Founder stories
AI-native code editor built on VS Code that uses frontier AI models to help developers write, edit, and understand code faster.
How Michael acquired customers
Tools used to build Cursor
Four MIT grads spent a year wandering in the desert building CAD tools. When they pivoted to an AI code editor, they built the fastest-growing SaaS product in history.
Four MIT computer science graduates β Michael Truell, Sualeh Asif, Arvid Lunnemark, and Aman Sanger β incorporated Anysphere in 2022 with a $400K pre-seed. They spent their first year "wandering in the desert," building AI tools for mechanical engineering and CAD. They lacked domain expertise, hit data scarcity walls, and kept gravitating back to what they actually knew: code.
In early 2023, they pivoted to building an AI-native code editor. They shipped their MVP alongside GPT-4's launch in March 2023. Initial hype brought users, but usage tanked after the novelty wore off. The turning point came months later when they shipped two features that actually stuck: Cmd+K instructed edits and Codebase Indexing β the ability to ask any question about your entire codebase.
What happened next defied SaaS orthodoxy. Cursor reached $100M ARR with reportedly zero marketing spend. No sales team, no ads, no content marketing. Developers at OpenAI, Midjourney, Perplexity, and Shopify adopted it individually and spread it to their teams. By end of 2023 they had $1M ARR. Twelve months later: $100M ARR β the fastest $1M-to-$100M ramp in SaaS history.
They hired obsessively slowly β one person every 2-3 months. They avoided networking events, conferences, and speaking engagements to focus entirely on product. By the time they hit $100M ARR, the core team was still tiny. Michael described their hiring filter as seeking "macro-optimists, micro-pessimists" β people who believe in the long-term vision but are never satisfied with today's execution.
Founder-market fit matters more than the idea β pivot to what you actually know
Obsessive product focus beats marketing when the product genuinely saves users time
Hire slowly and obsess over quality β one person every 2-3 months
Failed launches are just learning β the features that drove retention came months after launch
Fork existing infrastructure (VS Code) to focus on your differentiator instead of rebuilding from scratch
Inspired by Michael's journey? Generate a business idea in the AI / ML space using AI and real founder data.
Michael achieved 3 milestones on the path to $100K ARR
The journey, decisions, and context behind this milestone
See the complete breakdown: launch strategy, validation methods, startup costs, expert analysis, replication playbook, and more actionable insights.
Upgrade to PremiumInstant access to all founder journeys
Founders with similar journeys or strategies
In 2013, I sold all my possessions, packed a backpack and a laptop, and flew to Thailand to begin my digital nomad life. I was once a lost musician ea...
On March 1st 2023, OpenAI announced the ChatGPT API. Right on that day, I came up with the idea to create a new UI to solve my own pain points with th...
My journey took me from being a Paris waiter to an $80,000/month solopreneur over seven years of persistence. After 17 failed projects, I found succes...
Get more founder journeys like this delivered to your inbox every week.