
In this Y Combinator Startup School talk, Jake Heller, co-founder and CEO of Casetext, explains how his team stopped everything at $20M in revenue to build CoCounsel, the first AI assistant for lawyers, on early access to GPT-4 — a bet that led to a $650 million acquisition by Thomson Reuters. He lays out a practical playbook across three areas: how to pick an idea (target jobs people already pay for, in categories of assist, replace, or do the previously unthinkable), how to actually build reliable AI (map exactly what a professional does, turn each step into prompts or plain code, and grind on evaluations until you pass 97-99% of tests), and how to market and sell it (build a genuinely great product first, price to value, build trust with head-to-head comparisons, and watch out for pilots that never convert to real revenue).
Mastra is an open-source TypeScript framework for building production-ready AI agents, workflows, and retrieval-augmented generation pipelines. Built by the team behind Gatsby, it gives developers agents, durable workflows, memory, RAG, tool calling, evals, and observability in one stack - so JavaScript and TypeScript engineers can move from prototype to production without leaving their language. Mastra is developed by Kepler Software, a San Francisco startup, and reached 25,000+ GitHub stars and 300,000+ weekly npm downloads with adoption at companies like Brex, Sanity, Replit, and PayPal.