Andrej Karpathy, former head of AI at Tesla and researcher at OpenAI, is launching Eureka Labs, an “AI native” education platform. In tech speak, that usually means built from the ground up with AI at its core. And while Eureka Labs’s AI ambitions are lofty, the company is starting with a more traditional approach to teaching.
San Francisco-based Eureka Labs, which Karpathy registered as an LLC in Delaware on June 21, aims to leverage recent progress in generative AI to create AI teaching assistants that can guide students through course materials.
Eureka Labs envisions AI assistants or personalities that would work with a human teacher to allow “anyone to learn anything,” according to Karpathy, who posted the news on X. Teachers would still design the course material, but they’d be supported by this AI assistant. The startup does not yet appear to have built or tested the efficacy of integrating AI assistants into the classroom. At least one Georgia State University study found that AI teaching assistants helped some students get better grades.
Karpathy’s post points to a potential future where those assistants are based on real people – a la Meta’s weird celebrity chatbots or Character AI’s character chatbots. The post on X, which is mirrored on Eureka’s bare bones new website, doesn’t provide much information on this new startup, like whether this is just a MOOC with a chatbot or if this is a product that Karpathy would like to implement in, say, high schools.
Karpathy didn’t respond to TechCrunch’s request for more information.
Along with the post on X announcing the news, Karpathy included what is likely an AI generated image of a futuristic school, replete with a spaceship-like building, solar panels everywhere (even on the floor) and a smiling girl with…is that three hands?
Despite stating that Eureka Labs aims to build AI teaching assistants, Karpathy also noted that the new venture’s first product will be an AI course, LLM101n, an undergraduate-level class that will help students train their very own AI. This mini-me will be like a smaller version of the AI teaching assistant Eureka Labs hopes to build and scale, according to Karpathy. The AI pioneer wrote on X, and on Eureka’s bare bones new website, that the course materials will be available online, and that the startup will run both digital and physical cohorts of people going through the materials together.
The link for this AI course leads to a GitHub repository that hints at a different type of course than Eureka Labs is advertising – instead of “How to build an AI assistant,” the link leads to a how-to for building a “Storyteller AI Large Language Model (LLM).”
“Hand in hand, you’ll be able [to] create, refine and illustrate little stories with the AI,” reads the copy on the page. The class promises to teach eager AI students how to “build everything end-to-end from basics to a functioning web app similar to ChatGPT, from scratch in Python, C and CUDA, and with minimal computer science prerequisites.”
Whichever course Eureka Labs intends to introduce first, neither appears to be complete. A note posted on the GitHub page says the course will take time to build and there’s no specific timeline.
It’s also unclear if Karpathy has self-funded Eureka Labs or has received backing from investors, and what the startup’s business model is. There were no public filings of any investments related to Eureka Labs. The startup’s LLC filing with California’s Secretary of State is signed by Karpathy alone, and he hasn’t divulged whether he’s working with other high-profile leaders in the AI sector.
Karpathy noted on X that Eureka Labs is the culmination of his passion in both AI and education over the last two decades. Karpathy taught deep learning for computer vision at Stanford University until 2015, when he left to co-found OpenAI. Two years later, Karpathy moved on to Tesla to head up the automaker’s AI team, where he led the computer vision team of Tesla Autopilot. Autopilot is Tesla’s advanced driving assistance system that relies on cameras to ingest environmental data and then perform certain driving tasks like cruise control and automatic steering.
Karpathy left Tesla in 2022 and migrated back to OpenAI, where he led a small team related to ChatGPT. In February, the research scientist stepped down from his role at OpenAI, too. In both instances, Karpathy insisted that there was no drama or fallout that led to his decision to leave.
Throughout his career at Tesla and OpenAI, Karpathy has continued to be an educator. He currently leads an online course called Neural Networks: Zero to Hero that helps students learn to build neural networks from scratch in code. Karpathy also has a YouTube channel where he semi-regularly posts lectures on LLMs and AI.
