Andrej Karpathy enters the AI education startup field, dedicated to making it po
On July 16th local time, one of the founding members of OpenAI and former Tesla Artificial Intelligence Director, Andrej Karpathy, announced the establishment of a "new type of school" called Eureka Labs (official website link at the end of the article).
This new company, headquartered in San Francisco, USA, is committed to creating an "AI-native" educational experience, making it possible for "anyone to learn anything" through the cooperation of AI teaching assistants and human teachers.
Karpathy shared his vision on the X platform: "Eureka Labs is the culmination of my nearly 20 years of passion in the fields of artificial intelligence and education."
He stated that the ideal learning experience should be guided by subject matter experts who are passionate about teaching, extremely patient, and proficient in multiple languages.
However, such experts are very rare and cannot personally guide the 8 billion people around the world. Through generative artificial intelligence, this learning experience has finally become more feasible.Pioneering Course Design
At present, Eureka Labs appears to be a brand-new artificial intelligence learning platform that is still in its very early stages of development.
Advertisement
On its official website, it lists its first course, LLM101n, which is specifically designed for undergraduates and teaches students how to train their own large language models.
This course will guide students from the basics, gradually building a "storytelling artificial intelligence large language model," and ultimately creating a web application similar to ChatGPT. The course content will be provided online, and plans are in place to launch digital and physical learning groups in the future.
The course link on the official website points to a GitHub repository, showing that the course includes 17 chapters, with content ranging from easy to difficult, covering language modeling, machine learning and backpropagation, attention mechanisms and Transformers, optimization, datasets, fine-tuning, and multimodality, etc.The course introduction for LLM101n states: "We will start from the basics, building a web application similar to ChatGPT using Python, C, and CUDA, with no high level of computer knowledge required. By the end, you will have a relatively in-depth understanding of artificial intelligence, large models, and deep learning."
The Future of AI-Assisted Education
In the future, Eureka Labs will not just be an online learning platform. In Kapasi's vision, AI teaching assistants will collaborate with human teachers to provide personalized guidance and expand the accessibility of quality education.
Through this "teacher + AI symbiosis" model, human teachers are responsible for designing course materials, while AI teaching assistants are responsible for guiding students through their learning, thereby significantly expanding the scope and depth of education.
Kapasi points out that if successful, this will enable anyone to easily learn any knowledge, expanding the scope of education (a large number of people learning a certain knowledge) and depth (an individual learning a large number of subjects, beyond the current range without tutoring).When asked by a commenter whether Eureka Labs' products would be offered through subscription, be freely available, or both, Kapasi stated that he hoped Eureka Labs would become "a proper, self-sustaining enterprise, but I also don't want to limit educational content."
"Eureka (from the ancient Greek εὕρηκα) is a wonderful feeling of sudden enlightenment," he wrote in the comment section, "Our goal is to inspire such moments in people's minds."
Below Kapasi's tweet, Google's Chief Scientist Jeff Dean expressed congratulations: "Congratulations! Artificial intelligence will have a huge impact on education, and I know this is something you have always been passionate about."
However, the large-scale introduction of artificial intelligence technology in the field of education also faces some challenges and controversies.
For example, how to clearly define the responsibilities between artificial intelligence teaching assistants and human teachers? How to protect students' privacy and data security? These issues all need to be seriously considered and resolved in the development process of Eureka Labs.Behind Self-Employment
Kapasi has been deeply involved in the field of artificial intelligence for many years. He earned his Ph.D. in 2015 from Stanford University, under the guidance of Professor Fei-Fei Li, focusing on neural networks and their applications in computer vision and natural language processing.
He was one of the founding members of OpenAI and left in 2017 to join Tesla, where he served as the Senior Director of Artificial Intelligence and led the autonomous driving team.
In 2023, he returned to OpenAI, but he resigned again in February of this year to become a freelancer. He is also a recipient of the "35 Innovators Under 35" award from the MIT Technology Review in 2020.
Image | Kapasi selected as one of the "35 Innovators Under 35" by the MIT Technology Review in 2020 (Source: MIT TR)Kapasi stated that founding Eureka Labs is the epitome of his passion in the fields of artificial intelligence and education.
"My interest in education began with Rubik's Cube tutorials on YouTube, developed into founding the CS231n course at Stanford, and then to my most recent Zero-to-Hero AI series.
My work in the field of artificial intelligence includes academic research at Stanford, practical products at Tesla, and general artificial intelligence research at OpenAI," he wrote.
The vision of Eureka Labs is not limited to its initial artificial intelligence courses, but also implies a broader curriculum system, which may cover various disciplines.
Kapasi hopes to inspire people's interest in learning and enhance human potential through this platform. "Today, we focus on building the LLM101n course, but we look forward to artificial intelligence becoming a key technology to enhance human potential in the future. What else would you like to learn?"According to currently available public information, the funding source for Kapasi's founding of Eureka Labs could be self-financed, or it might have received support from other investors.
However, what is certain is that this pioneer in artificial intelligence is going all out, committed to bringing revolutionary changes to education through Eureka Labs. No matter how the future unfolds, this will be an important milestone in the integration of artificial intelligence and education.
If you want to get a sneak peek, the official website of Eureka Labs is here ().