Y Combinator

Unlearn the lab: Francois Chaubard on what PhDs must shed to found

Key points

Key takeaways from Y Combinator Startup School talk by Francois Chaubard, visiting partner at YC and founder and CEO of Focal Systems (November 2026):

Vanity metrics vs what people want. Academics optimise for grades, citations and talks; Chaubard quotes Paul Graham that the most damaging thing learned at school is getting good grades — founders must instead ask what people will actually pay for.

Interesting vs important. Publishable research is often irrelevant to daily life; in a startup, only product-market fit matters (per Marc Andreessen), so founders must settle for "it just works" and ship, empiricism included.

Plagiarism is a strategy. Recounting Joe Lonsdale's MIT-vs-Stanford distinction, Chaubard argues pioneers get shot in the back: Lyft vs Uber, Anthropic vs OpenAI and Instacart all won on execution of a boring customer problem, not novelty (Steve Jobs: start from the customer experience and work backwards).

There is no finish line. Researchers jump between five ideas at once; founders must fix one problem and crank on it for decades, because an academic's breadth-first search distracts teams from the core mission.

Get on a plane. Focal spent three months building people-tracking models for retailers, who then declined to pay — a human-in-the-loop check would have surfaced it in days; Chaubard advises AI founders to validate demand before spending millions on compute.

Speed beats perfect information. Bezos' 70% rule applies: waiting for 100% of the data loses the market, as Larry Ellison's SQL proved against Michael Stonebraker's technically superior Quill.

Don't delegate the hard stuff. Hiring an MBA and staying in the lab "on VC's dime" fails; founders need customer empathy from doing sales and marketing themselves, plus the stomach to be capitalistic — charging money, beating competition (Ellison: rivals want your company dead).

Nobody cares how smart you are. Academic credibility sells nothing; per Dale Carnegie, curiosity about people closes more deals than P vs NP — talk to users where they live, not investors who rarely know what makes a great product.

Read more: Y Combinator · Watch the talk

More top stories