Three years of Noahpinion

What happened to the world in 2023, and what I wrote about it

Photo by Xiang Gao on Unsplash

I’ve been blogging since 2011, but the Substack version of Noahpinion was created on November 23, 2020. Thanks to all of you amazing readers, Noahpinion has experienced steady growth — it’s now up to around 167,000 email signups and 11,500 paid subscribers. This is well beyond my initial goals, and a bigger audience than I ever expected to have. All I can say is that I’m incredibly grateful, and that I’ll do my very best to keep providing you all with analysis that’s both useful and entertaining.

This year has seen a few changes at Noahpinion. My list of things to write about was getting too long, so I started doing quasi-weekly roundup posts. I also started a new podcast called Econ 102, along with Erik Torenberg (the name comes from the fact that I was a teaching assistant in his Econ 102 class, long ago). This is in addition to my existing podcast Hexapodia, with Brad DeLong, which has slowed down due to Brad’s work schedule, but is still going. All podcast episodes of mine will be posted in the roundup posts. Finally, Noahpinion switched from the Substack domain to a custom domain, www.noahpinion.blog. I like the “.blog” domain because it reminds people that the blogosphere still exists, and in many ways is healthier than ever.

Anyway, every year on Noahpinion’s anniversary, I do a post reviewing seven big throughlines from the past year. The links below are all links to other posts I wrote over the past 12 months. So you can use this post as a sort of summary or index, especially if you want to know what my main topics of interest have been.

As usual, there was no shortage of big events to write about.

AI’s big coming-out party

I’m a big techno-optimist, but even I was surprised by the impact that generative AI has had over the last year. Chatbots had been progressing slowly for a while, but the release of ChatGPT on November 30, 2022 was a game-changer — a robot that talked like a human. And that was followed up pretty quickly in March with GPT-4, a robot that talked like a fairly smart human. At the same time, there were huge advances in AI art, with apps like Midjourney. There’s a distinct sense that the world is not the same as it was a year ago. We’ve lived through radical technological change our whole lives, but AI is a different sort of beast entirely. In some ways, it’s almost like magic — a whole new way of controlling the world without necessarily understanding the mechanisms of control.

Much of the discussion around AI has centered on the risk that it’ll attack and perhaps even destroy humanity. I doubt the current generation of models are capable of this, but what do I know? What does anyone really know? There’s not a lot most people can productively say about this “X-risk”, which of course hasn’t stopped people from saying quite a lot.

But one topic where an economics writer like myself hopefully can add productively to the discussion is the risk of massive job loss from AI. There’s a widespread fear that generative AI will replace human capabilities in such a wide variety of tasks that many humans become either completely unemployable, or employable only for starvation wages. This idea has been embraced by a lot of top AI engineers and some very prominent economists as well, but personally, I don’t think this is likely. OpenAI engineer Roon and I co-authored a post where we argued that the principle of comparative advantage implies that no matter how good AI gets at anything and everything, there will always be some valuable work for humans to do, even if the nature of that work changes. It’s theoretically possible for wages to go down as a result of automation, but the apocalyptic “obsolescence” scenario is just highly unlikely.

It’s also possible that AI might give wages a big boost, similar to what machine tools did for factory workers a century ago. One of the most consistent findings in the research economists have done on generative AI so far is that it’s a skill leveler — whether it’s coding or call centers or college essays, low-skilled workers get a much bigger boost from AI than high-skilled ones. This could open fields like software engineering to the masses. The fact that Americans mostly aren’t thinking about these positive scenarios is probably a form of collective PTSD from decades of rising inequality, in which new internet and computing technologies seemed to benefit high-skilled workers and managers while leaving everyone else behind. AI won’t necessarily be like that. Thus I think it’s a mistake for economists and policymakers to think they can “shape” the direction of AI innovation in order to replace fewer workers.

There is one area, though, where generative AI has so far been a disappointment to me personally — I can’t really use it to write. Even GPT-4 often imagines fake sources and facts, and gets big concepts wrong in its eagerness to sound authoritative. I hope this “hallucination” problem gets fixed soon, so I can spam the entire world with Noah Smith Thought. In the meantime, I do use GPT-4 to proofread my posts for typos, and it does a good job.

China finally stumbles

If AI was the biggest technology story of the past year, China’s slowdown was the biggest economic story. For decades, China had been the motor of global economic growth — an unstoppable engine of industrialization that sucked in raw materials, flooded the world with cheap and ever-more-sophisticated products, made a few lucky foreign capitalists rich, and put a lot of unlucky foreign workers out of a job. In recent years, China has accomplished industrial feats on a completely unprecedented scale, building the world’s largest high-speed rail network (by far) in just a few years, dominating global auto exports, and building more renewable energy than anyone else. Much of this was due to its locally integrated supply chains and its deep pool of engineering talent.

But there was always trouble brewing. In the high-tech realm, China’s industrial policies often reduced productivity instead of increasing it. Its “Belt and Road” initiative was mostly a failure, leaving many other countries disgruntled and in debt. Domestically, its growth was ever more dependent on a bloated real estate industry. And other countries were getting increasingly worried about Chinese power and military aggression, leading them to start trying to decouple their economies from China — an effort which has made gradual but real progress.

In 2023, those problems could no longer be papered over. China’s real estate industry, which had begun collapsing all the way back in 2021, finally started taking the rest of the economy with it. The government seems unwilling to do the sort of consumption-boosting fiscal stimulus that many observers recommend, and instead seems focused on slowing the collapse of real estate. But this seems likely to simply prolong the pain and entrench pessimistic expectations, especially among an increasingly disaffected young generation.

It looks as if China’s long, spectacular rise is reaching a plateau. I don’t expect the country to collapse into rapid decline, but I do think attention and money will start shifting to the rest of Asia. Fortunately, China’s slowdown is not very likely to hurt the rest of the world economically, though it could make China more willing to engage in aggressive military misadventures. Hopefully the recent mild thaw in U.S.-China relations means that they’re not heading down the same road as 1914 Germany.

Predawn in America

If China’s economy had a rough 2023, America’s economy had a stellar one. A year ago, I predicted slowing inflation, but I also thought that this would come at the expense of at least moderately slower growth and higher unemployment. So far, it has not. Amazingly, Fed rate hikes don’t seem to have hurt the real economy much, even though they probably helped bring down inflation (via lower inflation expectations). A drop in oil prices, and possibly also the end of the last-vestiges of post-pandemic supply chain snarls, gave the Fed an assist.

As a result, the U.S. economy has been firing on all cylinders, growing even faster than China. Low unemployment, low inflation, and rising real wages have put us in a Goldilocks situation since late 2022 — not even a bust in the tech industry could stop us. Along with rapid tech progress, the strong U.S. economy is reviving the hope of a Roaring 20s decade. And when we zoom out to look at the whole last decade, the long-term trends look good too — middle-class and working-class wealth rising strongly, inequality starting to fall, and the Millennial generation doing much better than most of us had feared.

The one strange thing about this great economy is that most people don’t seem to have recognized it yet. Consumer confidence is still in the dumps, and many analysts are at a loss as to explain it as anything other than “vibes”. But I suspect that as the Goldilocks situation persists, people will eventually realize that the inflation of 2021-22 is no longer with us, and that things are actually looking pretty good — and if the Fed feels safe enough to start cutting interest rates soon, this will happen even more rapidly. We’re not quite at “Morning in America” yet when it comes to American attitudes, but perhaps it’s predawn.

A world at war

War and geopolitical conflict was a big theme of 2022, with the Ukraine War and the escalation of U.S.-China tensions. 2023 was, unfortunately, no different. The Ukraine War has settled into a costly stalemate, and U.S.-China tensions have eased ever so slightly, but this October war exploded between Israel and Gaza. Unlike Ukraine, this is a war where the moral lines are far less clear. There are two sides that each want to control all of the territory “from the river to the sea”, instead of simply establishing three separate countries like they ought to. And both are prepared to use brutal methods to achieve their goals. Putin’s failure in the Ukraine war was a geopolitical coup for America and its alliances; the Israel-Gaza war, no matter how it goes, will be bad for both.

This was not a war the U.S. wanted. But in a world where American economic and military dominance has largely evaporated, and in which the U.S. has been stretched thin by conflicts or potential conflicts around the globe, it was not a war that America had the power to prevent. Expect more conflicts to pop up around the globe, now that there’s no longer a hegemon. The West is not failing, but the end of Pax America is not going to be pleasant.

At this moment of proliferating threats, we need to remember that the overarching competition between the U.S. and the China-Russia axis — which now includes Iran as well — is still the most important geopolitical story by far. Not all of the conflicts around the world, nor the reactions to them, are outgrowths of Cold War 2, but many are. This is a challenge the U.S. and its allies can’t afford to shy away from.

In particular, we need to make sure we’re able to deter a war with China over Taiwan or the South China Sea, even if that means we have to spend more on defense. This requires spending less attention on the Middle East and more on Asia. And it also means our government needs to be stepping up to make a forceful case for liberalism, instead of leaving public debate to a “marketplace of ideas” where the Chinese and Russian governments have the most powerful voices.

Indiamerica has arrived

2023 was, in some sense, the year America discovered India. A flurry of diplomatic initiatives, culminating in Prime Minister Modi’s visit to Washington, made it clear that the two countries are becoming quasi-allies with a shared interest in guarding against Chinese power. And multinational companies’ rush to diversify out of China has resulted in India getting a lot of attention as an investment destination.

But honestly, I think that China is only a modest part of why India and America are becoming more important to each other. A bigger factor is the flood of hundreds of millions of Indians into online spaces, following India’s rapid smartphone adoption in the last couple of years. India has four times the U.S.’ population — it’s now the most populous country on Earth, as of this year — and many of them speak English. It was therefore inevitable that Indians would become very important to English-language online media.

At the same time, India is one of America’s biggest sources of immigration, especially of the high-skilled variety; Indian Americans are making waves in business, science, and politics. The Vice President, and two GOP presidential candidates, are of Indian descent.

The big question for India now is whether it can successfully industrialize. Its growth has been fairly spectacular up until recently, and the de-risking wave gives it a big opportunity. But India needs to rise to that challenge with strong pro-growth policies, especially in education, deregulation, and a continued focus on infrastructure.

The financial crisis that wasn’t

One brief but exciting economic event this year was the collapse of Silicon Valley Bank, which touched off a small wave of mid-sized regional bank failures (mostly on the West Coast). This episode caused a lot of panic, but it also gave us econ writers the opportunity to explain how bank runs work, and how they can be stopped. Fortunately the government stepped in very quickly, doing exactly what it needed to do by temporarily guaranteeing unsecured deposits at SVB and other failing banks, and opening up a lending facility. This ended the banking panic, just as economic theory would predict.

What the SVB drama illustrated, besides the timeless wisdom of the Diamond-Dybvig model, was just how insulated the rest of the economy is from Silicon Valley. A few tech figures predicted that the financial system would collapse, everyone would switch to Bitcoin, etc. — nothing even remotely close to that happened. In fact, the U.S. economy as a whole shrugged off both the tech bank collapse and the more general bust that has been happening in the tech industry since late 2021. This macroeconomic irrelevance might hurt some tech figures’ pride, but it’s actually good for the tech industry, since it means that the U.S. government won’t have a reason to step in and regulate it heavily.

Bidenomics growing pains

2022 saw Biden’s big economic strategy come together. Attempts to reform the welfare state were largely stymied by inflation, but attempts to reshape the economy through industrial policy and strategic trade are beginning to have big effects. These efforts started out mostly as a reaction to challenges like climate change and military competition that the traditional “neoliberal” free-market approach couldn’t handle. But in recent years, some economists have started to think more deeply about why industrial policy might be a new organizing principle for a modern economy.

The new approach is hardly without its vocal critics, even if their criticisms so far tend to be a bit scattershot. And Biden’s efforts have run headlong into one big problem: the permitting regime that has prevented America from building much of anything in the last few decades. Distressingly, some progressives have tried to defend the existing system, but this stubborn status quo bias is only harming their own progressive goals.

What the Biden administration and its successors need to do is to push past these reflexive defenses of the old order. America’s leaders need to embrace FDR’s philosophy of “bold, persistent experimentation”, observe which policies are working and which aren’t, assess the reason for the failures, and make needed adjustments. They also need to beef up the civil service, boost skilled immigration, and do whatever else is necessary to increase the nation’s capacity to make industrial policies work.

Looking ahead to 2024

Obviously, and ominously, 2024 will be dominated by the U.S. election. Concerns about Biden’s age and lingering anger over the inflation of 2021-22 threaten to bring back Donald Trump, which would throw both U.S. society and global geopolitics into complete chaos. I expect this risk to inhibit the U.S.-led bloc’s attempts to help Ukraine, bolster Taiwan, and reshape the global economy in a durable fashion in 2024. That’s unfortunate, but it’s the price of living in a divided society such as ours.

Socially, I think the U.S. is still in a period of slowly cooling unrest, despite all the fighting over Israel-Gaza. But I think 2024 will be a bit like 1976 — a small resurgence of unrest that makes many people feel like the 2010s never ended. But just like the country truly calmed down after ‘76, I foresee the potential for a much more general and durable calming after 2024, especially if Trump loses again.

Economically, there’s the possibility that a long period of 5% interest rates could be more than industries like commercial real estate can handle. That could provoke a recession, or at least a mild slowdown like the one in 2015-16. But there’s a silver lining here — if growth and employment start showing even small signs of weakness, inflation will probably fall even more, and the Fed will probably cut rates faster, both of which could make Americans feel better about the economy.

Technologically, I’m pretty optimistic about 2024; I think it’s pretty clear that we’ll get continued progress in AI, batteries, space, robots, etc., as well as continued adoption of solar power and electric cars. Of course I’ll write about that in my annual “techno-optimism” post.

Anyway, it’s still an exciting, crazy, fascinating world out there. Stay tuned, and enjoy another year of Noahpinion.

---