College Dropout Builds $30M Startup While Fighting Cancer | Jenni AI David Park | EO

Eight years of failure, a cancer scare, then Jenni AI

Jenni AI is an academic AI assistant built by a founder who spent eight years failing first. The product found its shape as each new model arrived, moving from an agency that sped up human writers to software that replaced them, and survived by moving to Malaysia. Cancer and investor near-misses shaped how Park runs it now.

Takeaways

• Jenni AI grows 15 to 20 percent monthly. The academic AI assistant for research and writing has raised from Jason Calacanis and won the AI Grant from Nat Friedman and Daniel Gross. It runs at about $2.5m ARR with over 2m users.
• Eight years of failure preceded Jenni AI. He lost all his money at 16 on a clothing brand and later built an app that added users across four social networks. An AI dating app failed after a venture capitalist passed because users delete it once they find a partner.
• Each model jump reshaped the product. Jenni began when only GPT-2 existed, as an agency making human writers 10 percent faster. GPT-3 turned it into software without human writers, and a wave of student users pushed the team toward academics.
• Malaysia kept the company alive. A scout for Jason Calacanis invested $100,000, then growth stalled near $2,300 a month. Park and his cofounder flew to Malaysia, where everything costs a quarter of the price, and he says the company would have died otherwise.
• Viral traffic broke the product overnight. A Twitter thread brought 10 new users every second. Jenni burned its 5,000 OpenAI credits within an hour, and Park spent the night answering users on Intercom, his first feeling of any success.
• Cancer drove him to raise before surgery. Diagnosed with thyroid cancer in Korea and given a 50 percent chance, he wanted money raised so the startup would survive if he did not. He had a fever on surgery day and told the doctor to proceed.

The $250,000 that almost slipped away

1. He was invited to the K Startup Grand Challenge in Korea, where a top 10 finish wins money, and got a venture capitalist to commit $250,000.
2. He stayed out late the night before the meeting, and the investor pulled out.
3. Two months from running out of money, he found another backer at the last minute.
4. The near-miss taught him that founders carry enormous responsibility and must be more disciplined than employees.

Park's advice for early founders

• Talk to users more than anything else, and ask uncomfortable questions about their workflow and pain points.
• Do not guard the product. The lifeblood of a startup is knowing what to build, so giving it away is fine.