What should our leaders study?

An Odyssean Education, Wang’s Breakneck, and what we’re really missing

There’s a buzz around Dan Wang’s new book Breakneck. Its portrayal of China’s engineering-minded leaders, and America’s lawyerly ones, raises important questions: how should our political elites be trained? In particular, should the West emulate China and have more of them study engineering (or STEM)?

My take on this question: it’s less important what elites studied, and more important what their underlying political views are, and how their governments are structured.

Before we get into that, it’s worth revisiting the fullest expression of the STEM in government philosophy: An Odyssean Education, Dominic Cummings’s 200+ page case for transforming Britain’s ruling elite.

Evaluating An Odyssean Education

Writing in 2013, Cummings argued the world has shifted from simple systems to complex ones; from zero-sum, hunter-gatherer societies to market-based, technological, and modern nation states. Our political leaders aren’t cut out for this. They lack the knowledge and intellectual toolkit to make good decisions in the face of multifaceted and existential challenges. They don’t know enough STEM to succeed in the information age.

Cummings’s idea is to train synthesisers who can perform better in this world. A mixture of maths, physics, and biology, plus some political economy and philosophy – his Odyssean education – would create cross-disciplinary leaders with rare insight. They could understand increasingly important scientific issues and make good decisions about them.

The long document, written in 2013, is undoubtedly prescient given all that is now happening with AI. Written several years before the crucial ‘Attention Is All You Need’ paper, and the better part of a decade before the scaling era, An Odyssean Education would have left you looking in roughly the right place for the next big technological wave. The energy section, with its focus on solar, nuclear, and batteries, outlines some of the physical constraints now facing the AI revolution.

The biology section has proven a slower burn. We enjoyed the fruits of incredibly rapid vaccine development in 2021, and the impact of GLP-1 agonists is beginning to show up in American obesity rates, but, perhaps understandably, neither are mentioned. The paper does describe the fall in the cost of genome sequencing, which is having a huge impact on medicine. Links between specific alleles and traits are being drawn, enabling embryo selection and raising the spectre of designer babies. Noor Siddiqui’s much remarked upon appearance on Ross Douthat’s podcast is the opening salvo of a much wider debate, as foreseen in Cummings’s paper.

The biggest single mistake is the paper’s critique of orthodox economics in the aftermath of the 2008 crash. Cummings predicted that ‘agent-based modelling’ would take off instead, but I’ve never heard of it (nor have the friends I canvassed). You’re still better off learning regular, orthodox economics.

The promise and peril of the engineering state

There are situations where the skills Cummings outlined would be very helpful. The Covid pandemic is the most obvious and painful example, when a simple grasp of statistics eluded some of our top politicians and civil servants. Ironically it was Cummings himself who had to grapple with the consequences of a poorly trained elite, as discussed in his written evidence to the ongoing Covid inquiry.

Wang’s Breakneck outlines the positives of engineer leaders. Towering bridges in mountainous Guizhou are developing an impoverished and isolated state. Manufacturing at a grand scale in Shenzhen has lifted millions out of poverty and might have done the most to help China’s security by enabling it to build millions of drones. A state that fully believes in itself and in the goodness of its own mission, in contrast to the unambitious and self-abasing governments of the West.

But it also shows the negatives. Jailing hundreds of thousands of Uighurs to achieve political domination of Xinjiang province. A fixation on metrics leading to the murder of three hundred million children by abortion, thanks to the one child policy. A Zero Covid obsession leading to human misery at a grand scale, and threats to the legitimacy of the regime itself.

An engineering state with the wrong values is a bit like an unaligned AI – with the wrong reward function, it can pursue a line of reasoning and a course of action that ends up destroying a huge amount of value. Like an AI, the state won’t even be able to understand the full consequences of what it’s doing, as it only recognises one important indicator, like new Covid cases, number of children per family, or bridges completed. It’s high modernism on steroids.

An engineering state with the wrong values is a terrible sight to behold, shoving every part of society in the wrong direction. Taken in isolation, Cummings’s focus on the skills of our leaders only risks the same problem. Instead, we need our political class to have better values.

Getting the basics right

Much ink has been spilled over what good political values are. Let’s keep it simple: strong commitments to economic growth, individual freedom, and national identity.

At least in Britain’s case, technology, and political elites’ understanding of it, isn’t the most pressing problem right now: houses are undersupplied, energy is incredibly expensive, and whatever one thinks of immigration, it’s effectively out of control. Simply adding technical skills to our current politicians won’t solve these problems.

We can’t change the values of our leaders once they’re in office, as if this would even be possible. The real solution is for our higher education institutions – and other elite training mechanisms – to inculcate better values. A sufficient share of the Western public can be persuaded of these values to enact political change, if we try.

In a way, this is emulating China more than putting an engineer in charge of the country would. China is a communist country, and its elites put serious importance on Marxist-Leninist ideology. We should be highly committed to the values outlined above.

A good leader can get the country building, no matter what degree they studied. Lee Kuan Yew is the ultimate example – a lawyer by training, but a political leader who turned Singapore into a prosperous city-state. He didn’t need an engineering degree, as he could delegate road construction and port logistics and the planting of trees to those with technical expertise.

China’s leaders do not understand the finer details of the industries they’ve cultivated, and they don’t need to. It’s the wrong expectation for Western leaders too.

Time to think

Along with an updated set of beliefs, Western leaders would benefit from more time to think, the better to lead their respective countries. Here I focus on Britain, the political system I know best, but the principles apply to all governments.

From the Times last week:

[Baroness Casey] suggested to the prime minister that he might emulate his chancellor in delegating to a trusted enforcer the daily grind… [and] devote his focus to the “big picture”.

As Substack In the Sight of the Unwise recently argued, the machinery of government in Britain makes it incredibly difficult to actually govern. Details are not filtered out by the time they reach the prime ministerial desk or meeting agenda. More officials, or a handful of deputies, would lead to a better signal to noise ratio in the PM’s information flow, leaving him or her time to focus on other matters.

The time won back from the system could be spent thinking. The principal political leader of a country needs to have time to zoom out and consider matters that would be lost in the day-to-day. We don’t need that leader to be able to explain the finer details of a neural net; we do need them to understand the wider significance of the scaling era. (Studying history, not STEM, is perhaps the best degree background for this task.)

The Politburo Standing Committee in Beijing – China’s top leadership – holds regular study sessions, learning about the issues du jour from top academics. These sessions are used to set priorities across Chinese government and society.

John Bew, chief foreign policy adviser to four prime ministers, has proposed mandatory reading weeks for mid-career national security officials, and the return of essays to force people to really think through the important issues of the day. Why not senior career officials across many departments?

The point is not to emulate authoritarian systems, but to recognise that our top political leaders should have a government under them who take care of the details, allowing the leaders (with a strong belief in the values outlined above) to set the overall direction of the country and communicate it to the public – whether they’re engineers or not.