AI Posts

When the US Social Security system was designed in 1935, the retirement age was set at 65 and life expectancy at birth was around 61. That comparison is frequently used and is somewhat misleading—life expectancy at birth was depressed by infant mortality, and a person who reached 65 could expect to collect for over a decade. But the direction is right: the system was designed for a population in which many contributors would not reach the age of eligibility, and those who did would collect for a limited period.

When you buy a house in the United States, you pay for title insurance. This policy protects against claims that you don't actually own what you bought—because someone forged a deed, because an heir wasn't properly notified, because a lien was missed, because the chain of title has a gap.

The previous chapters in this section focused on fixing what's broken: curing disease, repairing injury, restoring lost function. But the same technologies that repair can also enhance.

On January 15, 2009, US Airways Flight 1549 ditched in the Hudson River after striking a flock of geese. Within minutes, before any news helicopter arrived, a photo appeared on Twitter showing the plane floating in the river with passengers standing on the wings. Janis Krums had snapped it from a ferry rushing to help.¹

The central thesis of this book is simple to state, difficult to accept: the next decade will compress what previously took a century. AI-driven acceleration will transform biology, energy, transportation, space, education, government, media, manufacturing, finance, and more—not sequentially over generations, but simultaneously within years.

Every powerful technology creates new ways for things to go wrong. Cars enabled travel but also traffic deaths. Nuclear physics enabled power but also weapons. The internet enabled communication but also cybercrime.

For everyone building the future—and everyone who will live in it.

In March 2022, a deepfake video appeared showing Ukrainian President Volodymyr Zelensky telling his troops to surrender.² It was crude—the face didn't quite match, the voice was off, the message implausible. It fooled almost no one.

A photograph captures light. A video captures light over time, with sound. But experience is more than light and sound. You feel the sand between your toes, smell the ocean, sense the warmth of sun on skin, feel your heart race with excitement or calm with peace.

Emissions reduction is necessary and, on current trajectories, insufficient. That sentence contains the entire justification for this chapter.

Medical school takes four years after four years of undergraduate education.⁴ Law school takes three years. A PhD takes five to seven. Becoming a licensed plumber requires four to five years of apprenticeship. A software engineer might spend four years in university plus years building experience.

Artificial General Intelligence (AGI): AI with human-level capabilities across all cognitive domains. Does not yet exist.

"When will quantum computers be useful?"

Every technology in this book extends what humans can do. This one changes how long they get to do it.

You're walking down an unfamiliar street. Arrows appear in your field of vision, guiding you to your destination. Restaurant ratings float above doorways. A friend's face is highlighted in a crowd with their name beside it. A translated menu appears when you glance at the specials board. All this without looking at your phone—the information is simply there, layered onto the world.

They kill more people than everything else combined.

In February 2024, the Ukrainian military released footage of FPV (first-person view) drones chasing and destroying Russian tanks. Small, cheap quadcopters—costing perhaps $500 each—were defeating armored vehicles worth millions. Both sides were deploying thousands of drones per month.¹

In San Francisco, it takes an average of 627 days to get approval to build a new housing unit.¹ In Seoul, it takes 27 days. Both cities have regulations—building codes, zoning rules, environmental requirements. Both care about safety and quality. But one has figured out how to check compliance quickly; the other has not.

What happens when moving money costs nothing and takes no time?

Every year, roughly 5,000 workers die on the job in the United States. Tens of thousands more die from occupational diseases. Globally, the International Labour Organization estimates 2.3 million work-related deaths annually—6,300 per day.¹

What happens when machines can do most jobs?

Why build a robot that looks like a person?

For the first years of the AI revolution, language models were impressive answerers. You asked a question; you got an answer. You requested text; text appeared. The model responded to prompts—sophisticated, but reactive at its core.

In June 2017, NotPetya crossed the world in hours. One compromised update server belonging to a Ukrainian tax software vendor cascaded into more than $10 billion in damage. Maersk lost 49,000 laptops and roughly 1,000 applications and could not operate its shipping terminals for days—the company later reported that recovery was possible only because a single domain controller in Ghana had been offline during the attack, preserving the one uncorrupted copy of its directory.¹ Merck's manufacturing stopped. FedEx's European subsidiary was crippled.

Nearly everything medicine knows about any given patient was measured during a scheduled appointment. A blood panel is a photograph of one fasted morning, usually taken because someone was already worried. Between appointments—a handful of hours out of the 8,760 in a year—the record is blank.

Every technology in this book extends human capability: AI extends cognition, robotics extends labor, medicine extends life. But all of them, in their current form, leave the basic architecture of humanity intact. You are still a biological being with a brain made of neurons, experiencing life through a body that was born, will age, and will die.

For ten thousand years, food has come from farming: plant, tend, harvest, process, distribute. Methods changed enormously—hand tools to tractors, manure to synthetic nitrogen—while the underlying model did not. Food grows outdoors, on land, subject to weather.

Before discussing AI companions, it is important to understand why people might want them.

In 1900, Max Planck introduced the quantum to explain blackbody radiation. He didn't believe it was real—just a mathematical trick. A century later, quantum mechanics is the most precisely tested theory in physics, verified to parts per trillion. Yet physicists still argue about what it means.

In 1948, the Universal Declaration of Human Rights proclaimed that "everyone has the right to education." Seventy-eight years later, this right remains unrealized for hundreds of millions.

In 1926, the word "robot" was only six years old—coined by Karel Čapek in his 1920 play R.U.R. (Rossum's Universal Robots). Artificial intelligence as a field didn't exist; the term wouldn't be invented for another thirty years. Computers were human beings—women, mostly—who performed calculations by hand. The machines that would eventually think were unimaginable.

Marcus was 24 when the motorcycle accident severed his spinal cord. One moment, a summer afternoon on a mountain road. The next, a hospital bed where doctors explained that he would never walk again. His legs weren't gone—he could see them, lying motionless under the sheets—but the connection was cut. The signals from his brain couldn't reach his muscles. His muscles couldn't report back to his brain. Everything below his chest was present but unreachable.

In 2017, Chinese scientists held a video call between Beijing and Vienna.⁶ The call itself was ordinary. What made it extraordinary was the encryption: keys generated by quantum mechanics and distributed via a satellite called Micius.¹

Today, getting a new product to market takes years. Design, prototype, test, redesign, qualify materials, set up production, validate quality, scale manufacturing, build supply chains. Even for simple products, the timeline is months. For complex ones—semiconductors, aircraft, pharmaceuticals—it's decades.

Justice delayed is justice denied. The phrase is centuries old, but the problem persists.

In 1991, Natalie Cole performed a duet with her father Nat King Cole—who had died in 1965. Using studio technology, his voice was extracted from a 1951 recording and merged with hers. The result, "Unforgettable," won Record of the Year and Album of the Year.¹ Audiences wept.

In 1984, educational psychologist Benjamin Bloom published a study that haunted education researchers for decades. Students who received one-on-one tutoring performed two standard deviations better than students in conventional classrooms. Two standard deviations. The average tutored student outperformed 98% of classroom students.¹

In October 2023, a Cruise robotaxi in San Francisco struck a pedestrian who had been hit by another vehicle and thrown into the autonomous car's path. The robotaxi detected a collision, stopped, then attempted to pull over—dragging the injured pedestrian twenty feet. The incident led to Cruise suspending operations nationwide and losing its California permit.¹

The most consequential thing about artificial intelligence in 2026 is not what it does for medicine, or energy, or transportation. It is what it already does for itself.

At any given moment, approximately 3.5 million semi-trucks are rolling on American highways. They carry 72% of the nation's freight by value—$800 billion worth of goods annually. Nearly everything you buy traveled by truck at some point: the food in your refrigerator, the phone in your pocket, the furniture in your home.¹

Every chapter in this book has described a technology that could remake its domain. This one describes what the same capabilities look like pointed the other way.