Andrej Karpathy, previously in charge of AI at Tesla and a researcher at OpenAI, is introducing Eureka Labs, an AI-native education startup.
In technical terms, this typically means it is developed with AI as a central focus from the beginning. Although Eureka Labs has ambitious goals for AI, the company is initially taking a more traditional approach to teaching.
The San Francisco-based Eureka Labs, established as an LLC in Delaware on June 21 by Karpathy, aims to utilize recent advancements in generative AI to develop AI teaching assistants capable of guiding students through course materials.
Eureka Labs has a vision of AI assistants or personalities that would collaborate with human teachers to make learning accessible to everyone, as stated by Karpathy in a post on X.
While teachers would still be responsible for creating course material, they would be supported by these AI assistants.
The startup has not yet developed or tested the effectiveness of integrating AI assistants into the classroom, but a study at Georgia State University found that AI teaching assistants improved grades for some students.
Karpathy’s post suggests a potential future where these assistants are based on real people, similar to Meta’s celebrity chatbots or Character AI’s strange chatbots.
The post on X, also found on Eureka’s basic new website, lacks details about the startup, such as whether it is simply a MOOC with a chatbot or a product intended for implementation in high schools.
Karpathy did not reply to the inquiry for additional details. In addition to the announcement on X, Karpathy shared what appears to be an image generated by AI depicting a futuristic school with a spaceship-like structure, solar panels covering the floor, and a smiling girl with an unusual number of hands.
Despite the focus on developing AI teaching assistants, Karpathy mentioned that Eureka Labs’ initial offering will be LLM101n, an undergraduate-level AI course designed to help students train their own AI.
This miniature version of the AI teaching assistant that Eureka Labs plans to create will be available online, and the startup will facilitate both digital and in-person cohorts for individuals to go through the course materials together.
The AI course link provided by Eureka Labs directs to a GitHub repository that suggests a different course focus. Instead of “How to build an AI assistant,” the link leads to instructions for creating a “Storyteller AI Large Language Model (LLM).”
The course description mentions the ability to collaborate with the AI to create and refine short stories. It promises to teach aspiring AI students how to develop a functioning web app similar to ChatGPT from scratch using Python, C, and CUDA, with minimal prerequisites in computer science.
Eureka Labs seems to be in the process of developing its first course, as indicated by a note on their GitHub page. However, the completion timeline is unspecified.
It is also unknown whether Karpathy has personally funded Eureka Labs or if the startup has received investments.
Additionally, details about the startup’s business model and potential partnerships with other AI industry leaders have not been disclosed.
Public filings related to investments in Eureka Labs have not been found, and the LLC filing with California’s Secretary of State is signed by Karpathy alone.
On X, Karpathy mentioned that Eureka Labs represents the culmination of his dedication to both AI and education for the past twenty years.
Until 2015, Karpathy was a professor at Stanford University, teaching deep learning for computer vision, before co-founding OpenAI. Following this, he joined Tesla to lead the automaker’s AI team, specifically overseeing the computer vision team for Tesla Autopilot.
Autopilot is an advanced driving assistance system by Tesla that utilizes cameras to gather environmental data and execute specific driving functions such as cruise control and automatic steering.
In 2022, Karpathy departed from Tesla and returned to OpenAI, where he headed a small team focused on ChatGPT.
However, he later resigned from his position at OpenAI in February. Karpathy emphasized that his decision to leave in both cases was not due to any conflict or disagreement.
Throughout his time at Tesla and OpenAI, Karpathy has maintained his role as an educator.
He is currently leading an online course titled “Neural Networks: Zero to Hero,” which aims to teach students how to construct neural networks from the ground up using code.
Additionally, Karpathy regularly shares lectures on LLMs and AI on his YouTube channel.
Stories You May Like