Here’s the thing about AI that nobody talks about: it’s not actually new. It’s just finally good enough to matter.
Take Demis Hassabis, the neuroscientist who won the Nobel Prize in Chemistry last year for building AlphaFold2—an AI that mapped 200 million proteins in hours. Before that, figuring out protein shapes took researchers years in the lab. Now it’s a coffee break.
That same deep learning technology? It’s now being applied to something way less noble but infinitely more profitable: picking stocks.
I’ve been building computer systems to find growth stocks since the 1970s, when most Wall Street guys thought I was insane for using a mainframe. Over 47 years, my quant models identified 676 stocks that doubled—Microsoft in 1987, Apple and Nike in 1988, Nvidia way before ChatGPT made it famous. A thousand bucks in Nvidia back then? Over a million today.
But here’s where it gets interesting.
Last year, I started working with TradeSmith to layer AI on top of my existing Stock Grader system. Not to replace it—to supercharge it. We’re talking about adding the same pattern-recognition technology that’s diagnosing pancreatic cancer three years early and designing drugs in hours instead of years.
The backtesting results are honestly ridiculous. A stock like DXP Enterprises that would’ve made you 615% with my old system? The AI version suggests 3,626%. Broadcom’s 292% becomes 6,284%. Same stocks, same time period, just smarter timing on when to actually buy and sell.
Across the board, we’re seeing up to 20 times more money than the original system alone.
Look, I get the skepticism. I felt it too. But when you’ve spent five decades hunting for edges in the market and you finally find one that actually works, you pay attention. This is the biggest edge I’ve seen in my entire career—and I’ve seen a lot.
The real kicker? Most investors still don’t realize what’s coming. Back in the 1970s, computers seemed absurd for stock picking. Today, they handle 80% of daily trading volume. AI is about to do something even bigger.
