Yuaiweiwu Lin Wang: From Education Going Online to Going Intelligent — When Will the AI-Powered All-Capable Personal Tutor Arrive? | Gaorong Future
"Yuaiweiwu" — dancing with love, and dancing with AI.
AI-powered applications in education represent a massive opportunity.
As large language models and related AI technologies enter a new phase of transformation, a team with battle-tested experience in both AI and education — and a track record of delivering results — has embarked on another journey of exploration and creation.
This time, they aim to "dance with love" and "dance with AI."
In 2023, Huaiting Zhang, former co-founder of Gaotu Techedu and former head of Baidu's Fengchao advertising platform, joined forces with Wei Liu, former general manager of Gaotu Class, and other core founding team members to launch Yuaiweiwu, an AI education company. To date, the company has secured investment from top-tier institutions including Gaorong Ventures, Qiming Venture Partners, Monolith, and HSG.
Recently, Lin Wang, co-founder and CTO of Yuaiweiwu, spoke at a joint livestream hiring event organized by Gaorong Ventures and BOSS Zhipin, sharing with young AI talent the company's founding vision and their thinking on the arrival of the AI education era, potential product forms, and growth paths. Wang previously led AI teams at Baidu, Gaotu, and Ant Group, and has published over ten papers at top AI conferences.
The following is Wang's presentation, edited for clarity:
When we named the company Yuaiweiwu, "ai" carried two meanings. First, the essence of education is love. Second, "AI" is a homophone of "ai" — our founding mission is to empower education through AI.
Why did we choose AI education as our entrepreneurial direction? First, China's education industry has completed its full digitalization, providing an excellent foundation. Second, education represents a massive, non-negotiable market in China. Third, our team brings together leading figures from both education and AI with substantial accumulated experience.
Over the past three decades, China's education and training industry has passed through three developmental stages. The first, beginning in 1993, was the offline education era represented by New Oriental. The second, around 2015, was the online education era represented by Gaotu and Yuanfudao. The third stage — the AI education era we are entering today — is now underway.
Each major stage solved one core problem.
The leap from offline to online education delivered value through greater accessibility. Internet technology enabled a good teacher to reach not just dozens of students, but thousands or even tens of thousands in a single session. Confucius spoke of "education for all without discrimination"; online education allowed every student more equal access to quality educational resources.
But online education did not solve the problem of personalization. When one teacher lectures to thousands of students online, the content is identical. We know that the best education is "teaching students in accordance with their aptitude" — this is the core problem AI must solve.
Excitingly, large language models have emerged as one of the biggest variables in education today. This generation of AI technology makes it possible to scale industries that previously required intensive personalized service. How to understand this? Private assistants used to be highly personalized services; ChatGPT essentially gives everyone a private assistant.
With large models, a personalized AI tutor dedicated to you has finally become possible. Yuaiweiwu aims to give everyone an AI tutor suited to their needs — one who is not only deeply knowledgeable but can also converse with you, accompany you through lessons, and walk you through problems.
So what does the future of education look like with AI? What is our goal?
Our answer: we hope to build for everyone an all-capable personal tutor who understands you, is always available, and is affordable — providing personalized educational companionship tailored to individual circumstances.
First, understanding you. Capable of providing personalized, dynamic solutions based on individual interests, developmental stage, efficiency, potential, state, personality, and more.
Second, always available. Compared to current online education's dual-teacher model, an AI tutor solves not just spatial constraints but temporal ones as well — a teacher available 24/7, responding in real time.
Third, affordable. As computing and model costs decline, AI tutor costs are expected to decrease further. Our estimate: one hour of daily learning with an AI tutor costs roughly $500 per year. We believe that in two to three years, this figure could drop by another order of magnitude, making a personal tutor accessible to everyone.
Finally, all-capable. Traditionally, different subject teachers have varying understanding of any given student. In the future, the all-capable tutor won't be limited to a single subject or body of knowledge, but will be able to cultivate comprehensive competencies and even provide lifelong companionship.
We believe an AI tutor meeting these capabilities could emerge within one to three years.
With our goal established, what is our development path? Yuaiweiwu's plan is to first build a closed business loop, then achieve AI-driven replacement, and ultimately form a new industry paradigm.
First, we build a closed online education business loop. As a startup, we have been focused on healthy, stable development from day one, hoping to generate stable cash flow early on. More importantly, we want to establish a core business loop from the outset.
Looking back at companies that ultimately succeeded through each technological transformation, beyond technology-driven growth, they all had core businesses. For example, in the 2012 deep learning revolution, companies that truly grew alongside recommendation algorithms and related technologies — Pinduoduo, ByteDance, Xiaohongshu — all had core businesses at their center.
From its founding, Yuaiweiwu had users, data, and revenue, entering a positive feedback loop on that basis.
AI-driven replacement will unfold in two phases. In the first phase, we use AI to optimize every link in the existing online education chain. We aim to rebuild each of these links with AI to improve operational efficiency and reduce costs.
In the second phase, we use AI to build a student-centered AI tutor, shifting from educational digitalization to educational intelligence.
Of course, we are advancing both phases in parallel. Early on we devoted more resources to the first phase; currently both directions are receiving heavy investment. Ultimately, we hope to land a new industry paradigm through this gradual approach, moving forward steadily.
To date, we have not only achieved scalable revenue growth over the past two years but also explored numerous AI application scenarios. Only with growing business can we continuously invest resources into new R&D and product directions.
Of course, applying AI technology in education today still faces notable shortcomings and challenges, such as the well-known problem of LLM hallucinations.
In our entrepreneurial journey, we focus on matching product and business needs with existing technical capabilities. First, in scenario selection, we emphasize alignment with current model capabilities, starting with domains where large models perform well. For example, we began with elementary school English AI classes, launching our AI English teacher product "Aixue."
Second, we build appropriate openness into curriculum design. On one hand, we maintain rigorous teaching systems and learning frameworks to ensure educational outcomes; on the other, we allow appropriate openness and personalization in teaching segments where large models can contribute.
We firmly believe technology develops exponentially. No one anticipated ChatGPT would emerge so quickly; after GPT-3, GPT-4 advanced dramatically in short order. We once thought mathematical reasoning remained beyond large language models' reach, but their ability to solve math and physics problems has now improved substantially.
As large models' reasoning capabilities and multimodal technologies advance, AI will create more surprises, unblocking pain points in the existing education chain and driving further explosive growth in industry applications.
From its earliest days, Yuaiweiwu defined its corporate culture around two core metrics — "love quotient" and "AI quotient." Love quotient: using love to help everyone become a better version of themselves, encompassing not only the students and parents we serve but every employee who joins us. AI quotient: becoming a company truly driven by AI technology.
"Love yourself, love your partners, love the world." Yuaiweiwu has already attracted outstanding talent from Apple, ByteDance, Alibaba, Tencent, and other companies. Our AI team has deep expertise in large model algorithms, natural language processing, speech technology, computer vision, platform architecture, and machine learning.
Yuaiweiwu is also recruiting more aspiring young people to join us on the journey of exploring the AI education era.
Resume Submission
