The mathlete’s secret

Every math olympiad champion in the last thirty years learned one thing first. And it wasn't a formula.

Olympiad problems are built so no memorized formula will crack them. No syllabus, no study guide. Winners are the ones who can look at a problem they've never seen and work out what it's actually asking, before they reach for a single technique.

They call it “reading the problem.”
And it’s the secret to building better solutions.

McDonald's spent three years teaching an AI to take drive-thru orders. But it put bacon on the ice cream and rang up hundreds of dollars in nuggets while customers begged it to stop. The clips went viral. In 2024, McDonald's shut it off and walked away from IBM.

Taking the order was never the hard part. The hard part is the parking lot — the noise, the accents, the kid in the back seat, the guy who changes his mind. McDonald’s automated the one step you can draw on a flowchart and left the mess underneath it untouched.

Wendy's read the same drive-thru differently. They named what actually breaks: background noise, regional dialects, a menu with billions of combinations. Then they built around it. A screen that shows the order before it reaches the kitchen. They pulled crew members off their headsets and onto the food. They ran a quiet rollout in their hometown until it worked, then the tech budget more than doubled to push AI ordering past 500 restaurants.

Sure, Wendy's used newer tech. But newer tech didn’t design a new order process and confirmation screen. Reading the problem did.

Two drive-thrus aren’t a trend, but McKinsey surveyed 1,993 companies and found redesigning the workflow was the strongest predictor of whether AI reaches the bottom line. The 42% who abandoned most of their AI projects last year probably skipped that part.

Champions know: Reading the problem is the problem.
Everyone else is hunting for a formula.