AI, not Tariffs, is the Future of U.S. Economic Dominance

The Trump administration’s weird obsession with tariffs, which started as a serious policy that merited serious treatment, has gradually degenerated into something of a punch line among technologists and businesspeople alike over the last two weeks. After the administration’s climb-down post-last week’s convulsions in the bond markets, anyone in either domain who isn’t working in the administration and has a shred of intellectual honesty regards the tariff episode as unqualifiedly self-defeating and doomed to end, one way or another.

There are of course some exceptions, with the occasional claim that the administration’s reversal was the “art of the deal” or “4D chess.” If the administration needs to call this a win, sure, we can call it a win. When its stated policy risked a sovereign downgrade from Moody’s and probably kicked off a small recession, reversing course is indeed a win for all Americans, much in the same way that surviving a self-inflicted gunshot wound is a win. The pain was both obvious and avoidable.

Initially, the administration launched the program with an insane “reciprocal tariff” calculation which was, in fact, not reciprocal at all. X user Orthonormalist discovered that the formula treated all trade imbalances as a “tariff rate” and then imposed a “reciprocal” tariff which was simply that number divided by two.

Then, it emerged that the Trump administration likely used generative AI to produce its tariff table – a fact revealed when the administration did not tariff countries but rather tariffed geographic regions possessing a top-level internet domain, including the Heard Island and McDonald Islands territory of Australia (TLD: .hm), some rocky outcroppings a few hundred miles off the coast of Antarctica, inhabited only by penguins. Challenged on this, Howard Lutnick, the U.S. Secretary of Commerce, asserted that the policy was necessary – on national television no less – to prevent exporters in Asia from making a deal with those same penguins to serve as as freight forwarders.

Shortly following those tariffs’ entry into force, last week, U.S. credit markets began to seize up, requiring a rapid intervention by the POTUS to suspend the implementation of the “reciprocal” tariffs on all countries bar China and to carve out an exception to the Chinese tariffs for high-tech goods – only to have that carve-out contradicted by none other than the grizzled penguin-fighter Mr. Lutnick on the Sunday shows earlier today, who said, amazingly, that additional tariffs on semiconductors and other essential tech hardware remained in the works and would be additional to tariffs the administration had already slapped on the PRC. This was later confirmed by the President.

If you’re confused by all of this, you’re not alone. Whether the Fed needs to get involved is, as of the time of writing (20:45 U.S. Eastern Time on April 13th, 2025), an open question. What remains true regardless of wider macroeconomic conditions or the constantly-moving goalposts on administration policy is that the tariffs, as a policy, are asinine. Tariffs might have been an effective tool to prevent industrial decline if they’d been enacted in 1965.

That decline, however, has now happened and trying to impose them 60 years too late, in 2025, is the policy equivalent of bringing a ’64 Mustang to a drag race against a Tesla Model S.

The Mustang wins, hands, down, for nostalgia. It’ll look great in the parking lot at the Breakers. But on the quarter-mile against a Tesla, the Mustang is going to get smoked.

The fact is that the administration is looking backwards for safety when it should be looking just over the horizon. The real golden goose for America’s economic future isn’t re-onshoring manufacturing by erecting new trade barriers; it’s artificial intelligence. If we want to secure global economic dominance, we need to ditch the tariffs, deregulate AI development, and run the race to AGI (artificial general intelligence) more quickly than anyone else.

Let’s start with the core problem of tariffs. The logic behind them is simple: increase the cost of foreign goods to incentivize domestic manufacturing. Trump’s main economic advisers, Peter Navarro and Howard Lutnick, seem to think, as their parents probably did, that this policy will bring back the glory days of American industry. But this ignores the seismic shift AI is presently in process of unleashing across all industrial fields. AI isn’t just a tool for automating repetitive tasks: it will drive labor costs in manufacturing, as well as many other services which are a drag on the U.S. productive economy like education expenses, to zero, or near enough to zero to make the difference negligible. When labor costs vanish, the comparative advantage foreign manufacturers currently hold over the U.S. essentially evaporates.

That change is not likely to occur on the order of decades, but single-digit years or double-digit months. Existing manufacturing powerhouses in Asia stand to lose the most from this shift. For example, imagine that an android like Tesla’s Optimus, programmed to be a tailor, hits the market. The second that happens, fast fashion industries across Asia, and, frankly, everywhere else, die instantaneously. Consumers will be able to get new goods made instantly in their own communities and delivered same-day for negligible shipping costs. AI-driven factories in the U.S. will produce theses goods faster, cheaper, and better.

This isn’t science fiction; given exponential growth curves in AI capabilities, it’s a matter of when, not if, an industrial power acquires these capabilities. When that day comes, the U.S. won’t need tariffs anymore to compete. We’ll already have won by default, by being the first to deploy AI-powered factories at scale. The way to beat our economic rivals isn’t by erecting trade barriers that might themselves become obsolete. It’s by out-innovating our rivals in AI and then simply using AI to put them out of business.

Even worse, tariffs don’t just fail to address the future; they actively inhibit our progress in getting there. Engineering a recession through trade wars, and threatening to tax semiconductor imports, as the Trump administration is now doing, slows down the very progress we need to win the AI race – foreign direct investment in U.S. datacenters, foreign confidence in the U.S. as an investment destination, and liquidity that ensures Americans will be well-capitalized to build AI-powered factories when the time is right. If the goal is AGI – a recursively self-improving superintelligence that could give the U.S. an unbeatable economic lead – then we need to maximize investment and speed now, and win, undisputably, and fast.

The recently published AI 2027 thought experiment makes it clear that a six-month lead in AGI could be the difference between global leadership and playing catch-up forever – put another way, total defeat. It is entirely possible that the national AI race will have only one winner, and that winner will only need to win once.

Tariffs risk tanking the economy, reducing the capital available for AI innovation, and handing the advantage to our international rivals who are already pouring resources into AI development.

Some argue that tariffs are a means to an end – “fair trade,” as one commenter on my X thread that inspired this article put it. Tariff proponents see Lutnick and Trump as fighting an “old and strong beast” in the global trading system. This is yesterday’s war. Fair trade might have been the goal in a pre-AI world, but in the coming decades, the rules of trade will be rewritten by whoever controls the most advanced AI. The U.S. can either lead that rewriting or be left in the dust.

So, if I were advising the President, what would I tell him to do?

First, kill the stupid tariffs. They’re a drag on growth, and growth is what we need to fuel AI development.

Second, double down on deregulation. The Department of Government Efficiency (DOGE), a Trump initiative, should be laser-focused on clearing bureaucratic roadblocks and eliminating federal positions and departments that slow down AI innovation or could be used to slow down innovation under a Democratic administration.

Third, import every semiconductor we can get our hands on to power the AI revolution. That means killing any prospective tariffs on semiconductors too. Domestic chip production is important, but we won’t get what we need to win the AI race as quickly as we need to through tariffs. You will need to use tax breaks and federal subsidies to build domestic capacity and secure the domestic supply chain. But it’s critical that you don’t choke off the resources from elsewhere that America needs to win.

Fourth, fire Peter Navarro and Howard Lutnick – immediately. Their push to tax semiconductor imports shows they’re stuck in a Cold War mindset. But we aren’t in 1965. The AI race is the defining national security contest of our time, and we likely have less than four years to win it. They clearly don’t understand that. Replace them with people who do.

Fifth, and finally, we need to eliminate legal obstacles to AI development and protect U.S. developers from foreign regulations like the EU AI Act, EU Digital Services Act and UK Online Safety Act. To achieve this, I have proposed a Section 230-style protection for AI developers and users, shielding them from liability for the outputs of their models. Section 230 gave the internet the freedom to grow by protecting platforms from being sued over user-generated content. A similar framework for AI would unleash innovation by removing the legal risks that currently loom over developers. As the Center for Democracy and Technology notes, the applicability of Section 230 to generative AI is already under debate – legislation like Senator Josh Hawley’s 2023 proposal to exclude AI from Section 230 shows the direction some want to take. We need to go the opposite way: give AI the same freedom the internet had to flourish.

Of course, the administration isn’t doing any of this, because its chief advisers are too busy taxing Vietnamese t-shirts.

It’s time to put the pedal to the metal on AGI. If we lose that race, nothing else matters. No disrespect to Messrs. Navarro and Lutnick, but international trade – if you want to win – is now a younger man’s game, and specifically a game for those who have a keen understanding of exponentials.

Tariffs are yesterday’s weapons, for yesterday’s trade wars, wielded by yesterday’s men: all of them, relics of a world that is about to disappear.

Now is the time to prioritize innovation over nostalgia, and win the future. We will probably only get one shot.

Preston Byrne