The most valuable leadership lesson isn't about finding your passion — it's about choosing to love what the job actually requires.
The Summary
- Jensen Huang called Abridge CEO Shiv Rao at midnight to unpack a founder challenge, delivering advice that flips the "follow your passion" script: fall in love with whatever the job demands, even if it's living on airplanes.
- Abridge, a healthcare AI startup building patient visit transcription tools, raised $300M at a $5.3B valuation in June with Nvidia as an investor.
- The real insight: successful founders bend their DNA to love the unglamorous parts, not just the heroic vision work.
The Signal
Huang's midnight call wasn't about product strategy or market positioning. It was about the psychological contract every founder makes with their job. Rao had emailed earlier that day seeking guidance. Huang called back the same night, on his way home from the office. That detail matters. The CEO of a $3 trillion company returning a midnight call to a portfolio founder says more about operational intensity than any keynote speech.
The advice itself cuts against decades of career guidance. "Fall in love with whatever the job is," Huang told him. Not find what you love. Not follow your passion. Fall in love with what's required. Rao's example: learning to love "living on an airplane." Five years ago, he would have designed around that constraint. Now he's convinced himself to enjoy it because that's what scaling a $5.3 billion healthcare AI company demands.
"Your job is to fall in love with whatever the job is. That is something you can do, and you can convince yourself."
This is the unglamorous truth about building in the agent economy. The sexy part is the vision: AI that frees doctors from paperwork, agents that handle clinical documentation while physicians focus on patients. Abridge's tools transcribe and summarize patient visits, turning messy conversation into structured data. That mission attracts founders. What doesn't attract them: the endless flights between hospital systems, the repetitive sales cycles, the regulatory navigation that makes software distribution feel like trench warfare.
Huang has said this before in public forums. It's easier to fall in love with what you do than to find what you love. But hearing it delivered at midnight to a founder in the weeds reveals the operational philosophy behind Nvidia's dominance. The company didn't become the infrastructure layer for AI by having people who only loved chip design. They needed people who loved supply chain optimization, customer support escalations, and partner enablement calls at 11pm.
The key pattern for agent-economy builders:
- The work compounds when you love the unglamorous parts, not just the mission
- Passion follows mastery, and mastery requires doing the boring stuff repeatedly
- The founders who win are the ones who convince themselves to love operational drudgery
Rao's journey mirrors what every AI infrastructure company faces right now. Healthcare is notoriously difficult to penetrate. Sales cycles are long. Compliance requirements are brutal. Integration timelines stretch for months. The companies that win won't be the ones with the best models or the cleanest UI. They'll be the ones whose founders genuinely love the grind of getting a hospital CIO on the phone for the seventeenth time.
The Implication
If you're building in AI, crypto infrastructure, or any part of the agent stack, audit what you're avoiding. The parts of the job you're designing around or delegating away might be exactly what the business needs you to fall in love with. Huang's framework isn't about grinding through misery. It's about retraining your reward system to find satisfaction in operational excellence, not just product vision.
Watch the founders who scale past $1 billion. They're not the ones with the best pitch decks. They're the ones who learned to love the midnight calls, the repetitive demos, the airplane food. That's the selection pressure in Web4: not who has the best idea, but who can bend their DNA to love what building actually requires.