Paul Graham’s Reminder to Work on What Truly Interests You
October 12, 2025
Paul Graham’s “How to Start Google” is written for curious teenagers, yet the lessons hit anyone who still wants to build something big. We pulled out the ideas that stuck with us.
Three Ingredients For A Google
PG says a Google-scale company takes three things: deep skill in a technology, an idea you care about, and co-founders you like struggling with. Starting a company does not spare you from work. It gives you more of it, so you might as well pick work you do not mind obsessing over.
People try to guess which technology will be hot next decade. PG thinks prediction is a dead end. Pick the topic you cannot stop thinking about. If nothing jumps out, start with programming. It shifts you from consuming to producing. Plenty of CS majors pass exams but still write weak code because they never built anything of their own.
His answer to “How do I get good?” is simple: build your own project. When you create something for yourself you leave the classroom mindset. You move from memorising facts to solving real problems. PG points to Steve Jobs learning calligraphy just because he liked it. Years later that detour gave the Macintosh its famous typography. The dots only connected because he followed what interested him.
A New Lens On The World
Mastering a tool changes what you notice. Mark Zuckerberg arrived at Harvard and could not believe there was no school-wide face book. He coded one. Apple began because Steve Wozniak wanted a better computer for himself and a few friends. Larry Page and Sergey Brin wrote a search engine because the ones in grad school felt clumsy. None of them drafted business plans first. They scratched their own itch and found out everyone else had the same itch.
PG offers a test: if you shut down your project tomorrow, who would complain? Friends will humour your idea. Real users will ask where it went. That is how you know it matters. So instead of building whatever is trending on Hacker News, list the problems that annoy you or the people close to you. Build the smallest fix. Ship it. A rough repo or a quick blog post is enough to find the others who care.
The work feels messy because curiosity does not follow a straight line. That is the point. Interesting ideas and rare skills grow in the mess.