In an interview with Chris Miller – professor at Fletcher school, he argues that nuclear weapons are easier to make than advanced chips. Nuclear technology hasn’t changed much since the 1960s, but chips are constantly changing and evolving to become smaller and more efficient over time. Chips are ultra small and inexpensive pieces that are used in almost anything electronic. They contain thousands or millions of tiny devices called transistors which are used to flip circuits off and on to create binary code.

This counterintuitive reality challenges our assumptions about technological complexity. “It’s so easy to make nuclear bombs, even the North Koreans can do it,” Miller explains, pointing out that nuclear weapons technology has remained largely static for decades. Meanwhile, the semiconductor industry has achieved something far more remarkable: making computing power so cheap and ubiquitous that it’s embedded in everything from dishwashers to cars.

The Invisible Revolution in Your Pocket

Every time you send a text message or scroll through social media, you’re witnessing the work of billions of microscopic switches operating at speeds that would have been unimaginable just decades ago. These transistors, carved into pieces of silicon no larger than your fingernail, create the endless streams of ones and zeros that power our digital world. Your Instagram likes, your GPS directions, your streaming music – all of it exists as binary code flowing through circuits smaller than viruses.

The scale of this miniaturization is staggering. Today’s most advanced transistors are measured in nanometers, making them only slightly larger than individual atoms. They’re smaller than bacteria, smaller than the mitochondria in our cells, and about half the size of a coronavirus particle. “There’s basically nothing we manufacture at such tiny scale,” Miller notes, emphasizing just how extraordinary this achievement is.

From Vacuum Tubes to Nanoscale Precision

The journey to this level of precision began in the mid-20th century when engineers at Bell Labs invented the first transistor. Before this breakthrough, computers relied on vacuum tubes – bulky, inefficient devices that generated so much heat they literally attracted moths. Early computer operators had to regularly “debug” their machines by removing insects that were drawn to the warm, glowing tubes.

The transition from individual transistors connected by wires to integrated circuits represented a fundamental shift. As Miller describes it, engineers realized they could eliminate the “jungle of connections” by placing multiple transistors on a single piece of semiconductor material. This innovation, developed simultaneously at Texas Instruments and Fairchild Semiconductor, created the first computer chips and set the stage for decades of exponential improvement.

The Economics of Getting Smaller

What drives this relentless pursuit of miniaturization isn’t just technological possibility – it’s economic necessity. Moore’s Law, named after Intel co-founder Gordon Moore, predicts that the number of transistors per chip doubles every couple of years. This isn’t a law of physics but rather a law of economics that has held true since the 1960s.

To illustrate the unprecedented rate of improvement in semiconductors, Miller offers a striking comparison: “If airplanes doubled in speed every two years from the 1960s up to the present, we’d be flying faster literally than the speed of light. But chips have done that.”

This exponential improvement has made computing power essentially free, enabling the explosion of smart devices throughout our daily lives. Your coffee maker, refrigerator, and car all contain multiple chips because the technology has become so inexpensive that manufacturers can afford to add computing power to virtually any product.

The $350 Million Machines

Achieving nanoscale precision requires some of the most sophisticated manufacturing equipment ever created. Inside semiconductor fabrication facilities, or “fabs,” humans are largely absent from the production floor because people are simply too imprecise for this type of manufacturing. Instead, machines costing $350 million each handle the delicate work of carving transistors into silicon.

These machines represent the pinnacle of human engineering achievement. They use the flattest mirrors ever made, the most powerful lasers deployed in commercial devices, in a process dealt with extreme precision. A ball of tin falls through a vacuum chamber and gets struck twice by a laser, creating a plasma that burns 40 times hotter than the surface of the sun. This plasma emits light at exactly 13.5 nanometers – the precise wavelength needed to carve transistors with atomic-level accuracy.

A Global House of Cards

The complexity of modern chip manufacturing means that no single country can produce cutting-edge semiconductors independently. The industry relies on a intricate global supply chain where Taiwan manufactures most advanced chips using tools from the Netherlands and Japan, chemicals from Japan, and designs from the United States.

This specialization has created both incredible efficiency and dangerous vulnerability. Today, just three companies worldwide can produce the most advanced processor chips, and some devices rely on chips that can only be made by a single factory in Taiwan. The Taiwan Semiconductor Manufacturing Company alone produces about 90% of the world’s most advanced processor chips – the ones that power smartphones, computers, and artificial intelligence systems.

The New Battleground

This concentration of manufacturing has turned semiconductors into a key battleground in the competition between global powers. China imports more chips than oil, making it completely dependent on foreign suppliers for the technology that powers its economy. Meanwhile, the United States has begun restricting sales of advanced AI chips to China, hoping to maintain its technological edge in artificial intelligence.

The stakes couldn’t be higher. Any disruption to chip production in Taiwan – whether from natural disaster, accident, or conflict – would create shortages that make the pandemic-era chip crisis look trivial. During COVID-19, missing chips forced automakers to leave thousands of completed cars in parking lots, waiting for semiconductors that cost just a few dollars each. A broader disruption could cripple the global economy.

The AI Acceleration

The rise of artificial intelligence has added new urgency to chip development. Training advanced AI systems requires enormous amounts of computing power, with companies like OpenAI and Anthropic spending millions of dollars primarily on purchasing the most advanced semiconductors available. The release of ChatGPT triggered a wave of investment as tech companies rushed to build massive data centers filled with cutting-edge chips.

This demand is driving even more specialization. While companies like Nvidia have dominated with general-purpose AI chips, a new generation of startups and tech giants are designing processors optimized for specific AI workloads. The goal is to make artificial intelligence cheap enough to be as ubiquitous as Google search – essentially free for users.

The Endless Frontier

Looking ahead, Miller sees no reason why Moore’s Law won’t continue for the foreseeable future. As chips become more powerful and less expensive, they’ll find their way into even more applications. Cars that currently contain around 1,000 chips might have 10,000 within a decade. The basic pattern remains constant: better chips lead to cheaper chips, which leads to more uses for chips.

This endless cycle of improvement represents one of humanity’s most remarkable achievements – turning sand into the foundation of the modern world through precision engineering that operates at the scale of atoms. While nuclear weapons may capture headlines, it’s the humble semiconductor that truly shapes our daily lives, one microscopic transistor at a time.

Interview Link: https://www.youtube.com/watch?v=foYWzdvajvo