Summary of Key Points
Jiuhen Technology (founded by Wang Yuan, former vice president of NetEase) has been established for less than two years. The company has evolved from a tool called remio, which focused on personal digital memory, to rOS, an agent operating system. Its goal is to become the “contextual infrastructure for the agent ecosystem” (an analogy: large models are like engines, while context is the fuel). The discussion covers aspects such as product differences, commercialization strategies, entrepreneurial logic, team changes, and investors' concerns:
- Product Upgrade: remio was a “personal digital memory library” that aggregated fragmented data from files, emails, meeting recordings, etc. on users’ computers, allowing AI to help with tasks like finding information and writing content. The upgraded rOS, however, functions more like an “agent operating system.” It can:
- Act proactively: Unlike traditional AI assistants that require constant prompting, rOS can automatically capture new emails, classify them, and respond without needing frequent queries from the large model.
- Have its own independent infrastructure: While other tools are temporary solutions with no dedicated databases, rOS’s native intelligent applications (aApps) have their own storage and can run in the background for extended periods.
- Coexist with traditional OSes: It doesn’t replace Windows or Mac; instead, it works on top of them, similar to how cloud operating systems operate on Linux, making AI assistants more efficient.
- Commercialization: Jiuhen targets a “to-customer” (ToC) market overseas and a “to-business” (ToB) market in China. It charges based on the amount of tokens used, which reflects the resource-intensive nature of AI processing. The company has started charging C-end users this year, with a 80%+ renewal rate in the second month and a 60%+ rate in the fourth month, indicating good user retention. However, B-end sales are more volatile.
- Entrepreneurial Logic: Wang Yuan left NetEase to start Jiuhen in 2024 at the right time, as the internet industry is entering a new phase driven by AI innovations (similar to how mobile phones opened up a new era). Small teams like Jiuhen can compete with giants by adapting quickly. With its smaller size (around 10 employees), management is more manageable, and Wang Yuan can now focus on coding instead of spending 70% of his time on administration.
- AI’s Impact on Work: AI has transformed team collaboration:
- Efficiency improvements: One person can handle multiple tasks previously requiring four people. The role of programmers has evolved from simply writing code to designing automated workflows.
- Reduced need for expansion: AI increases productivity, so fewer engineers are needed to achieve the same results. This means the core development team doesn’t need to expand rapidly.
- Investors’ Concerns: Investors worry that Jiuhen’s technology might be overshadowed by free features offered by larger companies. Jiuhen emphasizes that its “context management” service is a fundamental component of the agent ecosystem, essential for all intelligent agents, not just an optional add-on. They need to prove that their product is a necessary part of the industry.
In summary, Jiuhen Technology’s story illustrates how AI startups can seize technological opportunities and use flexibility to compete with giants by focusing on building essential infrastructure. Although commercialization is still in its early stages, the company’s differentiated products and user loyalty give it a competitive advantage. The key lies in proving that its services are indispensable within the agent ecosystem.