Stock Trader Study Says AI Should Copy the Brain's Wiring

A team of researchers posted a paper this week proposing a new way to describe how groups of decision-makers affect each other. Their idea uses "probability waves" — a math tool borrowed from physics — to model people whose choices ripple into one another instead of being made alone. When they tested it on Chinese intraday stock market data, they found 82% to 94% of trading decisions (89% on average) matched their model. Fewer than 5% looked purely independent, which is what standard economics has long assumed traders are.

So what does that have to do with AI? Today's AI runs on artificial neural networks — huge webs of trillions of dials that are expensive to run and hard to explain. The authors say adding what they call "adaptive entangled game modules" (rules that let AI systems influence each other like people in a crowd) could lead to smaller, cheaper "human-like processing units." They think this matters most for robots and machines that move around in the real world, where being small and efficient counts.

The paper also makes a bolder claim: that trader behavior is a window into the brain, and that the results support a hypothesis that nerve fibers in the brain are linked in ways science doesn't fully understand yet. That's a leap, and the authors present it as a hypothesis, not proof. The work mixes finance, quantum physics and neuroscience, which is unusual and will draw skepticism.

The catch: this is a preprint, meaning it hasn't been checked by other scientists, and it describes no product, no company and no timeline. The "quantum" language is largely metaphor. About 11% of behavior still isn't explained by their model. Treat this as a hint about which direction AI research might go — not a promise of something arriving soon.