Andrew Ng, the influential computer scientist and entrepreneur who gave the artificial intelligence industry some of its most enduring phrases — from “AI is the new electricity” to “agentic AI” — is now making a billion-dollar bet against the dominant fear that AI will destroy jobs. His new company, LearnVector, launched with a $100 million investment from Coursera (where Ng serves as chairman), aims to prove that AI can act as a personal tutor, helping adults reskill and stay valuable in a rapidly changing labor market.
Ng’s perspective stands in stark contrast to the recent wave of layoffs at major tech companies. Amazon, Meta, and Block have all cited AI-driven automation as a reason for cutting tens of thousands of positions. “Conventional wisdom holds that AI will replace people, and I believe in the opposite,” Ng told Reuters. Instead of a “job apocalypse,” he argues that AI will make workers more productive and create fresh demand for new skills, a vision that LearnVector is designed to realize.
The man behind the terms
To understand the significance of Ng’s latest move, it helps to look at his track record. Born in the United Kingdom and raised in Hong Kong and Singapore, Ng earned his PhD in computer science at the University of California, Berkeley, under the supervision of Michael I. Jordan. He later became a professor at Stanford University, where he co-founded the Google Brain deep learning project with Jeff Dean, Greg Corrado, and others. That work helped kickstart the modern deep learning revolution.
Ng later served as chief scientist at Baidu, where he built the company’s AI research team. Throughout his career, he has been a prolific educator: he co-founded Coursera in 2012 with Daphne Koller, creating a platform that brought university courses online to millions. He also founded deeplearning.ai, a series of accessible courses that taught a generation of engineers the fundamentals of neural networks. The phrase “AI is the new electricity” came from a talk Ng gave in 2017; it became a rallying cry for the technology’s transformative potential. “Agentic AI” is a more recent coinage, referring to AI systems that can take autonomous actions toward goals rather than merely respond to prompts.
Ng’s belief that AI will augment rather than replace human labor has been consistent. In a 2023 op-ed, he wrote that “AI will create more jobs than it eliminates” and called for a focus on reskilling. LearnVector is the most concrete manifestation of that philosophy to date.
What LearnVector will do
LearnVector is described as an AI-native learning startup. The plan is to build “agentic” AI tutors that act as personal instructors, adapting to each learner’s pace and drilling them until they can prove mastery of a skill. Unlike a generic chatbot that gives one-shot answers, these tutors will be persistent, rigorous, and tightly integrated with Coursera’s vast library of content and data on how people learn.
The initial target audience is white-collar workers and adults who need to update their skills as AI automates parts of their jobs. Ng plans to sell the platform to corporations, governments, and higher education institutions. The first courses are due in early 2027, leaving a three-year runway to develop the technology and curriculum. Pricing, course offerings, and the user interface are still being finalized.
The underlying AI models will come from a variety of third-party providers; LearnVector’s differentiator is not the model itself but the pedagogical structure and the proprietary data Coursera has accumulated from millions of learners. “The secret sauce is the data on how people learn,” Ng said, “combined with agents that can actually adapt the teaching in real time.”
The $100 million bet
Coursera’s investment gives it approximately a one-third stake in LearnVector, implying a valuation of about $300 million for a company that has no product yet. The sum is meaningful: it represents close to 13% of the $790 million in cash that Coursera held at the end of March 2025. Coursera CEO Greg Hart told Axios that the board saw a rare opportunity: “We had an opportunity to invest at the formation of a brand new company from one of the world’s foremost experts on AI. That was incredibly appealing.”
Because Ng is both chairman of Coursera and founder of LearnVector, the deal required special handling. A committee of independent directors reviewed the terms and approved the investment unanimously. The structure is fully disclosed, but a $100 million related-party transaction is unusual and deserves scrutiny. Still, Hart insists the process was rigorous and that the potential payoff justifies the risk.
The AI and jobs debate
Ng’s optimistic view enters a fiercely contested arena. On one side, studies from organizations like the World Economic Forum and McKinsey project that AI will displace millions of jobs while creating millions of new ones, often requiring higher-level skills. On the other side, researchers at MIT and elsewhere have found that the actual rate of job displacement due to AI so far has been lower than headlines suggest. A widely cited paper from 2023 noted that most AI adoption is still shallow — companies experiment with chatbots but rarely replace entire workforces.
Yet the anxiety is real. Surveys show that a majority of workers fear AI will make their roles obsolete. Recruiters are pivoting to specialize in AI-related hiring, while companies like IBM and Google have slowed hiring for non-technical roles, citing automation. In the past year, Amazon cut 18,000 jobs, Meta shed 21,000, and Block laid off 1,000 — all with explicit references to AI-driven efficiencies.
Ng counters that this is a temporary disruption. “Every major technological shift — the steam engine, electricity, the internet — was accompanied by predictions of mass joblessness,” he said. “In each case, productivity gains led to new industries and more employment overall. AI will be no different.”
The key, he believes, is reskilling. Workers need to learn not just how to use AI tools, but how to develop skills that complement AI: critical thinking, creativity, emotional intelligence, and domain expertise that machines cannot easily replicate. LearnVector aims to provide structured, effective pathways for that reskilling.
Education sector skepticism
LearnVector’s launch comes at a time when the role of AI in education is deeply contested. Teachers worry that tools like ChatGPT enable cheating and undermine traditional assessments. Many school districts have banned generative AI outright, while others struggle to craft effective policies. Parents are divided, with some embracing AI tutors for homework help and others fearing the erosion of foundational learning.
By targeting adults rather than schoolchildren, LearnVector sidesteps the most visceral debates. Yet it enters the reskilling market, where evidence of effectiveness is mixed. Massive open online courses (MOOCs), which Coursera helped pioneer, have famously low completion rates — often below 10%. Many learners who start a course never finish it. A structured, adaptive tutor could improve engagement, but the product doesn’t exist yet to prove it.
Moreover, corporate reskilling programs have produced uneven results. A 2022 study from Harvard Business School found that many employees who are told to “upskill” fail to find time or motivation, especially if the training is not directly tied to promotion or compensation. LearnVector will need to overcome these cultural and structural barriers, not just technical ones.
Why not just use a chatbot?
A natural question is why anyone would need a specialized AI tutor when they can already ask ChatGPT or Claude to explain anything. Hart’s answer is that a one-shot answer from a general chatbot is less effective than a system designed around a serious pedagogical view of a specific skill. “A tutor doesn’t just answer your question; it makes you solve problems, it corrects your misconceptions, it paces you through a curriculum,” he said. “That’s what LearnVector will do.”
The analogy is to the difference between reading a textbook and having a one-on-one teacher. Research consistently shows that adaptive tutoring can be highly effective — the famous Bloom’s two sigma problem noted that students taught with mastery learning and one-on-one tutoring performed two standard deviations better than those in conventional classrooms. The challenge is scaling that approach affordably. AI offers a way to simulate that experience at low cost.
Ng is well aware of the hurdles. “Building an agent that can truly adapt to a human learner in real time is one of the hardest problems in AI,” he admitted. “But if anyone can do it, the team we’re assembling — with the data from Coursera and the latest research — can.”
LearnVector is not Ng’s first bet on AI in education. His earlier company, deeplearning.ai, has trained over a million students. He also co-founded Landing AI, which focuses on industrial computer vision. Each venture has advanced his broader agenda of making AI accessible and useful. LearnVector might be the most ambitious yet because it stakes a claim on how people will navigate the very world AI creates. Andrew Ng gets until 2027 to show the tutor works, and the $100 million bet suggests he is confident the market is ready.