infrastructure (i.e. 5G) to be built by a firm with close ties to the Chinese state. This week a controversial paper from Christopher Balding (the US academic whose note on leaving China was one of the most-read items in TiB last year) claims to show significant overlap in personnel between Huawei and Chinese intelligence organisations. 

Balding’s written more about this here and there’s coverage in Forbes here. It’s not remarkable that Huawei would hire ex-military personnel. What’s striking is Balding’s claim that a number of Huawei employees’ CVs match at the metadata level with specific incidents of known state-sponsored industrial espionage - that is, they list the right expertise, in the right place, at the right time. Not everyone is convinced (see Balding’s reply), but it’s further ammunition for the Huawei hawks.

As I’ve said before, the situation is a perfect microcosm of so many of questions the world will face in the coming decades as technology and geopolitics collide. It’s sometimes hard to keep up: last week Trump seemed to signal a u-turn, a few weeks after having banned US companies from exporting tech to the Chinese firm - a move that Huawei says could cost it $30bn of revenue. But Congress may block any softening; being anti-Huawei is one of the few bipartisan positions left in America.

Modern Monetary Theory (MMT) is one of the most controversial ideas in economics today. It’s complicated (see here for a good explainer), but the basic idea is countries that control their own currencies should be willing in certain circumstances to print money to increase public spending. You can imagine why that’s a popular idea in some parts of the left and very unpopular with mainstream economists.

It’s surprising, then, that MMT has found a (qualified) supporter in Ray Dalio, the billionaire founder of Bridgewater, one of the world’s largest (and oddest) hedge funds. In a recent pair of articles, Dalio first argues that American capitalism is no longer working for most people (above all, inequality is rising and opportunity falling) and second that MMT - or a variant that he calls Monetary Policy 3 (or MP3) - may be part of the answer.

Dalio’s argues that during the next economic crisis we’ll likely still be at close to zero interest rates, so classic monetary policy (“MP1”) won’t work and quantitative easing (“MP2”) is regressive and a causes of growing inequality. He thinks coordinating fiscal and monetary policy - i.e. printing money to invest in infrastructure and education - may be the answer. Even more surprising, though, is that when a fellow billionaire, John Mauldin, chimed in to disagree, it was to decry MMT - but also to endorse Dalio’s claim that capitalism isn’t working. Interesting times. 

Thank you for reading. Thoughts in Between is growing fast, thanks to your referrals - please do forward to a friend who might enjoy it. And just reply if you have any comments or talk to me on Twitter.

Until next week,

Matt