On March 19th, Honor unveiled its new Honor Magic6 series smartphones, featuring Honor MagicOS 8.0 that comprehensively empowers platform-level AI, ushering in a new era of human-machine interaction.
Following the event, Honor CEO Zhao Ming was asked about the possibility of Honor's in-house developed cloud-based large models being implemented this year. Zhao Ming responded by stating that Honor will not pursue generic cloud-based large models; instead, its core focus remains on building edge AI capabilities.
In the future, cloud-based models will assist Honor's edge AI applications, but the core focus will remain on the edge.
For instance, platforms like ChatGPT, AIGC, and Sora are not currently the main focus of Honor's development. Despite allocating 11.5% of its budget to R&D, even if doubled, there is still a distinction in the strategic logic and development focus.
Zhao Ming also mentioned that for such generic large models, Honor would opt for collaboration with partners like Baidu, Tencent, Alibaba, and industry-leading networks and generic network model partners. The primary focus remains on the first, second, and some parts of the third layer of AI applications at the edge, with the fourth layer not being the core focus.
Furthermore, Zhao Ming believes that when building a platform-level AI, what can initially be presented to consumers is likely to be considered a silent cost because it involves underlying components. Honor has already introduced precise service recommendations based on geofencing, and moving forward, more AI services are expected to be introduced.
