Lenny's Podcast · Mar 13Inside Bolt: From near-death to one of the fastest-growing products in history | Eric Simons
Spent seven years building WebContainer, an in-browser operating system, and survived by keeping burn very low and raising nothing for two or three years. It only ever sold one $9 plan before Bolt.
Tips
- Replace the single flat price with tiered plans that sell more inference▶ 31:51Users used up $9 of inference in 48 hours and begged to pay more; competitors later copied the tiered model.
- Run compute in the user's browser to get a generous free tier▶ 23:47A Spotify clone was built and running in about 60 seconds, and deploying to Netlify cost nothing in server compute.
- Keep burn low until customers pull the product from you▶ 27:38They skipped the 2020 and 2021 hiring surge; he says tripling the company would have meant no Bolt.
- Make a deep technology bet and survive long enough for it to land▶ 34:59Bolt was an overnight success seven years in the making, launched when the company was about to shut down.
- Staff with fewer people holding more context, and hire from your community▶ 35:48A core group of five or six people has stayed five years or more.
- Wait for the model threshold before building an AI product▶ 1:07:10Bolt was built in 90 days after Sonnet and grew rapidly right away.
