Eureka Labs is an AI-native education company founded in 2024 by 📝Andrej Karpathy, former Director of AI at 📝Tesla Motors and founding team member of 📝OpenAI. The company builds courses in which a human teacher designs the curriculum and an AI teaching assistant scales personalized instruction, with the stated long-term goal of letting anyone learn anything. Its first product is LLM101n, an undergraduate-level course that teaches students to train a small large language model.
Eureka Labs was registered as a Delaware LLC on June 21, 2024 and announced publicly on July 16, 2024 via Karpathy's X post. The framing — "a new kind of school that is AI native" — positions the company against both traditional MOOC platforms and the broader cohort of AI tutoring tools by treating the teaching assistant as the core product rather than a feature layer. Karpathy has framed it as the culmination of two decades of his own work in AI and education, including the deep learning course he designed at Stanford that became one of the school's largest classes.
Eureka Labs is the clearest applied test of whether generative AI can actually scale expert teaching rather than just generate plausible-sounding answers. It belongs in the library because Karpathy is using it as a public laboratory for ideas — context engineering, LLM-maintained knowledge bases, agentic instruction — that overlap directly with the MythOS thesis.