Tip
Brian Tolkin talks about launching uberPOOL in China: “I’ve slept on the floor in China before launching uberPOOL, and when you reflect the stress onto your teams, everybody tenses out. It counterintuitively doesn’t produce better outcomes.”
Turns out AI implementation crises work the same way.
Your AI agent hallucinated in production. Customer service is getting angry calls. Your CEO is asking for a postmortem. The board wants an emergency update. Everyone’s looking at you. If you panic, the whole team panics. Engineers start making desperate changes without thinking. Product managers start promising fixes they can’t deliver. Everyone tenses up and performance actually degrades.
Younger managers think showing urgency means showing stress—they mirror the pressure they feel. You know better. You’ve been through enough product failures, security incidents, and PR fires to recognize the pattern: when leaders reflect stress onto teams, decision quality drops. People stop thinking clearly. They make reactive choices instead of strategic ones.
That calibration—staying calm when everything is on fire—comes from living through enough crises to know panic doesn’t help. You’ve seen teams that stayed composed solve problems faster than teams that freaked out. You know the difference between “this feels like an emergency” and “this IS an emergency that requires calm, systematic response.”
This isn’t about suppressing emotion or pretending things aren’t serious. It’s about knowing that your team takes emotional cues from you. When you project calm focus instead of panic, they stay sharp. That’s pattern recognition from watching enough crises play out.
Context
Brian Tolkin was Head of Product at Opendoor and spent nearly five years at Uber starting as employee 100. He launched uberPOOL globally, experiencing multiple high-stakes crises including the China launch where nothing worked 12 hours before go-live.
The insight about not reflecting stress came from watching teams perform differently under pressure—calm leaders produce better outcomes. For experienced leaders navigating AI implementation challenges, this pattern recognition is the advantage: you know panic degrades performance because you’ve seen both calm and panicked responses enough times to recognize the pattern.