QUALCOMM: Smart Computing

QUALCOMM, a world leader in the development and commercialization of core technologies for the wireless industry, highlighted intelligent computing as a global trend.

Advances in processor technologies have enabled the distribution of complex workloads across the network, with more computing performed at edge devices where data is generated. 

Given the proximity to raw data, edge computing enables smarter processing, which reduces response time, improves privacy and security, and allows for greater personalization. 

Thus, according to QUALCOMM, with increased processing power, mobile devices are becoming a ubiquitous AI platform, with large and complex generative AI algorithms running on the device, enabling contextual and on-demand AI use cases. 

As 5G wireless connectivity complements on-device generative AI, edge devices enable productivity-enhancing use cases, while intelligently processing and sharing data with cloud-based applications as needed. 

Intelligent Computing

Leveraging the smartphone base and the scale of mobile devices, QUALCOMM envisions generative AI becoming ubiquitous and expanding beyond smartphones into industries and applications such as computing, IoT, XR and automotive.

Also, as part of intelligent computing, complex large language models (LLMs), big vision models (LVMs) and other generative AI models that can generate new content are beginning to change the user experience landscape. 

LLMs (e.g., GPT-4 and Llama2) are useful for text-based natural language processing applications, such as answering queries, summarizing, and creating documents, while LVMs (e.g., Stable Diffusion and ControlNet) are useful for image and video processing. 

They are revolutionizing traditional methods of search, content creation, recommender systems and personalized digital assistants, offering significant improvements in consumer usability and productivity.

QUALCOMM believes that a variety of innovative business and consumer use cases will emerge from generative AI, especially as LLMs and LVMs run on devices and incorporate multiple modalities (such as text, voice, camera, camera, infrared, RADAR and LiDAR sensors) and bring the benefits of immediacy, privacy, security and personalization to consumers.