When the Guardrails Come Off: What F1's Biggest Rule Change Taught Me About AI Transformation.
An F1 weekend in Miami revealed the gap between AI pilots that succeed in sandboxes and rollouts that fail at scale, and what it takes to perform when the guardrails come off.
Last Friday I flew out of Atlanta at 5 a.m., spent a day at the Miami Grand Prix weekend, and was back home by 11:30 p.m. I went as someone who had never really followed Formula One. I came home a fan, and I've been thinking about why that happened.
It wasn't the spectacle alone, though standing close enough to feel a car through your chest at 180 miles per hour does something to you. It was the proximity to a level of preparation and precision that most people never get to observe up close. Being inside that environment, even briefly, changed how I understood it. I watched the sprint race and the Grand Prix from my couch on Saturday and Sunday, and I was paying attention in a completely different way than I would have been a week earlier.
That shift, from passive observer to genuinely invested, is exactly what most organizations are failing to create in their people around AI. But I'll get to that.
A 19-year-old, a penalty, and a Sunday recovery
Kimi Antonelli arrived in Miami as the youngest championship leader in F1 history, an Italian teenager already carrying two consecutive Grand Prix wins. I watched Sprint Qualifying from the stands. Antonelli took P2 behind Norris. The stewards were already strict on track limits, dropping Albon five places after a mid-session breach.
Then Saturday happened.
Saturday morning Sprint
A penalty drops Antonelli from P4 to P6. McLaren's Lando Norris wins.
Saturday afternoon Qualifying
Antonelli takes pole position for Sunday's Grand Prix.
Sunday Grand Prix
Antonelli wins his third consecutive race.
Three poles, three wins, three weekends.
I watched both days from home, but I understood what I was watching in a way I wouldn't have without Friday.
What changed between the Sprint and the Grand Prix wasn't his talent. What changed was his ability to read what had gone wrong, make targeted adjustments, and execute cleanly when it mattered. His team gave him the right information and the space to recalibrate without overreacting. That combination is harder to build than it sounds, and most organizations trying to navigate AI change don't have it.
Something most people missed about the Miami race
This wasn't a normal race weekend. The 2026 Formula One season already represents the sport's largest rule overhaul in decades, with new power units, new aerodynamics, and cars essentially rebuilt from scratch.
The biggest challenge is probably that we are starting from scratch on everything, new tyres, new fuel, new engine, new chassis, new sporting regulations, new everything.
Then, just before Miami, the FIA convened an emergency meeting and agreed to a second round of changes, implemented at the race itself.
Teams arrived at one of the season's biggest events with rules that had just shifted again. The teams that handled it well weren't necessarily the richest or the most resourced. They were the ones that had built cars and cultures capable of absorbing a change without losing their footing. There's an obvious parallel to how organizations are responding to AI right now, and it's not a flattering one for most.
What the simulator actually showed me
Before the race, I had the chance to get into a full Formula One simulator. Racing line on. Automatic shifter on. Two assists that no actual driver uses in competition.
I ran three laps and still struggled.
The drivers I watched preparing for the race had logged dozens of qualifying and practice runs, all without those aids, just to be ready to compete in a 57-lap Grand Prix. The gap between my simulator experience and what I saw on track wasn't mainly about reflexes. It was about accumulated reps under real conditions, and the ability to make hundreds of small decisions per lap from a place of deeply internalized understanding. And yes of course, innate inhuman talent to drive a $15 million car at 200+ miles per hour.
This is where most AI change programs are quietly failing. The pilot worked. The sandbox was successful. The guardrails held. Leadership called it a win. Then the organization tried to scale it, the guardrails came off, and the gap showed up. Not because the tools were wrong, but because the organization never actually logged the laps.
What separates the teams that adapt
Antonelli didn't win on Sunday because he forgot what happened Saturday. He won because he had the infrastructure around him to process what went wrong, adjust without overcorrecting, and execute under pressure.
The organizations doing well in AI transformation right now share something similar. They didn't just buy tools and run training programs. They changed how decisions get made, how work gets structured, and how quickly they can identify when something isn't working and adjust. That's a different kind of investment than most companies are making, and it's the one that actually shows up in results.
The rules are always changing. Miami showed us that even the most complex teams can adapt. The question isn't whether your organization has the guardrails in place, or the 19 year old prodigy is behind the wheel of your codebase.
It's whether you've put in enough real reps to perform when those guardrails aren't there.