Crowns, Wars, and AI Fortunes
Posted October 27, 2026
Chris Campbell
In 1776, a Boston schoolteacher sat in a British prison, eating scraps from the officers’ table.
His crime? He’d been spying on the British army.
Badly, it turned out.
He wrote everything down and sent it to Dr. Joseph Warren, the guy who sent Paul Revere on his midnight ride. Then Warren got killed at Bunker Hill, and the British emptied his pockets.
Rule one of espionage: don’t put it in writing.
Rule two: if you must, put it in code.
Rule three: don’t sign it.
Lovell broke all three.
The British threw him in a cell for 18 months.
Five years later, as fate would have it, he sat in Congress with a pile of stolen British letters.
One problem: they were gibberish. Rows of complete nonsense. The original cryptography.
The Original AI
The man’s name was James Lovell. Maybe you've heard of him. I hadn’t until recently.
Apparently, nobody trained Lovell to break codes. He taught himself. (A man who spends eighteen months in jail over a letter tends to take an interest.)
When British mail came in, it landed on his desk. Scrambled. Unreadable. He started with what he knew about English.
He knew "e" shows up more than any other letter. He knew "the" appears everywhere. He knew what a British officer would probably say next.
Guess a letter. Check it. Guess the next one.
The gibberish cracked. By the fall of 1781, Washington was outside Yorktown, reading Cornwallis’s letters. Lovell had sent him the key.
Here's why I’m telling you this: Lovell basically invented ChatGPT. Stay with me. Because this weird fact answers the question everyone’s asking: is AI a bubble?
From Quill to Chatbot
Jump ahead 160 years.
Alan Turing breaks the Nazi Enigma code. The machines built for that job lead to the first computers.
Then Claude Shannon, a Bell Labs mathematician who worked on wartime codes, runs a strange experiment.
In 1951, he handed people a piece of English text and asked them to guess the next letter. Then the next. Then the next. He wants to measure how predictable language is.
Answer: very.
Lovell's method. Now with math.
Fast-forward 70 more years. Every AI you use runs on that same idea. ChatGPT. Claude. Gemini. Predict the next word. Then the next.
That's the whole trick.
Five Codes, One Key
With ChatGPT, everyone expected a fancy autocomplete. Then the predictions got good enough to write software. Draft contracts. Pass the bar exam.
Here's the thing about codes…
They’re everywhere.
- Software is a code. AI now writes a big chunk of it. Ask any programmer.
- DNA is a code. Google DeepMind's AlphaFold cracked how proteins fold. Its creators won the 2024 Nobel Prize in Chemistry.
- Law is a code. Contracts repeat patterns. So do court rulings.
- Medicine is a code. Symptoms. Scans. Lab numbers.
- Markets are a code. Earnings calls. Filings. Price charts.
For 250 years, cracking any of these took a Lovell. Rare talent. Years of training. Hours per job.
That scarcity set the price. It's why a specialist costs what a specialist costs. Few people could read the code, so the few who could got paid.
AI now reads all of them.
Cheaply. Fast. Tonight.
Crowns, Wars, and Elon’s Fortune
Is AI a bubble? Here’s my test: a bubble is when the price runs far ahead of the use.
So look at the use. At Google, AI now writes three of every four new lines of code. The FDA has cleared over 1,600 AI medical devices. Half a million people a week already pay a robot to drive them home. Scientists used it to crack protein folding and won a Nobel.
That’s use.
Some prices will run too far ahead. But the use won’t go backward. Why? Because while bubbles come and go, codebreaking never dies.
Every cracked code in history has handed someone a crown, a war, or a fortune. They put America into World War I, won Midway, and cost a queen her head.
On Wall Street, one of the most valuable codes right now is the buyout. When a deal goes public, the target’s stock can soar in a morning. Most investors read about it after the jump.
But big deals leave a trail. Filings. Hiring sprees. Patents. Executive flights. Too many signals for any human to read.
James now has a new AI system that reads all of it. While Wall Street sleeps, it’s combing through every major company, hunting for the next deal.
This week, the trail led to Elon Musk.
In June, SpaceX raised more than $75 billion in the biggest IPO in history. Now Elon has to spend it.
James’s system has flagged the company he believes Elon buys next. And you can own it today, before Wall Street catches on.