- Yao described proprietary context and user trust as potential sources of durable AI product advantages.
- His WeChat example was hypothetical, not an announcement of new access to personal information or a product rollout.
Context beyond model scores
Tencent chief AI scientist Yao Shunyu argued that lasting competitive barriers are more likely to emerge in AI products than in the underlying models, according to Yicai’s original November 8 report on a dialogue at the 2026 Future Science Prize.
Asked by Zhang Yaqin what could keep users loyal to model companies, Yao pointed to information and context that competitors may not possess. He used a hypothetical AI inside WeChat to illustrate how knowledge of a user’s circumstances could differentiate a service even without the strongest model.
Trust and release choices
Yao also emphasized trust as agents move from answering questions to acting for users. The report treats permission, privacy and safe handling of information as unresolved requirements, rather than automatic benefits of an existing user ecosystem.
Asked whether a future leading Hunyuan model would remain open, Yao gave no direct yes-or-no commitment. He discussed practical incentives, community benefits and safety considerations, suggesting that open and closed models could continue to coexist.
Sources & context
Go to the original material. Company claims remain attributed to their sources.
01
