Ex-OpenAI and Tesla engineer Andrej Karpathy announces AI-native school Eureka Labs
Karpathy's new company will create a “teacher+AI symbiosis” where human-written materials will be scaled and guided by an AI assistant.
venturebeat.com
Ex-OpenAI and Tesla engineer Andrej Karpathy announces AI-native school Eureka Labs
Taryn Plumb @taryn_plumbJuly 16, 2024 3:23 PM
Credit: VentureBeat made with Midjourney v6
Any questions about what OpenAI and Tesla alum Andrej Karpathy might be cooking up next have been put to rest: The prominent AI researcher and computer scientist took to his account on X today to announce his new venture, Eureka Labs, which he described as a new kind of AI-native school.
The company aims to provide a “teacher + AI symbiosis” where human expert-written course materials will be scaled and guided with an AI Teaching Assistant.
As Karpathy wrote on X: “@EurekaLabsAI is the culmination of my passion in both AI and education over ~2 decades…It’s still early days but I wanted to announce the company so that I can build publicly instead of keeping a secret that isn’t.”
Expanding education in reach and extent
The ideal experience for learning anything new, Karpathy noted, is under the guidance of subject matter experts who are “deeply passionate, great at teaching, infinitely patient and fluent in all of the world’s languages.”For instance, American physicist Richard Feynman would be the best possible teacher for a course on physics.
But those types of experts “are very scarce and cannot personally tutor all 8 billion of us on demand,” Karpathy noted.
But with generative AI, this type of learning experience feels “tractable,” he noted. With Eureka Labs, teachers will still design course materials while the AI Teaching Assistant will steer students through them, allowing an entire curriculum of courses to run on a common platform.
“If we are successful, it will be easy for anyone to learn anything, expanding education in both reach (a large number of people learning something) and extent (any one person learning a large amount of subjects, beyond what may be possible today unassisted),” Karpathy forecasted.
The company’s first product will be the “world’s obviously best AI course,” LLM101n, he noted. The undergraduate-level class will help students train their own AI.
The course materials will be available online, Karpathy noted, but Eureka also plans to run digital and physical cohorts of students running through the program together.
When asked by a commenter on X whether the products would be available by subscription, free to all or a mix, Karpathy noted that he wants Eureka Labs to be “a proper, self-sustaining business, but I also really don’t want to gatekeep educational content.”
Most likely, the content would be free and permissively licensed, he said, with the revenue coming from everything else — such as running digital/physical cohorts.
“Eureka (from Ancient Greek εὕρηκα) is the awesome feeling of understanding something, of feeling it click,” he posted in a comment thread. “The goal here is to spark those moments in people’s minds.”
Sharing ‘a gift’
Excitement for Karpathy’s new venture is palpable; the announcement on X was met with tens of thousands of likes, comments, reposts and reactions, with users calling him a visionary, a gifted teacher and even the GOAT (greatest of all time). The company’s X page, which launched today, already has nearly 10,000 followers (as of the posting of this story).“This is going to be so damn impactful!” exclaimed one user.
Another commented: “Instead of going off onto an island and [counting] his fortune the man returns the promethean flame of wisdom and knowledge to the next generation.”
Sequoia Capital partner Shaun Maguire praised Karpathy for his “gift for teaching,” adding, “thank you for sharing and scaling that gift!”
Another X user gushed, “The most amazing contribution from you, @karpathy is going to be this. @EurekaLabsAI is going to be phenomenal. You are really one of the most distinguished selfless person.”
Finally at his ‘real job’
The OpenAI founding member and former Tesla AI scientist said his involvement in education transcended from “YouTube tutorials on Rubik’s cubes” to CS231n at Stanford, a 10-week deep learning for computer vision course. He also leads an independent “ Neural Networks: Zero to Hero” course on building neural networks from scratch.Karpathy noted that he has been involved in academic research, real-world products and AGI research throughout his career, but that all of his work has “only been part-time, as side quests to my ‘real job,’ so I am quite excited to dive in and build something great, professionally and full-time.”
After co-founding OpenAI in 2016, Karpathy initially left the then-nonprofit in 2017 to serve as senior director of AI at Tesla, where he led the computer vision team of Tesla Autopilot, according to his website. He then rejoined OpenAI in 2023 shortly after the earth-shaking release of ChatGPT, then departed the company again this February and has been a free agent ever since.
At the time, he explained on X that his “immediate plan is to work on my personal projects and see what happens. Those of you who’ve followed me for a while may have a sense for what that might look like ;)”
Karpathy earned his PhD from Stanford University under the tutelage of AI godmother Fei-Fei Li, focusing on convolutional/recurrent neural networks and their applications in computer vision, natural language processing (NLP) and their intersection. He also worked closely with venerable AI researcher Andrew Ng.
The AI leader is no doubt passionate about the cross-section of AI and education, commenting: “We look forward to a future where AI is a key technology for increasing human potential.”
