Qualcomm’s CEO, Cristiano Amon, unveiled a vision of the future of smartphones during the opening day of the Snapdragon Summit in Maui, Hawaii. Amon emphasized the pivotal role of artificial intelligence (AI) in the next generation of smartphones, describing how AI would run continuously on devices, opening up new possibilities and enhancing the user experience. The event showcased Qualcomm’s commitment to enabling powerful Generative AI models to run locally on smartphones, eliminating the need for a constant internet connection.
Amon recalled the evolution of smartphones, from their early days focusing on voice and messaging to the 4G era when they transformed into powerful computing platforms. He highlighted how AI was driving the next major transition in the smartphone industry, likening it to the profound changes seen with mobile devices.
In a market where smartphone sales have plateaued, and differentiation among devices has become challenging, Qualcomm is turning to Generative AI as a means to stimulate demand for high-end devices. Rather than just emphasizing raw performance, Qualcomm is focusing on the on-device AI capabilities of its Snapdragon 8 Gen 3, the latest flagship mobile processor, to revolutionize the user experience.
The Snapdragon 8 Gen 3 processor’s generative AI capabilities took center stage during the event. The chipset’s AI Engine supports multi-modal generative AI models, enabling features like “Stable Diffusion,” where users can suggest prompts and receive custom images in seconds. It also features “Photo Expansion,” which uses generative AI to expand image backgrounds, and “Video Magic Eraser,” capable of removing subjects from videos. Moreover, it supports a chatbot trained on Meta’s Llama 2 and can accept text, image, and voice inputs. Most notably, these features can be used even when the device is offline, highlighting the shift towards on-device AI.
While on-device AI features have existed for some time, they were limited in their capabilities, with advanced generative AI features only available through cloud-based solutions. Qualcomm’s approach to on-device AI offers numerous advantages, including enhanced user privacy, optimized battery life, rapid response times, and personalized experiences. Additionally, it is more cost-effective to implement on-device AI than to maintain large servers for complex machine learning models.
Though running complex AI models on devices is challenging at present, experts anticipate significant progress in the field as Generative AI continues to mature. Qualcomm is not alone in its pursuit of on-device Generative AI; tech giants like Google, Microsoft, and Apple are also making significant efforts to harness this emerging technology.
Rick Osterloh, Google’s Senior VP of Devices and Services, expressed enthusiasm for on-device Generative AI, describing it as a more distilled version of cloud models running locally on devices. As the technology develops, it promises to offer exciting possibilities for the future of smartphones, making them more personalized and capable than ever before.