Larry Ellison on being told you’re nuts, errors in conventional wisdom, and shipping before it’s ready
Larry Ellison on being told you’re nuts, errors in conventional wisdom, and shipping before it’s ready
The Z Fellows Newsletter - November 2, 2026
Welcome back to the Z Fellows newsletter! Every Monday we share 3 ideas - to help you build companies, ship products, and create your life's work.
1: Expect to be told you’re nuts
“When you innovate, you’ve got to be prepared for everyone telling you you’re nuts.” — Larry Ellison, co-founder of Oracle
In 1977, Ellison and his partners came across an obscure IBM research paper by Edgar Codd describing a “relational” model for databases, a radically more flexible way to store and retrieve data. IBM had the theory but treated it as research. The prevailing wisdom, even inside IBM, was that relational databases could never be made fast enough to be commercially viable. Ellison simply disagreed and set out to build one before IBM productized its own idea.
His own account of what followed: “In the very beginning, people said you couldn’t make relational databases fast enough to be commercially viable. I thought we could, and we were the first to do it. But we took tremendous abuse until IBM said, ‘Oh yeah, this stuff is good.’” The abuse, in his framing, wasn’t a sign he was wrong, it was the standard toll for being early.
Ellison treats derision as a leading indicator. If a serious idea attracts no ridicule, it’s probably not far enough ahead of the consensus to be worth much. The “you’re nuts” isn’t noise around the signal; sometimes it is the signal.
The takeaway: Being told you’re nuts is the entry fee for building something genuinely new; it is not evidence you’re wrong. Expect the abuse, budget for it emotionally, and keep shipping until the doubters ratify you.
2: Hunt for errors in conventional wisdom
“When you find errors in conventional wisdom — when everyone says A and A is not true — you gain competitive advantage. Only a few times do you have to find errors in conventional wisdom to make a living.” — Larry Ellison
Ellison frames his entire method as a search for consensus mistakes. He credits his success not to raw intelligence but to a disposition: “questioning conventional wisdom, doubting experts and questioning authority.” It’s an uncomfortable stance socially as it strains relationships with parents, teachers, bosses, but it’s the engine of every edge he’s found.
Crucially, you don’t need many. You don’t have to be contrarian about everything, or even most things; you have to be right about one big place where everyone says A, and A is false. Relational databases were that error where the whole industry said, “too slow to be practical,” and Oracle’s empire grew in the gap between that consensus and the truth.
This reframes contrarianism from a personality trait into a scarce, high-value hunt. The goal isn’t to disagree reflexively; it’s to find the specific, load-bearing belief that everyone holds and that happens to be wrong, and then to bet everything there.
The takeaway: You only need to be contrarian in one place that matters. Don’t disagree for sport, hunt patiently for the single widely-held belief that’s actually false and concentrate your bet on that one error.
3: Ship before it’s ready
“If you do everything that everyone else does in business, you’re going to lose. The only way to really be ahead is to be different.” — Larry Ellison
Oracle’s early strategy was aggressive to the point of controversy: get a working relational database to market fast, make it run on the minicomputers companies already owned, and win the head. Ellison bet on being first and improving in the open beat waiting for perfection while a competitor shipped. The company’s first major customer was the CIA, on a project code-named “Oracle,” which eventually became the company’s name.
The philosophy generalizes across his career: a bias toward action over analysis and momentum over polish. Speed to market, in his view, usually beats flawless execution because the market teaches you what to fix faster than any internal review can; a shipped-and-improving product compounds a lead that a perfect-but-late product never catches.
This is not an argument for shoddiness, it’s an argument about sequencing. Get the thing that’s genuinely differentiated into the world, then harden it under real use.
The takeaway:First-and-improving usually beats perfect-and-late. If your product is genuinely different, ship it before you feel ready and let real customers, not internal reviews, tell you what to fix.
See you next Monday,
- The Z Fellows Team
Any feedback?
Want to dive deeper? Listen to our mentor interviews here.
Interested in becoming a Z Fellow? Apply here.
Update your email preferences or unsubscribe here
© 2026 Z Fellows
228 Park Ave S, #29976, New York, New York 10003, United States
Powered by beehiiv