What I think happens next.

The degree was always a proxy for capability. AI made the proxy obsolete. When the work a graduate was hired to do gets done by a model, the four-year six-figure credential stops penciling out. The demographic cliff was already coming. AI turns a slow decline into a collapse.

Owning a car is about to look like owning a horse. Autonomous fleets make the license pointless for a generation that can summon a ride cheaper than insuring a used Civic. The status symbol flips. Not driving becomes the default, not the exception.

Tesla is worth about $1.6 trillion today. The car business doesn't get you to $10 trillion. Robotaxi and Optimus do. If autonomy and humanoids land anywhere near the timeline, the cars become the smallest part of the story. That's a 6x in ten years. Aggressive, not crazy.

He lost his house in the fires and turned it into the most-watched comeback in the city. Name recognition, sympathy, and a media machine beat a résumé in LA politics every time. Stranger people have won bigger offices.

It's around $65,000 today, well off its peak. A million is roughly 15x. The case isn't magic. It's a supply that can't expand meeting demand from institutions and governments that keep showing up. One country puts it on the balance sheet and the rest scramble to follow.

It's around 7,600 today, already at record highs. The bull case isn't complicated. AI drives a productivity step-change, earnings expand, and the multiple holds. That's a double in under five years, roughly the pace of the 1982 to 2000 run. People will call this aggressive. The last secular bull says it's the base case, not the ceiling.

Personal car ownership is a habit, not a law. Once a robotaxi ride is cheaper and safer than the car sitting in your garage, the math ends the debate. The shift won't feel gradual. It'll feel like everyone switched at once, the way they did with smartphones.

Uber is worth about $146 billion today, already down from its highs as the market wakes up to the threat. The whole model depends on human drivers Uber doesn't have to pay to sit idle. When the cars drive themselves, whoever owns the fleet owns the margin, and it isn't the app that just dispatches them. Uber either buys its way into hardware or gets disintermediated by it.

Waymo got there first and built the better system. It won't matter. Tesla makes the cars, owns the factory, and ships millions of vehicles a year that can be turned into the fleet. Nobody else controls the whole stack. First to scale beats first to launch.

There were a few thousand a year ago. That sounds like a reason to doubt the number. It's the opposite. This is what the early part of an exponential looks like right before it stops being slow. Labor shortages, plus falling robot costs, plus AI that finally works, equals a wall of demand. A million is the floor, not the ceiling.

The first ones cost six figures. The ten millionth won't. Same curve as every piece of hardware that ever mattered, just faster, because the robot's brain rides on chips that are already on that curve. When a machine that works 20 hours a day costs less than a used Camry, the question stops being whether you buy one and becomes how many.

Filing taxes is a solved problem pretending to be hard. Your income, your deductions, and your forms already sit in a handful of databases. An AI that reads them, asks you three questions, and files in ten minutes makes the $300 preparer and the lost Saturday with tax software look insane. H&R Block and the rest are charging for a problem AI just deleted.

Tuition has gone up every year for forty years. That streak ends. When AI can teach a course better than the median professor, and the degree itself is worth less in the job market, schools lose their pricing power all at once. This is the first nominal price drop in the history of American higher ed. Not adjusted for inflation. Actually cheaper.

Most of what a primary care doctor does is pattern matching on symptoms, history, and labs. AI is already good at the pattern part, and it never forgets, never rushes, and costs almost nothing to run. You'll see a human when something is actually wrong. For the sore throat, the refill, and the "is this normal," you'll talk to a model.

US electricity demand barely moved for twenty years. That era is over. Data centers, AI training, and millions of robots that all need charging turn power into the scarcest input in the economy. The bottleneck on the whole boom stops being chips or capital. It becomes whether the grid can keep up.