Another celebrity founder enters the AI podcast arena, and Sequoia China makes a bet. Can they create waves this time?

Previously, in our topic selection, we have observed multiple cases of "celebrity entrepreneurs" entering the AI podcast field, including ChatPods launched by Zhang Yueguang and Laifu launched by Jiao Ke.

The dual - platform download volume of ChatPods and Laifu in the past 90 days | Image source: Diandian Data

However, as time passed, judging from the data, the performance of both products was not ideal. Diandian Data shows that ChatPods had 35,000 global downloads in September. Due to its small scale, third - party data did not capture active users, and its monthly revenue was less than $100. Laifu, which was launched later, had even lower data, with only about 2,000 downloads in September, and no MAU and revenue data were captured.

The work resume of Ethan KJ Li | Image source: LinkedIn

Even though the former encountered setbacks, it did not dampen the enthusiasm of later entrants. Recently, Ethan KJ Li, who once served as the CEO of ByteDance's smart education business line, has also entered the "AI podcast" field, but his approach is different from the previous two.

Ethan KJ Li's product is called Aibrary. It was launched for testing on the US App Store on April 23 this year and officially launched on September 23. According to the introduction on the official website, the core function of Aibrary is to transform/reshape books into personalized podcasts and serve personal learning scenarios through customized learning paths and interactive tutoring. Before this product was launched, its parent company, Ouraca, had completed multiple rounds of financing. Investors include well - known VCs such as Sequoia Capital China, Initium Capital, and Factorial Fund. With the background of "celebrity entrepreneur + well - known VC", Aibrary was born with a "silver spoon in its mouth".

From previous analyses, neither ChatPods' attempt to use AI to improve the efficiency of podcast listening nor products like Laifu's attempt to directly generate podcasts with AI to compete with human hosts seem to be viable. However, changing the approach and using AI podcasts to serve knowledge acquisition scenarios has been proven effective by products like NotebookLM. How can Ethan KJ Li's attempt differentiate itself from star products like NotebookLM?

1. This is the most reasonably designed AI podcast product we've ever experienced

Overall, Aibrary is mainly targeted at personal learning and improvement scenarios. Compared with products like NotebookLM and Speechify that we've observed, Aibrary has two core differences: 1. A relatively complete recommendation system and content system; 2. Using AI podcasts as the main content - carrying method.

The recommended content on the Aibrary homepage (Figures 1, 2, 3), and the second primary tab, Nova (Figure 4)

When entering Aibrary, users need to complete a 6 - step registration process. In addition to the usual age, gender, and topics of interest, users are also asked to choose their admired celebrities and their goals for using Aibrary. The user's choices will determine the recommended content on the Home page. According to the sharing of co - founder Wu Jundong, Aibrary has established a recommendation algorithm to recommend content according to user preferences. In addition, a search function is provided on the home page. Users can search for books by using book titles or topics of interest as keywords and then further study the books.

In addition to the home page with the above - mentioned recommendation and search functions, Aibrary's primary tabs also include the Chatbot "Nova" and personal profiles. In "Nova", users can have a conversation with the ChatBot (as shown in Figure 4) and put forward more personalized requirements. Aibrary's ChatBot has three roles with different divisions of labor: Nova, which mainly interacts with users; Orion, which provides knowledge management services (such as inquiring about book - related content and building knowledge graphs); and Atlas, which is mainly responsible for breaking down goals and helping users with reviews.

Based on my experience, most of the responses are completed by Nova. Only in some relatively complex questions do Orion and Atlas offer their own insights from their respective perspectives, and their participation is not high. As for why the task of sharing context is completed by three Chatbots, we haven't clearly understood the reason yet.

Orion and Atlas will intervene in more complex answers

Whether it's the home - page recommendations, search results, or ChatBot Q&A content, they are all divided into three types: book lists that may meet user needs, details of specific books selected by users (including written introductions, abstracts & two audio segments of podcasts, and book purchase links), and the idea Twin Podcast generated by the system when users want to further understand a specific topic in the selected book.

Book lists (Figures 1, 2), book details (Figure 3) | Image source: Aibrary

After users select a book, the details are not the traditional e - books but "two audio segments + purchase links" (as shown in the right - hand picture above). Both audio segments are about 8 - 10 minutes long. One is a "monologue" summary, and the other is a "dialogue" podcast. The former is a brief introduction to the book's outline, and the latter analyzes the outline in the form of a podcast. Both audio segments are provided by the system.

In terms of copyright, since Aibrary only generates summaries and podcasts based on the content of the books and does not use the original book covers, professional lawyers say that this is in a legal gray area, and the copyright risk is low both in China and overseas.

After listening to the two audio segments, users will have a preliminary understanding of the book. If they are interested, users can choose to click the purchase link below to jump to Amazon to buy the physical book or click "Make idea Twin Podcast" to use AI to generate another more personalized podcast that better suits their interests and gain a deeper understanding of the book content.

Book selection (Figure 1), AI host voice selection (Figure 2), user information and voice cloning (Figures 3, 4)

This AI - generated podcast is called "Idea Twin Podcast", which is also the biggest difference we've noticed between Aibrary and other similar products.

The Idea Twin Podcast is a real - time generated two - person podcast. The two participants are an AI host and the "user's avatar". Users first need to select a book and choose the voice of the AI host, then fill in personal information such as educational background, major, occupation, MBTI, personal description, etc., and clone their own voices. The system will generate an "avatar" for the user based on the input information to have a conversation with the AI host as the other role. Generating an Idea Twin Podcast costs 100 Credits. Non - subscribed users can only generate one podcast.

The Idea Twin Podcast follows a model where the host (AI voice) asks questions and the "avatar" (user's voice) answers. Users can think about the questions raised by the AI while listening to the Q&A and refer to the answers given by their "selves", which promotes users' active thinking.

According to the developer, "On Aibrary, you're no longer just listening to a podcast; you're co - hosting a show with AI." Based on my experience, the content of the "avatar"'s answers does refer to the filled - in personal information, and the user's own voice enhances the sense of immersion. However, this sense of immersion is still a bit superficial at present. During the listening process, I still feel like two AIs are having a conversation, and it's difficult to put myself in the role of the "avatar".

The corresponding relationship between Aibrary's functions and content | Produced by Baijingchuhai

In previous topics, we have observed multiple "AI + audio podcast" products. We created a quadrant chart based on the purpose (knowledge acquisition/creation tool) and the way AI is integrated (generation/understanding) to clarify the positioning of each "AI + podcast" product | Produced by Baijingchuhai

From the two charts of product design and industry positioning, Aibrary has indeed found a relatively blank position. Previously, products like Doubao had difficulty digesting long books with short podcasts. Compared with Speechify's word - for - word reading, which can absorb and digest long books, Aibrary provides a "from short to long" guiding experience and lowers the threshold of "reading" through podcasts.

The above is basically the stage where the product guides users into learning. However, the functions of formulating learning paths and interactive tutoring mentioned on the official website and in the podcasts, which are used to continuously supervise users to achieve learning goals, are not very obvious. As of now, the project still seems to be in an "unfinished" state. However, as a product that was officially launched at the end of last year, it has shown certain differentiation and a reasonable path after the experience. It has shown certain attractiveness compared with previous products. Behind this is the in - depth thinking of several founders who have been deeply involved in the education industry for many years.

2. Why did the star founders with deep roots in K12 education take a fancy to the lifelong learning track?

Ethan KJ Li, the founder of Aibrary, developed the K12 education collaboration platform "Jike Big Data" for schools and educational institutions in his early years. It was later acquired by ByteDance, and he joined ByteDance as the CEO of the smart education business. After leaving ByteDance in 2022, he briefly served as a partner at Initium Capital. In January 2025, he founded Ouraca in Silicon Valley. The other two co - founders are also deeply involved in the education industry, including Wu Jundong, the former director of international investment and strategic development at TAL Education Group, and Zhang Qiming, the former head of the education middle - platform at ByteDance.

From the resumes of the three founders, they were all "pioneers" in the domestic education and training industry. In the AI era, instead of continuing in the K12 track, they focused on the relatively niche track of "adult lifelong learning". This change stems from their thinking about learning in the AI era.

Founder Ethan KJ Li said in an article:

"In the near future, AI can almost answer everything, so 'answers' are gradually becoming less important. 'Questions' and the thinking process behind them are more valuable. Therefore, education in the AI era also needs to change from 'content instillation' to 'cognitive reshaping'. With this in mind, how to stimulate and guide users to think for themselves becomes particularly important.

On the other hand, in addition to thinking, feedback is also very important. Ethan KJ Li also mentioned: "Nowadays, inputting more content to users may not necessarily lead to real improvement in their abilities. Instead, designing an effective feedback mechanism can form a closed - loop learning system and allow people to grow on their own."

——Paraphrased from Ethan KJ Li's "100 Days in Silicon Valley, Written Before the Official Launch of Aibrary | Initium Sharing"

After reading Ethan KJ Li's sharing, we can summarize that he believes that in the AI era, personalized content, stimulating user thinking, establishing a rapid feedback mechanism, and establishing a closed - loop of 'content - action - feedback - growth' are the iterative directions for education products. According to the sharing of co - founder Wu Jundong in the podcast, based on these three core goals, the iteration of Aibrary is also divided into three stages: establishing a personalized recommendation mechanism, generating personalized content that stimulates thinking, and establishing a customized learning process.

Based on the above analysis of Aibrary, currently, Aibrary has established a personalized content distribution mechanism through a graded recommendation system and content system. Choosing the form of podcasts takes into account the fragmented usage habits of contemporary users. The Idea Twin Podcast corresponds to personalized content that stimulates thinking. How to establish a good feedback mechanism and use AI to enrich more personalized content that stimulates thinking will be the focus of the team's next iteration.

Conclusion

Combined with the founders' sharing, Aibrary actually reflects and tests the ideas of three senior education practitioners. In terms of the monetization strategy, most books on Aibrary can only be accessed by subscribed users. The Idea Twin Podcast even requires users to consume credits to generate, and there are quantity limits even for subscribed users. After the experience, it's basically impossible to use the product without a subscription.

In terms of pricing, compared with the traditional audio - book app Audible (annual subscription price: $159.99), Aibrary's annual subscription price is much lower ($6.99 per week, $89.99 per year, with a 7 - day free trial). Moreover, it also has a process of guiding users to learn. However, whether Aibrary can truly attract users to stay and pay for continuous use with these differences remains to be further observed.

Reference Articles

1. 100 Days in Silicon Valley, Written Before the Official Launch of Aibrary | Initium Sharing

2. Ouraca: We Want to Redefine Lifelong Learning with AI

The data comes from third - party platforms such as SimilarWeb, Diandian Data, Semrush, and