Negotiating Our Economic Future
Technology, particularly AI, is changing the world. It's changing how trade and diplomacy work too, and both concepts have a role to play in mitigating the risks.
Last month I sat down with Geoff Pigman, an expert in international trade and diplomacy, to discuss how technology was changing his fields of work, and how his work can help us mitigate the defining challenges of the 21st century. The episode is on Spotify and you can grab a free copy of the book by DMing Geoff on Twitter.
Blending the academic perspective with the business one, Geoff is an excellent person to talk to about trade. After his PhD at Oxford, he lectured and has received fellowships at universities across the world, whilst he’s spent the other half of his career advising businesses, governments, and NGOs.
We spoke about three big topics: how trade and diplomacy intersect, how technology is changing this, and how trade and diplomacy can help us mitigate today’s big challenges.
Tl;dr:
Trade matters and is becoming increasingly important
Technology is creating enormously exciting potential future worlds – an enormously terrifying ones
Diplomacy matters to managing these risks
DM Geoff on Twitter for a free copy of the book (excl. postage)
How trade and diplomacy works
Diplomacy is how countries interact with each other. They negotiate, exchange, and make agreements either bilaterally (one with another) or multilaterally (as a group). And it’s important to remember that how countries negotiate, their agendas and preferences, depends on their own context. Specifically, countries are concerned about keeping themselves safe, satisfying their people (particularly but not only in democracies), and the concerns of individual negotiators and leaders (particularly but not only in autocracies).
And whilst it’s pretty clear how diplomacy affects trade – countries make deals on tariffs and regulations like via the World Trade Organization – it’s also the case that trade affects diplomacy. When businesses from different countries trade, they become invested in the revenue and jobs their businesses generate from exports as well as the productive purposes that they use their imports for (e.g. raw materials or food).
Ecosystems of businesses, employees, and consumers build up that depend on trade relationships over time. European consumers have grown to depend on Russian oil, Western businesses are very keen to sell to the enormous Chinese middle class, British employees whose jobs depended on trade with Europe were highly opposed to Brexit.
By changing the terms of trade that your country has with another, you can thus affect how countries negotiate, creating a direct link from trading relationships to diplomacy more broadly. Hence why Western countries and businesses have sanctioned and boycotted Russia since the country invaded Ukraine: to harm Russia’s economic interests (such as the collapse in value of the Ruble and excluding Russia from SWIFT) and push Russian oligarchs and consumers to oppose the consequences of Putin’s invasion, and thus the invasion itself.
Technology is making this more complex
Of course, Russia is invading Ukraine despite this; there are many factors lying behind diplomatic outcomes beyond trade. And technological developments over the past half century have further complicated the trade-diplomacy relationship.
Technology has allowed a lot more international trade. Over recent decades, it has become dramatically cheaper to ship goods from one continent to the other, whilst the internet has made it much easier to buy things from abroad. Trade made up 15% of global GDP in 1993; twenty years on in 2013 it reached 25%. Last year it reached $29 trillion, just over 30%. Ecommerce accounts for one seventh of that, now a $4.2 trillion that is only possible with the internet. You would have thought that this would give countries more ability to influence each other….
But technology has also made it harder for countries to use trade to affect diplomacy. Spending on digital products and in digital ecosystems like Ethereum, where people transacted more money than Visa last year ($11.6 trillion), has ballooned; software and cloud based services alone were worth $396bn as a sector in 2021. And when you’re spending online, it’s much harder for governments to reach you, which makes trade a more blunt tool for affecting diplomacy.
Technology creates new actors, because many technological businesses tend towards large monopolies, which can wield enormous power. The most profitable and valuable companies in the world are Microsoft, Apple, Alphabet, Amazon, and Meta (excluding financial services firms and state-owned enterprises)1. Tech firms naturally evolve into monopolies because in their respective industries, they build digital platforms that consumers come to love and depend on, creating barriers to potential competitors.
On Windows, you have everything from Word to OneDrive to your work emails. Google is your personal basic executive assistant (e.g. search, email, calendar and maps, and images and YouTube), all bundled with Chrome and Android. If a retailer wants to sell to consumers, they have to be on Amazon. Facebook is noticeably smaller than those titans, but it is the influential social network because no user wants to set up an account on something else that doesn’t have their friends2.
In addition to making global platforms possible, technology gives us access to data of enormously higher quantity and quality. In turn, this allows organisation which possess that data to optimise their services better than others, and in turn attract more users, giving them more data. Hence why everyone still uses Google Search and why Meta and Alphabet still dominate digital advertising. (As an aside, it’s interesting that the only two of these companies which have opted to change their names are those built around monetising your personal data.)
The upshot of this is that the business world is more concentrated than it has been in decades (in China, Baidu, Alibaba, and Tencent are the tech monopolists instead)3. And as these firms grow, so does their importance as global actors. With valuations larger than many country’s GDPs, these firms are able to leverage governments to extract better deals, at a scale of influence that few companies have ever had, let alone a handful all at once. In the US, the city of Dallas offered to build a dedicated airport hangar for Amazon if the company chose to build its new HQ in the city; New Jersey offered the company $7bn cash.
Just like national governments, firms also have their own agendas and preferences, and now have the weight to force countries to negotiate with them in meaningful ways. In the past few months for example, Meta has threatened to block Facebook services in Australia over news content and the EU over data privacy. Countries are no longer just negotiating with each other, but with businesses as well.
Overall, I think governments have more diplomatic influence through trade than in recent decades. The growth in trade is genuine and here to stay, even if you ignore the growth areas that have that added complexity. Meanwhile, growth in influential businesses is genuine but less likely to stay in the same way: companies fight each other a lot more often than they fight governments, which operate leagues above them in terms of legitimate power regardless. So trade is still worth thinking about, not just because it has become more complex, but because it has become more important.
Technology as solution and technology as problem
As well as changing how trade and diplomacy works, technology can help us mitigate the big challenges of the 21st century. Specifically, Geoff focuses on climate change and how solutions lie at the intersection of diplomacy, trade, and technology; I agree that technology has a huge role to play. I think it’s unlikely that the solution to the climate crisis is one that demands everyone reduces their energy usage and quality of life. Rather, the solution is creating clean and efficient energy sources which eventually pave the way to energy abundance, rather than energy scarcity.
We’re already on the way. Solar, wind, fusion, geothermal, and hydrogen energy technologies have all advanced significantly even in the past decade. Researchers at the International Renewable Energy Agency (IREA) found that the price of solar power, globally, has fallen by over 80% since 2010. In that time, solar power has gone from being too pricey basically everywhere to being incredibly cheap basically anywhere: building new solar plants is now cheaper than building new fossil fuel plants in India, Chile, Mexico, Spain, and in US states like Arizona, Nevada, Colorado, and Texas. Since 1976, the price of solar energy has fallen by 99.6%, and installations have soared: solar is now the fastest growing energy source around the world.
Other renewable sources can have similar impacts. The same IREA study found that offshore wind costs have dropped by 70%, and onshore wind energy generation costs have fallen by a similar amount. What lies behind these dramatic falls? Wright’s law: the notion that for every time the number of manufactured units of a technology doubles, its price falls by a constant proportion. The idea is that every technology has a ‘learning curve’, where as more and more people produce a technology, like ‘solar panels’, they figure out better and better ways to do it, and share their knowledge with the world. New technologies in hydrogen fuel, geothermal energy, energy storage and batteries, and carbon removal near the top of their learning curves right now, and are experiencing rapid price declines.
This isn’t to say that the climate crisis is easy to solve. Agriculture and land use contribute a quarter of greenhouse gas emissions and aren’t as easy to solve. Cows should scare you as much as than coal. How we heat our buildings is another major challenge. But technology is completely changing the game when it comes to climate change.
Geoff also outlines how enormous attention needs to be paid to artificial intelligence. AI has continued to progress leaps and bounds and continues to amaze. Most recently, OpenAI have released DALL·E 2, an AI system that can create realistic images from text. It’s incredible.
As with art, AI has the potential to create an incredible future. DALL·E 2 is an example of artificial narrow intelligence (ANI), i.e. AI that is really good at a narrow group of tasks, like converting text into art. For something that’s meant to be narrow, ‘creating art’ is pretty broad, but it’s not as broad as artificial general intelligence (AGI), which is the theoretical concept of machine being able to understand the world as well as a human does, and learn any task. This includes Superintelligence, a computer that is much smarter than any human in practically every field, and could contribute enormously to human prosperity.
Geoff writes that “technological utopia will not happen by itself” — we have to work hard to create a better future — but technological dystopia easily could come by accident. Humanity’s superior intelligence is pretty much the sole reason that it is the dominant species on the planet. If machines surpass humans in intelligence, then just as the fate of gorillas currently depends on the actions of humans, the fate of humanity may come to depend more on the actions of machines than our own. It is not hard to imagine how a superintelligent computer could have interests which differ from those of humankind4.
A more important problem is that of goal alignment, which careers organisation 80,000 Hours outlines particularly well here. Whilst superintelligent general AI systems are a theoretical concept, unsafe applications of AI are possible with a narrow AI system that simply misunderstands the task it has been set. The following (highly-simplified) example comes from that link and illustrates how goal misalignment, with or without superintelligent AI, could cause enormous problems:
A pharmaceutical company uses AI to generate new organic compounds.
As the AI improves, it becomes increasingly impractical to keep humans involved in its work – the humans’ ideas are usually worse anyway. As a result, the AI is granted more and more autonomy to run experiments on new compounds.
Eventually, the AI is given the goal to “reduce the incidence of cancer” and offers a compound that initial tests show is highly effective at preventing cancer. Several years pass, and the drug comes into universal usage as a cancer preventative…
…until one day, years later, a molecular clock in the compound causes it to produce a potent toxin that suddenly kills anyone with the substance in their bodies.
It turns out the AI had found that the compound most effective at reducing cancer rates was one that killed humans before they could grow old enough to develop it in the first place. The AI also predicted that its drug would only achieve this goal if it were widely used, so it combined the toxin with a helpful drug that would incentivise the drug’s widespread adoption.
Of course, the concern isn’t about this example specifically, but rather unintended consequences more generally. Back in 1908, the New York Times reported on a dog that regularly pushed children into the Seine, only to rescue them for the treats he would receive at the end. When it comes to AI alignment, all of humanity are the children, the dog is the AI, and the stakes are much higher.
If we get pushed in the river, we might not get pulled out. So if we want to ensure we get to enjoy a lovely walk by the water, we have to ensure we don’t get pushed in.
Advances in artificial intelligence could lead to extremely positive developments, presenting solutions to now-intractable global problems, but they also pose severe risks. Nick Bostrom warns of “a society of economic miracles and technological awesomeness, with nobody there to benefit … a Disneyland without children.”
All it takes is for one superintelligent machine to receive a poor instruction for it to pose a large risk. That becomes much more likely without effective diplomatic relations between states to ameliorate any potential AI arms race. The capabilities of AI are well known: as Putin has observed from Russia, “whoever becomes the leader in (AI) will become the ruler of the world”; countries may take excessive risks, cutting corners on safety and alignment, in an effort not to be left behind.
Technology offers immense opportunities for the future. Immensely positive visions of that future are already being built. It is eternally up to us to ensure that we maximise the benefits of technological change whilst minimising the bad.
Edit, July 2023: I have written a follow-up to the idea that technology creates a plethora of potential futures, with an analysis of how we ensure we get the best possible one.
Meta (Facebook), Tesla, and Nvidia rank highly on valuation but have lower revenues; Samsung and the Taiwan Semiconductor Manufacturing Company are also up there with higher revenue but lower valuations. Berkshire Hathaway and JP Morgan have higher valuations and revenues but are financial conglomerates, whilst Saudi Aramco, the oil company, is state-owned; the Industrial and Commercial Bank of China is both.
Facebook’s ‘network effects’ have eroded for two reasons. Primarily, users began to use social media for more than keeping up with friends, mainly getting news and entertainment. This means you don’t need your friends to be on Instagram, TikTok, or Snapchat to download it, you just need journalists and creators from anywhere in the world to be there. Second, as people began to use social media for more things (with friends, news, entertainment), social networks began to specialise into providing certain uses, and networks that tried to be ‘catch-all’ were out-competed by specialists in every niche. The future of social media is conglomerates which provide multiple services and cross-leverages user data to optimise the user experience. It’s also going to be decentralised, but more on this next month.
Edit, July 2023: this month Meta released Threads, a Twitter-clone (algorithmic feed, text-based content) that became the fastest-growing app in history, growing to over 100 million users in a matter of days. It’s proof of how tech platforms tend to monopoly, and my claim of how the future of social media is conglomerates which provide multiple services (Meta runs Facebook, Instagram, and Threads) and cross-leverages user data to optimise the user experience (Threads users could import their data across from Instagram). For what it’s worth, Threads is also planning to build on the decentralised social network protocol, ActivityPub.
Industry consolidation is happening everywhere, it’s just in tech sectors it leads to supersized companies that can dictate to governments. This actually makes me optimistic, because it suggests that the dominance of technology firms is more the result of a temporary trend than anything inevitable, and that regulators have half simply been caught off-guard about how platform power and data leads to monopoly. More broadly, technology does not necessarily lead to consolidation.
I’ve mentioned OpenAI a lot in this post. I want to stress that they are one of many organisations working on AI and one of those most aware of this alignment problem. As they note, “solving this problem is of critical importance if we want life to go well.” They have an excellent video and explainer on their page about alignment here. I don’t believe that narrow AI models like DALL·E 2 or GPT-3 give us much reason to worry about AGI.
Great stuff. Quality content. The reason I subscribed. Brief balanced and to the point piece on the state of geopolitics and technology. I am intrigued to read your next one.