Michael Truell: Building Cursor at 23, Taking on GitHub Copilot, and Advice to Engineering Students
Timestamps are as accurate as they can be but may be slightly off. We encourage you to listen to the full context.
Podcast Summary
In this episode, Michael Truell, co-founder of Cursor, shares his journey from building mobile games in middle school to creating one of the fastest-growing AI coding tools today. He traces his path through failed startup attempts—from mechanical engineering CAD tools to encrypted messaging systems—before discovering the massive opportunity in AI-powered coding (12:00). Truell reveals how Cursor went from zero to $100 million in revenue within just two years, initially building their own code editor from scratch before pivoting to focus entirely on AI capabilities (13:56). The conversation explores the "long messy middle" of AI transformation in software development, where human programmers will increasingly work alongside AI as colleagues rather than tools, fundamentally reshaping how we build software while still requiring deep technical understanding and collaboration skills.
Speakers
Michael Truell
Co-founder and CEO of Cursor, the AI-powered code editor that grew from $1M to $100M ARR in 2024. MIT graduate who started programming in middle school, built neural networks from scratch in high school, and previously founded multiple startups including AI projects in mechanical engineering and robotics.
Host
Interviews high-growth startup founders and tech leaders, focusing on the origin stories and pivotal decisions behind breakthrough companies. Known for deep-dive conversations exploring the journey from initial idea to market success.
Key Takeaways
Bet on the Long-Term Vision, Not the Competition
When everyone thought GitHub Copilot had "won" with $100M+ revenue, Cursor's founders saw the bigger picture: all of coding would change in 5 years, not just get incrementally better. They realized existing players were "making great products a bit better" instead of "aiming for a world where all of coding gets automated." (11:27) Don't compete on features—compete on transformative outcomes.
Build Your Own Infrastructure When Distribution Demands It
Cursor initially built their own editor from scratch, taking 3 months to ship publicly. While this seemed insane, it gave them control over the AI integration experience that extensions couldn't deliver. (13:35) Sometimes the "inefficient" path creates the moat—especially when your core value prop requires deep platform control.
Resist the Loudest Users, Follow Your Conviction
Early users pulled Cursor toward non-coders and tech-stack-specific solutions, but the team resisted both directions. They had "a really loud segment of users that didn't know how to code at all" and others wanting narrow vertical tools. (19:02) Customer feedback is crucial, but transformative products often require ignoring the loudest voices to serve the silent majority.
Product Quality Drives Exponential Growth in Technical Markets
Cursor grew from 1M to 100M users in 2024 by obsessing over product improvements—making it code base aware, predicting next actions, increasing speed and accuracy. "If you make the product better, you see it in the numbers immediately," Michael explains. (20:28) In technical markets, superior product performance creates word-of-mouth flywheels that marketing can't replicate.
Stay Small and Focused Until Product-Market Fit Is Undeniable
Despite growing to millions of users, Cursor ended 2023 with fewer than 10 employees. The founders were "very patient early on and focused on hiring a lot less than we probably should have." (24:11) Premature scaling kills more startups than under-hiring—especially when your core team can execute the entire value chain.
Statistics & Facts
- GitHub Copilot was generating approximately $100 million in revenue by 2022, despite being a relatively new product in the AI-powered coding space. (11:16)
- Codex (the model behind GitHub Copilot) cost approximately $100k in training costs, demonstrating that training effective AI models was more accessible than many believed at the time. (16:54)
- YC batch composition shifted dramatically from single-digit percentage of companies using Cursor in 2023 to roughly 80% adoption in 2024, showing rapid market penetration among top builders. (21:31)
