MediaTek’s next-generation flagship processor may come with the ability to support generative artificial intelligence (AI), making smartphones more secure and faster. The chipmaker recently announced a collaboration with Meta under which MediaTek will equip with chips with Llama 2 large language model (LLM) to make them faster than ever. Not only generative AI will increase privacy on a device running MediaTek’s upcoming chipset, but it will also allow for better processing of data.
Korean publication ITHome has reported that MediaTek’s first Llama 2-enabled chipset could be called Dimensity 9300. It will feature a software stack optimised to run Meta’s LLM and enhanced hardware for the proper functioning of generative AI-based applications. In other words, the chipset will use additional architectures to allow room for generative AI to work without compromise. In addition to ensuring better privacy, generative AI in the chipset will give users of the associated device more control over their personal information because the applications will be able to run locally instead of needing to send data to the cloud for processing.
According to the report, the MediaTek Dimensity 9300 chipset will bring a radical change in design, favourable for the inclusion of Arm Cortex-X4 and A720 CPU cores, along with the Immortalis-G720 graphics processor. The change in the technical specifications of the chipset as compared to previous counterparts could allow MediaTek to take a leap in terms of flagship performance. However, its biggest rival Qualcomm is also on the same track. The world’s biggest mobile chip manufacturer is also working with Meta to implement Llama 2 into its chips.
But, irrespective of the competition between MediaTek and Qualcomm, customers buying a phone using processors from either manufacturer can look forward to secure mobile ecosystems, faster device performance, and protection against online threats, which is the biggest challenge for the Android operating system. Generative AI will also allow devices to become more interactive, minimising the navigation that is currently required to carry out a task on a phone. It will also help digital voice assistant companies leverage the capabilities of local processing to offer faster responses to customers without relying on the cloud.
Author Name | Shubham Verma