Jensen Huang is the 7th richest person alive. Nvidia is the first company in history to hit a $5 trillion market cap. And his management system is structurally designed to keep him stuck in operational detail.
60 direct reports. That's 10x the typical CEO span of control. No 1-on-1s ever. Every piece of feedback happens in front of the entire leadership team. His reasoning: if only one person learns from the mistake, the other 59 are waiting to repeat it.
He works seven days a week. He told Patrick Collison he's either working or thinking about work every waking moment. He was a busboy at Denny's as a teenager. He cleaned the toilets. He still brings it up decades later.
He founded Nvidia in 1993 inside that same Denny's in San Jose. 33 years later, the flat structure means fewer layers between him and the engineers than most 50-person startups have between the founder and the interns.
The structure IS the philosophy. You cannot operate at 60 direct reports and simultaneously believe grunt work is beneath you. Every layer a CEO adds is a vote that certain problems aren't worth their time. Jensen added zero layers for 33 years.
The CEOs who thought operational detail was beneath them are now running their companies on his chips.
What Pedro has built at Brex re: AI is nothing short of spectacular. I think we're going to look back on this time as one where Pedro went deep on a topic and put out a spike that will shake how all orgs run.
> be Peter Steinberger
> build a dev tool company
> sell it for $100M+
> spend 3 years in an existential crisis figuring life out
> start lifting, get absolutely jacked
come out of “retirement” just to experiment
> vibe code 40+ random projects that go nowhere
> project #44 becomes ClawdBot
devs on the internet love it
> it starts going viral
> then Anthropic shows up with trademark lawyers
> ClawdBot has to be renamed → MoltBot
> crypto scammers hijack your accounts within minutes
> quietly rebrand again → OpenClaw
> the open source community rallies around it
> repo explodes to 180K+ stars on GitHub
> suddenly the big labs are watching
months later…
acquired by OpenAI
- meet Cursor founders
- Michael Truell, Aman Sanger, Sualeh Asif, Arvid Lunnemark
- 4 MIT dropouts
- didn't invent a new language
- didn't rebuild the internet
- didn't create new hardware
- saw VS Code dominating dev workflows
- asked a simple question:
- "what if AI was the editor itself?"
- built Cursor
- looks like a normal code editor
- works like VS Code
- but AI is deeply integrated, not just a plugin
- no complex theory
- just clean execution
- shipped early
- before AI copilots became mainstream
- before every tool went "AI-powered"
- built under "Anysphere"
- scaled insanely fast
- raised $2.3B
- reached $29B+ valuation
- each founder owns ~4.5%
- turned into billionaires
- from "just another editor"
- to one of the biggest dev tools in the world
absolute legends 🐐