How To Go From Researcher to Founder

Academic habits that hold researcher founders back

Francois Chaubard argues that the habits that make a successful researcher can actively hurt a startup founder. Academia rewards novelty, rigor, publication and individual intelligence, while startups reward customer validation, copying proven ideas, fast decisions and selling. He draws on his own move from Fei-Fei Li's lab to founding Focal Systems, where months of model training failed because retailers would not pay. The fix is to leave the office, talk to users and adopt a capitalist, people-focused mindset.

Takeaways

• Academic skills can hinder founders. Francois Chaubard was a researcher in Fei-Fei Li's lab and says those skills held him back as a startup CEO. Academic-led startups often fail because what research rewards can be antithetical to founding, and becoming a founder means killing part of your academic identity.
• Vanity metrics mislead academics. Academia optimizes grades and citation counts. Founders must ask what people will actually pay for, not what they say they would pay for. Paul Graham said the most damaging thing school taught was getting good grades.
• Important beats interesting for founders. Academics get published by explaining mechanisms, while founders must be comfortable shipping something that just works. Marc Andreessen's view that only product market fit matters applies here.
• Copying is often the right strategy. Academia rejects non-novel papers, but Lyft copied Uber and Rappi followed Instacart. Joe Lonsdale's Stanford-style startup starts with one boring customer problem and iterates forever, as Instacart has for 15 years.
• Startups have no finish line. Academia has submission dates and finals, but startups do not. Founders need long-horizon commitment, while researchers spread themselves across five ideas at once and never finish anything.

Customer validation playbook

1. Get out of the office and get on a flight.
2. Meet users where they live and work.
3. Ask whether they would pay for the solution, not just whether the problem exists.
4. Validate demand before spending millions on compute. Start with a human in the loop.
5. Remember competitors are already doing this, and they will beat you if you do not.

Sales, speed and people

• Hyperbole is acceptable in sales, never lie. The '15 minutes could save you 15%' slogan is beautifully inexact yet hugely successful.
• Use the 70% rule. Bezos said most decisions need about 70% of information. Waiting for certainty lets competitors launch first.