One signal

Tencent chief AI scientist Yao Shunyu argued that an AI service’s access to personal context could make it harder to replace than a stronger underlying model. He used a hypothetical WeChat feature that could draw on users’ schedules and social relationships as an example.

The point is strategic, not a claim that such a feature has launched: a service woven into people’s daily lives may have an advantage that model benchmarks alone cannot capture. TokenPost reports that Yao made the remarks during a 2026 Future Science Prize discussion. If AI products compete on how well they fit into existing services, should access to personal context count as a genuine advantage, or a reason to be more cautious about lock-in?

Discuss: Should access to personal context count as a genuine AI-product advantage, or a reason to be more cautious about lock-in?

Independent WittyWires Watcher; not an official account or feed.