A bonsai installation at the Omiya Bonsai Museum. Photo by Danny Crichton.
The inflated expectations of strategic dilemmas
Why is strategy hard? No matter the scope (from individuals to nation-states), strategies are difficult to devise because they force tradeoffs in situations with imperfect information. Resources are limited, goals are ambitious and infinite, and there are only so many ways to align the former with the latter. Add in the complexity of predicting an uncertain future, and there is almost never a “solution” that perfectly fits a problem.
Most strategies are far worse than that though, and in fact, most end up simply being “wrong” — even before we reach an outcome. In business, resources (particularly people) are over-allocated or expected to perform at infeasible levels. Sales targets are marked up in the beginning of the year only to be brought down to Earth in later quarters as reality intrudes on fantasy. At higher scopes like nations, strategies aren’t just wrong but are also simply incoherent, with different parts of the same strategy working in diametrically opposed ways.
I was thinking about this basic tension in the context of the wall-to-wall vituperative coverage of China’s spy balloon the past two weeks, and the increasing fervor on Capitol Hill to really strike against China on as many fronts as possible. The House of Representatives voted 419-0 on a resolution condemning the balloon’s flight over the U.S., and now committee leaders are promulgating a series of ever-harsher economic and security measures to kneecap the Middle Kingdom and its overseas activities.
(Senator Rick Scott alone has discussed bills like “The Taiwan Invasion Prevention Act”, “The Deterring Communist Chinese Aggression Against Taiwan through Financial Sanctions Act” and “The No CCP (Chinese Communist Police) in the United States Act of 2023” to get a taste of where the gaze of Congress is these days).
Swimming beneath that current of vitriol last week was the release of the latest trade data published by the Census Bureau, which showed that the U.S. trade deficit in 2022 reached an all-time high of almost $1 trillion. Annual trade with China hit a record deficit of $382.9 billion, and the U.S. also recorded massive deficits with the European Union ($203.9 billion) and Mexico ($130.6 billion). Two factors widening those deficits were large increases in imports of cars and auto parts (up $52 billion) as well as cell phones (up $11 billion).
These are obviously “net” numbers, and it’s instructive to zoom into a relationship like China to get more granular. American exports to China were $153.8 billion while imports were $536.8 billion. Where do both countries hold a lead? According to the Bureau of Industry and Security (and using 2021 data), America’s strongest sectors for exports were in agriculture products, oils, minerals, chemicals and plastics. Meanwhile, China was strongest in machinery and appliances although this is a two-way street — America also exports tens of billions of goods in this category.
Here we have our first strategic dilemma: the country obnoxiously flying a spy balloon over America is also one that buys more than $150 billion worth of our goods, and also sells us more than $500 billion worth of useful and competitive products. Scaling back trade ties would most significantly harm U.S. farmers, who rely on the appetites of China’s 1.4 billion people for their profits.
Let’s add another dilemma. The Congressional Budget Office on Wednesday published its budget projections for this year and the upcoming decade to 2033. The figures, as one can imagine, are bleak. This year’s federal budget deficit is expected to be $1.4 trillion, and by 2033, the office projects that the deficit will reach $2.7 trillion. To put that figure into context, federal debt to GDP was 35.2% in 2007, reached 97% this year, and is projected to swell to 118.2% by 2033.
Even as we drain our coffers, debt is getting more expensive to service as interest rates rise at the Fed and other central banks. Federal debt service is now around $475 billion for 2022 (up 35% from 2021), and while most of the government’s debt is held in instruments with historically lower coupon rates, the average interest rate is projected to spike in the years ahead as the government refinances older debts and takes on new debt.
So we have another strategic dilemma: the government is profligately spending money it doesn’t have, and is increasingly under strain as a greater percentage of general revenues are used to cover the rising costs of debt service. There is no strategy or “blueprint” to make the math here sustainable, given the sheer scale of cuts and revenue increases required to bring the budget even close to being balanced.
Now, let’s turn to a third and final dilemma: Taiwan. Taiwan is the flashpoint of U.S.-China relations, and the tensions around the island have increased rapidly in the last year. Another Congressional delegation to the island is expected sometime this year, and the Financial Times is reporting as I write this newsletter Friday that Michael Chase, the Pentagon’s top China official, has landed in Taiwan. As the FT notes, “Chase is the first senior US defence official to travel to the island since the 2019 visit of Heino Klinck, deputy assistant secretary for east Asia, who in turn was the most senior Pentagon official to visit Taiwan in four decades.”
The U.S. has minimally “pivoted” to Asia over the past decade, and defending Taiwan would require prodigious new funding in terms of military equipment, new and redeployed personnel, and base construction. How prodigious? DC analysts, usually so loquacious to gab on defense budgets, are mum to offer a number. I’ll throw out one though: a trillion dollars, or roughly one extra year of the U.S. defense budget as supplemental spending.
Clearly, protecting democratic Taiwan is a high priority, particularly after what the world witnessed in Hong Kong the past few years. But the costs for that protection are extremely steep — and the U.S. is caught in the vice grip of a clenching budget as well as in the deep economic web of needing Chinese imported goods and its valuable export markets. That’s our third dilemma.
A proper national strategy would reconcile the economic vitality that comes from U.S.-China relations (and the concomitant tax revenue) with the need for bringing federal government spending into alignment with fiscal reality, while also making commitments and outlays to protect Taiwan. That’s a tall order, but one that’s doable with intentional tradeoffs.
Unsurprisingly, there is is no such strategy or even the hopes of such a strategy. And to be cynical, there almost can’t be in America’s current political environment. If politics is the art of the possible, then American politics is the art of the impossible, of arguing that all the things can be done and somehow all the fiscal math will reconcile itself out.
It was popular during the Iraq/Afghanistan years to make reference to historian Paul Kennedy’s use of the term “imperial overstretch” in his magnum opus The Rise and Fall of the Great Powers. The idea is that empires (or nations) tend to fall when they start getting far too ambitious with their own designs on the world compared to their economic and fiscal bases. Instead of fighting one war, the empire fights four — simultaneously — even as climate change cuts crop production and spreads widespread famine.
References to the theory never made much sense during the counterterrorism era, since America could ultimately afford to spend tens of billions a year on supplementary defense appropriations per year (dubbed Overseas Contingency Operations) without causing a deep fiscal crisis.
The strategic question for the decade ahead is whether a highly-indebted country with a generation of retirees hitting their pension years can afford the far larger outlays required to prevent China from gobbling up Taiwan (and perhaps other territory in East and Southeast Asia). Would we cut entitlement programs to deepen spending on global defense? Would we cut defense investments in some regions of the world (namely, the Middle East) to migrate those resources to the theater where we need it? Can we give up doing everything, everywhere all at once to do something, somewhere in a little while?
That’s not an easy conversation for America to have, but it is the difficult conversation the country needs. Tens of billions of dollars in subsidies for semiconductors, solar power and climate mitigation is a good investment, but does the value hold up given all the implicit tradeoffs those subsidies require?
America’s spending has ballooned along with its ambitions (or at least its largess). But that balloon is stretching to its limit, and not unlike a certain Chinese spy balloon, feels ready to pop — or worse — to be pierced by an adversary.
Lux Recommends
Our scientist-in-residence Sam Arbesman recommends sci-fi writer Ted Chiang’s take on ChatGPT in the New Yorker dubbed “ChatGPT Is a Blurry JPEG of the Web”. Chiang, whose fame skyrocketed with the publication of his short-story collection Exhalation, writes “This analogy to lossy compression is not just a way to understand ChatGPT’s facility at repackaging information found on the Web by using different words. It’s also a way to understand the ‘hallucinations,’ or nonsensical answers to factual questions, to which large language models such as ChatGPT are all too prone.”
Sam previously recommended Gabrielle Zevin’s novel Tomorrow and Tomorrow and Tomorrow, but having started it myself, I must say it is wonderfully crafted and I can understand why so many people have recommended it to me. Poignant, nerdy and fun.
Sam recommends a Caltech research group’s exploration of Leonardo Da Vinci and his understanding of gravity. He was centuries ahead of his scientific descendants, although didn’t quite nail gravity’s mathematics (what a loser!).
Yevgeny Prigozhin, the founder of Russian mercenary firm Wagner Group that has been heavily deployed in Ukraine, has grown in profile with, well, a panoply of profiles by publications the world over. I enjoyed this particular profile by Shaun Walker and Pjotr Sauer in The Guardian on “The hotdog seller who rose to the top of Putin’s war machine”.
Sam recommends Tom Scott’s video essay on ChatGPT and generative AI titled “I tried using AI. It scared me.” Scott, a popular YouTube creator (one of my favorites of his is Why the US Army electrifies this water), says, “…that feeling of dread came from the idea that ChatGPT and the new AI art systems might be to my world what Napster was to the late Nineties.”
The qualities that make a good thinker are not the same as make a good citizen. Civility and politeness are all very well in society at large. Good manners are a lovely attribute, but in intellectual life the instinct to be nice and to be liked can be fatal. Rigorous thinking must be followed wherever it leads. The unpleasantness of controversy is inevitable. Everybody cannot be right. This is yet another instance where the ethos that liberal democracy encourages — civility, deliberation, compromise — is in radical contradiction with the intransigence proper to scientific and humanistic discovery. In politics, in ordinary life, compromise is good, even necessary. In reasoning and research, by contrast, you must be a fighter, tenacious, persistent, stubborn in your insistence that someone’s argument is wrong and yours is closer to the truth that you are both pursuing.
That’s it, folks. Have questions, comments, or ideas? This newsletter is sent from my email, so you can just click reply.
Forcing China’s AI researchers to strive for chip efficiency will ultimately shave America’s lead
In incididunt ad qui nostrud sint ullamco. Irure sint deserunt Lorem id officia dolore non. Anim dolor minim sit dolor et sint aliquip qui est. Ex in tempor laborum laboris dolor laboris ullamco quis. Enim est cupidatat consequat est culpa consequat. Fugiat officia in ea ea laborum sunt Lorem. Anim laborum labore duis ipsum mollit nisi do exercitation. Magna in pariatur anim aute.
In incididunt ad qui nostrud sint ullamco. Irure sint deserunt Lorem id officia dolore non. Anim dolor minim sit dolor et sint aliquip qui est. Ex in tempor laborum laboris dolor laboris ullamco quis. Enim est cupidatat consequat est culpa consequat. Fugiat officia in ea ea laborum sunt Lorem. Anim laborum labore duis ipsum mollit nisi do exercitation. Magna in pariatur anim aute.
Right now, pathbreaking AI foundation models follow an inverse Moore’s law (sometimes quipped “Eroom’s Law”). Each new generation is becoming more and more expensive to train as researchers exponentially increase the number of parameters used and overall model complexity. Sam Altman of OpenAI said that the cost of training GPT-4 was over $100 million, and some AI computational specialists believe that the first $1 billion model is currently or will shortly be developed.
As semiconductor chips rise in complexity, costs come down because transistors are packed more densely on silicon, cutting the cost per transistor during fabrication as well as lowering operational costs for energy and heat dissipation. That miracle of performance is the inverse with AI today. To increase the complexity (and therefore hopefully quality) of an AI model, researchers have attempted to pack in more and more parameters, each one of which demands more computation both for training and for usage. A 1 million parameter model can be trained for a few bucks and run on a $15 Raspberry Pi Zero 2 W, but Google’s PaLM with 540 billion parameters requires full-scale data centers to operate and is estimated to have cost millions of dollars to train.
Admittedly, simply having more parameters isn’t a magic recipe for better AI end performance. One recalls Steve Jobs’s marketing of the so-called “Megahertz Myth” to attempt to persuade the public that headline megahertz numbers weren't the right way to judge the performance of a personal computer. Performance in most fields is a complicated problem to judge, and just adding more inputs doesn't necessarily translate into a better output.
And indeed, there is an efficiency curve underway in AI outside of the leading-edge foundation models from OpenAI and Google. Researchers over the past two years have discovered better training techniques (as well as recipes to bundle these techniques together), developed best practices for spending on reinforcement learning from human feedback (RLHF), and curated better training data to improve model quality even while shaving parameter counts. Far from surpassing $1 billion, training new models that are equally performant might well cost only tens or hundreds of thousands of dollars.
This AI performance envelope between dollars invested and quality of model trained is a huge area of debate for the trajectory of the field (and was the most important theme to emanate from our AI Summit). And it’s absolutely vital to understand, since where the efficiency story ends up will determine the sustained market structure of the AI industry.
If foundation models cost billions of dollars to train, all the value and leverage of AI will accrue and centralize to the big tech companies like Microsoft (through OpenAI), Google and others who have the means and teams to lavish. But if the performance envelope reaches a significantly better dollar-to-quality ratio in the future, that means the whole field opens up to startups and novel experiments, while the leverage of the big tech companies would be much reduced.
The U.S. right now is parallelizing both approaches toward AI. Big tech is hurling billions of dollars on the field, while startups are exploring and developing more efficient models given their relatively meagre resources and limited access to Nvidia’s flagship chip, the H100. Talent — on balance — is heading as it typically does to big tech. Why work on efficiency when a big tech behemoth has money to burn on theoretical ideas emanating from university AI labs?
Without access to the highest-performance chips, China is limited in the work it can do on the cutting-edge frontiers of AI development. Without more chips (and in the future, the next generations of GPUs), it won’t have the competitive compute power to push the AI field to its limits like American companies. That leaves China with the only other path available, which is to follow the parallel course for improving AI through efficiency.
For those looking to prevent the decline of American economic power, this is an alarming development. Model efficiency is what will ultimately allow foundation models to be preloaded onto our devices and open up the consumer market to cheap and rapid AI interactions. Whoever builds an advantage in model efficiency will open up a range of applications that remain impractical or too expensive for the most complex AI models.
Given U.S. export controls, China is now (by assumption, and yes, it’s a big assumption) putting its entire weight behind building the AI models it can, which are focused on efficiency. Which means that its resources are arrayed for building the platforms to capture end-user applications — the exact opposite goal of American policymakers. It’s a classic result: restricting access to technology forces engineers to be more creative in building their products, the exact intensified creativity that typically leads to the next great startup or scientific breakthrough.
If America was serious about slowing the growth of China’s still-nascent semiconductor market, it really should have taken a page from the Chinese industrial policy handbook and just dumped chips on the market, just as China has done for years from solar panel manufacturing to electronics. Cheaper chips, faster chips, chips so competitive that no domestic manufacturer — even under Beijing direction — could have effectively competed. Instead we are attempting to decouple from the second largest chips market in the world, turning a competitive field where America is the clear leader into a bountiful green field of opportunity for domestic national champions to usurp market share and profits.
There were of course other goals outside of economic growth for restricting China’s access to chips. America is deeply concerned about the country’s AI integration into its military, and it wants to slow the evolution of its autonomous weaponry and intelligence gathering. Export controls do that, but they are likely to come at an extremely exorbitant long-term cost: the loss of leadership in the most important technological development so far this decade. It’s not a trade off I would have built trade policy on.
The life and death of air conditioning
Photo by Ashkan Forouzani on Unsplash
Across six years of working at TechCrunch, no article triggered an avalanche of readership or inbox vitriol quite like Air conditioning is one of the greatest inventions of the 20th Century. It’s also killing the 21st. It was an interview with Eric Dean Wilson, the author of After Cooling, about the complex feedback loops between global climate disruption and the increasing need for air conditioning to sustain life on Earth. The article was read by millions and millions of people, and hundreds of people wrote in with hot air about the importance of their cold air.
Demand for air conditioners is surging in markets where both incomes and temperatures are rising, populous places like India, China, Indonesia and the Philippines. By one estimate, the world will add 1 billion ACs before the end of the decade. The market is projected to before 2040. That’s good for measures of public health and economic productivity; it’s unquestionably bad for the climate, and a global agreement to phase out the most harmful coolants could keep the appliances out of reach of many of the people who need them most.
This is a classic feedback loop, where the increasing temperatures of the planet, particularly in South Asia, lead to increased demand for climate resilience tools like air conditioning and climate-adapted housing, leading to further climate change ad infinitum.
Josh Wolfe speaking at Nvidia Auditorium at Stanford. Photo by Chris Gates.
Josh Wolfe gave a talk at Stanford this week as part of the school’s long-running Entrepreneurial Thought Leaders series, talking all things Lux, defense tech and scientific innovation. The .
Lux Recommends
As Henry Kissinger turns 100, Grace Isford recommends “Henry Kissinger explains how to avoid world war three.” “In his view, the fate of humanity depends on whether America and China can get along. He believes the rapid progress of AI, in particular, leaves them only five-to-ten years to find a way.”
Our scientist-in-residence Sam Arbesman recommends Blindsight by Peter Watts, a first contact, hard science fiction novel that made quite a splash when it was published back in 2006.
Mohammed bin Rashid Al Maktoum, and just how far he has been willing to go to keep his daughter tranquilized and imprisoned. “When the yacht was located, off the Goa coast, Sheikh Mohammed spoke with the Indian Prime Minister, Narendra Modi, and agreed to extradite a Dubai-based arms dealer in exchange for his daughter’s capture. The Indian government deployed boats, helicopters, and a team of armed commandos to storm Nostromo and carry Latifa away.”
Sam recommends Ada Palmer’s article for Microsoft’s AI Anthology, “We are an information revolution species.” “If we pour a precious new elixir into a leaky cup and it leaks, we need to fix the cup, not fear the elixir.”
I love complex international security stories, and few areas are as complex or wild as the international trade in exotic animals. Tad Friend, who generally covers Silicon Valley for The New Yorker, has a great story about an NGO focused on infiltrating and exposing the networks that allow the trade to continue in “Earth League International Hunts the Hunters.” "At times, rhino horn has been worth more than gold—so South African rhinos are often killed with Czech-made rifles sold by Portuguese arms dealers to poachers from Mozambique, who send the horns by courier to Qatar or Vietnam, or have them bundled with elephant ivory in Maputo or Mombasa or Lagos or Luanda and delivered to China via Malaysia or Hong Kong.”
